System

Abstract

A system is provided.SOLUTION: A system comprising: means for capturing a user's telephone conversation in real-time as voice data and converting the captured voice data to text data; means for analyzing the text data to determine a likelihood of fraud; means for notifying the user if a likelihood of fraud is determined; means for providing information to police if a likelihood of fraud is determined; and artificial intelligence agent means for continuing the conversation on behalf of the user.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The technology of the present disclosure relates to a system. [Background technology]

[0002] Patent document 1 discloses a persona chatbot control method performed by at least one processor, the method including the steps of receiving a user utterance, adding the user utterance to a prompt including an instruction sentence related to a description of the chatbot character, encoding the prompt, and inputting the encoded prompt into a language model to generate a chatbot utterance in response to the user utterance. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Japanese Patent Publication No. 2022-180282 Summary of the Invention [Problem to be solved by the invention]

[0004] In modern society, telephone fraud is becoming more sophisticated every year, and the damage caused by it is increasing. Elderly people and single-person households are particularly vulnerable, and self-defense alone has its limits. This has resulted in many people suffering financial and psychological damage. This invention aims to prevent users from becoming victims of telephone fraud and to deter fraud crimes by facilitating rapid police response. [Means for solving the problem]

[0005] The present invention provides a system that includes a means for capturing a user's telephone conversation as voice data in real time and converting the captured voice data into text data. It also includes a means for analyzing the text data to determine the possibility of fraud, and a means for notifying the user if a fraud is determined to be possible. The system also includes a means for providing information to the police if a fraud is determined to be possible, and an artificial intelligence agent means for continuing the conversation on behalf of the user. In this way, the system supports a rapid police response while protecting the user's safety.

[0006] "User" means an individual or organization that uses a telephone.

[0007] "Voice Data" means data that is a digital recording or representation of a user's speech.

[0008] "Text data" refers to voice data converted into character information.

[0009] "Natural language processing technology" refers to a series of technologies that enable computers to understand, interpret, and generate human language.

[0010] "Means for determining the likelihood of fraud" refers to algorithms or technologies that analyze text data and detect signs of fraud.

[0011] "Means for notifying the user" refers to a function or device for issuing a warning to the user when it is determined that there is a high possibility of fraud.

[0012] "Means for providing information to the police" refers to functions and devices for contacting and providing relevant information to the police when it is determined that there is a high possibility of fraud.

[0013] "Artificial Intelligence Agent" means a system or program that simulates natural conversation on behalf of a user and maintains a dialogue with a fraudster. [Brief explanation of the drawings]

[0014] [Figure 1] 1 is a conceptual diagram showing an example of the configuration of a data processing system according to a first embodiment. [Figure 2] 1 is a conceptual diagram showing an example of main functions of a data processing device and a smart device according to a first embodiment. [Figure 3] FIG. 10 is a conceptual diagram showing an example of the configuration of a data processing system according to a second embodiment. [Figure 4] FIG. 10 is a conceptual diagram showing an example of main functions of a data processing device and smart glasses according to a second embodiment. [Figure 5] FIG. 10 is a conceptual diagram showing an example of the configuration of a data processing system according to a third embodiment. [Figure 6] FIG. 11 is a conceptual diagram showing an example of main functions of a data processing device and a headset-type terminal according to a third embodiment. [Figure 7] FIG. 10 is a conceptual diagram showing an example of the configuration of a data processing system according to a fourth embodiment. [Figure 8] FIG. 10 is a conceptual diagram showing an example of main functions of a data processing device and a robot according to a fourth embodiment. [Figure 9] 1 shows an emotion map onto which multiple emotions are mapped. [Figure 10] 1 shows an emotion map onto which multiple emotions are mapped. [Figure 11] FIG. 3 is a sequence diagram showing a processing flow of the data processing system according to the first embodiment. [Figure 12] FIG. 10 is a sequence diagram showing the flow of processing in the data processing system in Application Example 1. [Figure 13] FIG. 10 is a sequence diagram showing the flow of processing in the data processing system according to the second embodiment when an emotion engine is combined. [Figure 14] FIG. 10 is a sequence diagram showing the flow of processing in the data processing system in Application Example 2 when an emotion engine is combined. DETAILED DESCRIPTION OF THE INVENTION

[0015] An example of an embodiment of a system according to the technology of the present disclosure will be described below with reference to the accompanying drawings.

[0016] First, the terms used in the following description will be explained.

[0017] In the following embodiments, a coded processor (hereinafter simply referred to as a "processor") may be a single arithmetic device or a combination of multiple arithmetic devices. Furthermore, a processor may be a single type of arithmetic device or a combination of multiple types of arithmetic devices. Examples of arithmetic devices include a CPU (Central Processing Unit), a GPU (Graphics Processing Unit), a GPGPU (General-Purpose computing on Graphics Processing Units), and an APU (Accelerated Processing Unit).

[0018] In the following embodiments, a coded RAM (Random Access Memory) is a memory in which information is temporarily stored and is used as a working memory by a processor.

[0019] In the following embodiments, the coded storage is one or more non-volatile storage devices that store various programs, various parameters, etc. Examples of non-volatile storage devices include flash memory (SSD (Solid State Drive)), magnetic disks (e.g., hard disks), and magnetic tapes.

[0020] In the following embodiments, a communication I / F (Interface) with a symbol is an interface including a communication processor, an antenna, etc. The communication I / F controls communication between multiple computers. Examples of communication standards applied to the communication I / F include wireless communication standards including 5G (5th Generation Mobile Communication System), Wi-Fi (registered trademark), Bluetooth (registered trademark), etc.

[0021] In the following embodiments, "A and / or B" is synonymous with "at least one of A and B." In other words, "A and / or B" means that it may be only A, only B, or a combination of A and B. Furthermore, in this specification, the same concept as "A and / or B" is also applied when three or more things are expressed connected by "and / or."

[0022] [First embodiment]

[0023] FIG. 1 shows an example of the configuration of a data processing system 10 according to the first embodiment.

[0024] 1, a data processing system 10 includes a data processing device 12 and a smart device 14. An example of the data processing device 12 is a server.

[0025] The data processing device 12 includes a computer 22, a database 24, and a communication I / F 26. The computer 22 is an example of a "computer" according to the technology of the present disclosure. The computer 22 includes a processor 28, a RAM 30, and a storage 32. The processor 28, the RAM 30, and the storage 32 are connected to a bus 34. The database 24 and the communication I / F 26 are also connected to the bus 34. The communication I / F 26 is connected to a network 54. Examples of the network 54 include a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0026] The smart device 14 includes a computer 36, a reception device 38, an output device 40, a camera 42, and a communication I / F 44. The computer 36 includes a processor 46, a RAM 48, and a storage 50. The processor 46, the RAM 48, and the storage 50 are connected to a bus 52. The reception device 38, the output device 40, and the camera 42 are also connected to the bus 52.

[0027] The reception device 38 includes a touch panel 38A, a microphone 38B, and the like, and receives user input. The touch panel 38A detects contact with an indicator (for example, a pen or a finger) to receive user input by the touch of the indicator. The microphone 38B detects the user's voice to receive user input by voice. The control unit 46A transmits data indicating the user input received by the touch panel 38A and the microphone 38B to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the data indicating the user input.

[0028] The output device 40 includes a display 40A and a speaker 40B, and presents data to the user 20 by outputting the data in a form of expression that the user 20 can perceive (for example, audio and / or text). The display 40A displays visible information such as text and images in accordance with instructions from the processor 46. The speaker 40B outputs audio in accordance with instructions from the processor 46. The camera 42 is a compact digital camera equipped with an optical system including a lens, aperture, and shutter, and an imaging element such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor.

[0029] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 control the exchange of various information between the processor 46 and the processor 28 via the network 54.

[0030] FIG. 2 shows an example of the main functions of the data processing device 12 and the smart device 14.

[0031] 2, in the data processing device 12, a specific process is performed by the processor 28. A specific processing program 56 is stored in the storage 32. The specific processing program 56 is an example of a "program" according to the technology of the present disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific process is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific processing program 56 executed on the RAM 30.

[0032] The storage 32 stores a data generation model 58 and an emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[0033] In the smart device 14, the processor 46 performs the reception output process. The storage 50 stores a reception output program 60. The reception output program 60 is used in conjunction with the specific processing program 56 by the data processing system 10. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output process is realized by the processor 46 operating as the control unit 46A in accordance with the reception output program 60 executed on the RAM 48.

[0034] Next, a description will be given of the specific processing performed by the specific processing unit 290 of the data processing device 12. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0035] The present invention is a system for monitoring users' telephone conversations in real time and automatically detecting fraud risks. Detailed embodiments for specifically implementing the system are described below.

[0036] Voice Input Processing

[0037] When a user makes or receives a call, the terminal automatically captures the call's audio data in real time, which is then converted and compressed into an optimal format by the terminal and then sent to the server.

[0038] Converting audio data to text data

[0039] The server receives the voice data sent from the device and converts it into text data using speech recognition technology, which includes a highly accurate natural language processing (NLP) engine.

[0040] Conversation analysis

[0041] The server analyzes the text data and extracts contextual information, entities (such as names of people and places), sentiment, and intent from the conversation. This analysis process is performed using natural language processing techniques. The analyzed information is then passed to a fraud detection algorithm.

[0042] Fraud Detection Algorithms

[0043] The server evaluates the likelihood of fraud based on the analyzed text data. Fraud detection algorithms look for specific keywords and phrases (e.g., "urgent," "transfer," "urgent") and check whether they match previous fraud cases. If potential fraud is detected, it is flagged by the server.

[0044] User notification function

[0045] If the server determines that there is a high possibility of fraud, it immediately sends a warning signal to the device, which then displays a warning message on the screen or makes an audio notification based on the received warning signal.

[0046] Police contact function

[0047] If the server determines that there is a possibility of fraud, it will promptly provide the police with relevant data (e.g., the content of the conversation, the caller's phone number, the date and time of the conversation, etc.) This information will be automatically forwarded to the police's receiving system in an appropriate format.

[0048] AI conversational agent

[0049] If it determines that there is a high possibility of fraud, the server sends an instruction to the device to activate an AI conversation agent. The device then activates the AI ​​conversation agent based on the received instruction and continues the conversation on behalf of the user. The AI ​​conversation agent generates natural conversation and continues to converse with the fraudster, buying time and supporting police response.

[0050] Specific examples

[0051] For example, if a user answers a phone call and the caller says, "I'm your son, please send me money now," the device immediately sends the voice data to the server. The server converts the voice data into text data and analyzes it using an NLP engine. If keywords such as "son" and "please send me money" are detected, it is determined to be a fraudulent call. A warning is immediately issued to the user and the police are notified. Furthermore, an AI conversation agent is activated to continue the conversation on the user's behalf, allowing time for the police to respond.

[0052] Thus, the present invention provides a system that keeps users safe from telephone fraud and allows for fast and effective police response.

[0053] The processing flow will be explained below.

[0054] Step 1:

[0055] A user makes or receives a call. The terminal captures the voice data of this phone conversation in real time. The voice data is compressed within the terminal and prepared to be sent to the server for further processing.

[0056] Step 2:

[0057] The device transmits the captured audio data to the server, and this communication uses a secure protocol to ensure data confidentiality.

[0058] Step 3:

[0059] The server receives the received voice data and converts it into text using speech recognition technology. The speech recognition model utilizes a highly accurate natural language processing (NLP) engine.

[0060] Step 4:

[0061] The server analyzes the text data and uses an NLP engine to extract contextual information, entities (people's names, places, etc.), sentiment, and intent from the conversation. The results of this analysis are passed to a fraud detection algorithm.

[0062] Step 5:

[0063] The server runs fraud detection algorithms to evaluate potential fraud patterns. If certain keywords or phrases (e.g., "urgent," "transfer," "urgent") are detected, they are compared against a database of past frauds to determine the likelihood of a match.

[0064] Step 6:

[0065] If the server determines that there is a high possibility of fraud, it will raise a warning flag and record the possibility of fraud. This flag will be immediately notified to the user.

[0066] Step 7:

[0067] When a warning flag is raised, the server immediately sends a warning signal to the terminal, which then displays a warning message on the screen or notifies the user by voice.

[0068] Step 8:

[0069] The server collects relevant information about the flagged call (e.g., the caller's phone number, the content of the conversation, the date and time, etc.) and prepares it for automatic provision to the police. This information is then forwarded in an appropriate format to the police's receiving system.

[0070] Step 9:

[0071] If the server determines that there is a high possibility of fraud, it sends an instruction to the device to activate an AI conversation agent. Based on this instruction, the device activates the AI ​​conversation agent and continues the conversation on behalf of the user.

[0072] Step 10:

[0073] The AI ​​conversational agent generates natural conversations on behalf of the user and continues the dialogue with the fraudster, buying time and assisting law enforcement. For example, the AI ​​agent can ask, "Okay, where should I send the money?" to elicit additional information from the fraudster.

[0074] Step 11:

[0075] All conversation data is recorded by the server and stored in a format that can be analyzed later. Newly detected fraud patterns and techniques are added to the database and used as training data to improve future detection accuracy.

[0076] Through these steps, the system can protect users from telephone fraud in real time and assist police in responding quickly.

[0077] Example 1

[0078] Next, a description will be given of Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0079] Telephone scams are becoming more sophisticated every year, making vulnerable individuals such as the elderly particularly vulnerable to fraud. Current manual countermeasures often delay detection and response to scams, often failing to detect them until actual harm has occurred. Furthermore, continuing to talk to scammers is mentally stressful for users, so a fast and effective response method is needed.

[0080] The specific processing by the specific processing unit 290 of the data processing device 12 in the first embodiment is realized by the following means.

[0081] In this invention, the server includes: means for capturing a user's telephone conversation as voice data in real time; means for converting and compressing the captured voice data into an optimal format and transmitting it to the server; means for receiving the voice data and converting it into text data using voice recognition technology; means for analyzing the text data and extracting contextual information, entities, emotions, and intent; means for assessing the possibility of fraud based on the text data; means for notifying the user if a possibility of fraud is determined; means for providing information to the police if a possibility of fraud is determined; and an artificial intelligence agent for continuing the conversation on behalf of the user. This makes it possible to detect telephone fraud in real time, quickly warn the user, and contact the police if necessary, thereby preventing damage before it occurs. Furthermore, by having the artificial intelligence agent continue the conversation with the fraudster, it is possible to reduce the user's mental burden and ensure time for the police to respond.

[0082] "User" refers to a person who uses the system.

[0083] A "telephone conversation" refers to a voice communication that a user has over the telephone.

[0084] "Voice data" refers to the digital recording of a user's telephone conversation.

[0085] "Format" refers to the data structure and file format used to save audio data.

[0086] "Compression" refers to a process for reducing the data size of audio data.

[0087] "Server" refers to a computer system for processing and analyzing audio data.

[0088] "Speech recognition technology" refers to technology for analyzing voice data and converting its contents into text data.

[0089] "Text data" refers to character string data converted from audio data.

[0090] "Contextual information" refers to information about the background and situation of the conversation content.

[0091] "Entity" refers to proper nouns or specific information (e.g., people's names, place names, organization names, etc.) that appear in the conversation content.

[0092] "Emotion" refers to the speaker's emotional state (e.g., joy, anger, sadness, etc.) extracted from the conversation content.

[0093] "Intention" refers to the purpose or aim that a speaker is trying to convey through a conversation.

[0094] "Fraud Potential" refers to the result of assessing whether the content of a conversation poses a risk of fraudulent activity.

[0095] "Notification" refers to a message or signal that displays a warning or information to a user.

[0096] "Police" refers to a public security agency or its affiliates.

[0097] "Means of providing information" refers to the methods and functions for transmitting relevant information from the server to the police.

[0098] An "artificial intelligence agent" refers to an intelligent program that automatically converses on behalf of a user.

[0099] The present invention is a real-time monitoring system for protecting users from telephone fraud. This system monitors users' telephone conversations in real time and automatically detects and responds to fraud risks. Detailed embodiments for specifically implementing the present invention are described below.

[0100] Voice Input Processing

[0101] When a user makes or receives a call, the device automatically captures the call's audio data in real time, converts the captured audio data into an optimal format (e.g., from WAV to MP3), compresses it, and then sends it to the server.

[0102] Converting audio data to text data

[0103] The server receives the voice data sent from the device and converts it into text data using voice recognition technology (e.g., a high-precision voice recognition engine). Software used here includes the Google Cloud Speech-to-Text API.

[0104] Conversation analysis

[0105] The server analyzes the text data. Natural language processing techniques (e.g., SpaCy) are used to extract contextual information, entities (people's names, places, etc.), sentiment, and intent. The analyzed information is then passed to a fraud detection algorithm for evaluation.

[0106] Fraud Detection Algorithms

[0107] The server evaluates the likelihood of fraud based on the analyzed text data. Fraud detection algorithms look for specific keywords and phrases (e.g., "urgent," "transfer," "urgent") and check whether they match previous fraud cases. Potential fraud is detected and flagged.

[0108] User notification function

[0109] If the server determines that there is a high possibility of fraud, it immediately sends a warning signal to the device, which then displays a warning message on the screen or makes an audio notification based on the received warning signal.

[0110] Police contact function

[0111] If a possible fraud is identified, the server will promptly provide the police with relevant data (e.g., the content of the conversation, the caller's phone number, the date and time of the conversation, etc.) This information will be automatically forwarded to the police's receiving system in an appropriate format.

[0112] AI conversational agent

[0113] If it determines that there is a high possibility of fraud, the server sends an instruction to the device to launch an AI conversation agent. The device then launches the AI ​​conversation agent based on this instruction and continues the conversation on the user's behalf. The AI ​​conversation agent generates natural conversation and continues to converse with the fraudster, buying time and supporting police response.

[0114] Specific examples

[0115] For example, if a user answers a phone call and the caller says, "I'm your son, please send me money now," the device immediately sends the voice data to the server. The server converts the voice data into text data and analyzes it using an NLP engine. If keywords such as "son" and "please send me money" are detected, it is determined to be a fraudulent call. A warning is immediately issued to the user and the police are notified. Furthermore, an AI conversation agent is activated to continue the conversation on the user's behalf, allowing time for the police to respond.

[0116] Prompt Sentence Examples

[0117] By inputting the prompt "Please explain in detail how the system will respond if a user encounters a phone scam," the generative AI model will explain the detailed operation of each processing step. Furthermore, by using the prompt "Please tell me how to convert voice data into text and assess the risk of fraud," a detailed explanation of the speech recognition and fraud detection algorithms will be obtained.

[0118] The present invention provides a system that keeps users safe from telephone fraud and enables fast and effective police response.

[0119] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0120] Program processing flow

[0121] Step 1: Getting voice input

[0122] When the device detects a user's phone call, it captures audio data in real time through the microphone. The input is the user's voice, and the output is digital audio data. The device converts this captured audio data into an optimal format (e.g., WAV format). After conversion, the data is compressed (e.g., converted to MP3 format).

[0123] Step 2: Sending audio data to the server

[0124] The device sends the compressed audio data to the server. The input is the compressed audio data obtained in step 1, and the output is the data to be sent to the server. This data is sent using a secure protocol (e.g., HTTPS).

[0125] Step 3: Converting audio data to text data

[0126] The server receives the voice data sent from the device and converts it into text data using a speech recognition engine (e.g., Google Cloud Speech-to-Text API). The input is compressed voice data, and the output is text data. Specifically, the server analyzes the voice file and outputs the spoken content as a string of characters.

[0127] Step 4: Analyzing the conversation

[0128] The server analyzes the text data. It uses natural language processing techniques (e.g., SpaCy) to extract contextual information, entities, sentiment, and intent. The input is text data, and the output is analyzed data. Specifically, the server tokenizes the text data and analyzes the meaning of each token.

[0129] Step 5: Fraud Potential Assessment

[0130] The server evaluates the likelihood of fraud based on the analysis results. It checks whether certain keywords or phrases (e.g., "urgent," "transfer," "urgent") match a database of past frauds. The input is the analyzed data, and the output is an assessment of the likelihood of fraud. Based on the assessment, if there is a high likelihood of fraud, it is flagged.

[0131] Step 6: User Notification

[0132] If it is determined that there is a high possibility of fraud, the server immediately sends a warning signal to the terminal. The input is the fraud evaluation result, and the output is the warning signal to the terminal. Based on the received warning signal, the terminal displays a warning message on the screen or issues an audio notification.

[0133] Step 7: Contact the police

[0134] If it is determined that there is a high possibility of fraud, the server will provide the police with relevant data (e.g., the content of the conversation, the caller's phone number, the date and time of the conversation, etc.). The input is the fraud assessment result and relevant data, and the output is the data to be sent to the police. The server formats the data in an appropriate format and forwards it to the police's receiving system.

[0135] Step 8: Launching the AI ​​Conversation Agent

[0136] The server sends an instruction to launch the AI ​​conversation agent to the device. The input is the fraud evaluation result, and the output is the startup instruction to the device. The device then launches the AI ​​conversation agent based on the instruction and continues the conversation on behalf of the user. Specifically, a natural conversation is conducted based on the prompt sentences generated by the AI ​​conversation agent.

[0137] Through the above processing steps, the system of the present invention monitors telephone fraud in real time and responds quickly and efficiently.

[0138] (Application example 1)

[0139] Next, a description will be given of Application Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0140] In recent years, telephone fraud has been on the rise, and countermeasures are urgently needed. However, with current systems, it takes time for users to realize they have been scammed, which often results in greater damage. Furthermore, delays in contacting authorities such as the police make it difficult to take prompt action. Furthermore, the pressure of continuing a conversation with the fraudster places a heavy psychological burden on users. Therefore, there is a need for a system that can detect telephone fraud in real time, immediately warn users, and, if necessary, contact authorities such as the police and take action on their behalf.

[0141] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 1 is realized by the following means.

[0142] In this invention, the server includes: means for acquiring a user's telephone conversation as voice data in real time; means for converting the acquired voice data into text data; means for analyzing the text data to determine the possibility of fraud; means for monitoring the voice data in real time and assessing the risk of fraud; means for determining the risk of fraud based on the detection of specific keywords and contextual information and displaying a warning; means for notifying the user when a possibility of fraud is determined; means for providing information to authorities such as the police when a possibility of fraud is determined; and artificial intelligence agent means for continuing the conversation on behalf of the user. This allows the user to know the risk of telephone fraud in real time and respond quickly. Furthermore, since authorities such as the police can be contacted quickly, a quick response can be expected. Furthermore, by having the artificial intelligence agent continue the conversation on behalf of the user, the user's mental burden can be reduced.

[0143] A "user" is someone who uses the system.

[0144] A "telephone conversation" is a dialogue conducted through voice communication.

[0145] "Voice data" refers to information that has been digitized and recorded from a user's telephone conversation.

[0146] "Text data" is voice data converted into character information.

[0147] "Analyzing text data" means using natural language processing technology to understand the content of the text data and extract specific meanings and information.

[0148] "Determining the likelihood of fraud" means that the system evaluates whether there is a risk of fraud based on the analyzed text data.

[0149] "Notifying the user" means that if the system determines that there is a possibility of fraud, it will provide the user with information to warn or caution them.

[0150] "Providing information to police or other authorities" means transmitting the content of the conversation and related information to police or other law enforcement authorities if it is determined that there is a possibility of fraud.

[0151] An "artificial intelligence agent" is a program that carries on a telephone conversation on behalf of a user and mimics natural dialogue.

[0152] "Monitoring voice data in real time" means instantly monitoring a user's telephone conversation and analyzing the voice data as needed.

[0153] "Assessing the risk of fraud" means determining the possibility of fraud based on the content of the text data.

[0154] "Detecting specific keywords and contextual information" means finding important words and context related to fraudulent activity from the analyzed text data.

[0155] "Display a warning" means that the system will display a warning message on the user's terminal if a fraud risk is detected.

[0156] The present invention relates to a system for monitoring a user's telephone conversations in real time and assessing the risk of fraud. Detailed embodiments for specifically implementing this system are described below.

[0157] System Configuration

[0158] The system mainly consists of the following components:

[0159] 1. User's device

[0160] 2. Server

[0161] 3. Artificial Intelligence Agents

[0162] User's device

[0163] The user's device is a communication device such as a smartphone, and captures the user's telephone conversation as voice data in real time. The device converts the captured voice data into an optimal format, compresses it, and sends it to the server. As a specific example, the device uses the "Google Cloud Speech-to-Text API" as voice recognition technology to convert the voice data into text data.

[0164] server

[0165] The server receives the voice data sent from the device and analyzes it in real time. The voice data is converted into text data using speech recognition technology, and then analyzed using a natural language processing (NLP) engine. The server performs the following processes based on the analyzed text data:

[0166] Detect specific keywords and contextual information to assess fraud risk.

[0167] If the risk is deemed high, a warning will be sent to the user.

[0168] Providing information to police and other authorities as necessary.

[0169] The server detects specific keywords (e.g., "urgent," "transfer," "urgent") and contextual information based on an algorithm to assess the risk of fraud. This determines whether there is a risk of fraud. Natural language processing technologies used in this process include Google Cloud Natural Language API.

[0170] Artificial Intelligence Agent

[0171] An AI agent is activated when a high risk of fraud is detected and continues the conversation on behalf of the user. The AI ​​agent mimics natural dialogue and maintains a conversation with the fraudster, freeing up police response time.

[0172] Specific examples

[0173] For example, if a user answers a phone call and the caller says, "I'm your son, please send me money now," the device immediately sends the voice data to the server. The server converts the voice data into text data and analyzes it using an NLP engine. If keywords such as "son" and "please send me money" are detected, it is determined to be a fraudulent call. A warning is immediately issued to the user, and the police are notified. An artificial intelligence agent is then activated to continue the conversation on the user's behalf.

[0174] Prompt Sentence Examples

[0175] Here is an example prompt for creating an AI model:

[0176] "Create an NLP model that converts phone voice data into text in real time and detects fraud risk."

[0177] "Convert the audio data into text data using the Google Cloud Speech-to-Text API."

[0178] As can be seen, the present invention provides a system that keeps users safe from telephone fraud and allows for a fast and effective police response.

[0179] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0180] Step 1:

[0181] The device receives the user's phone conversation in real time, capturing voice data through a microphone and converting it into a digital format.

[0182] Input: User's phone conversation (voice data)

[0183] Data processing: capturing audio data and converting it to a digital format

[0184] Output: Digital audio data

[0185] Step 2:

[0186] The device converts the acquired audio data into the optimal format and compresses it.

[0187] Input: Digital audio data

[0188] Data processing: format conversion and compression

[0189] Output: Compressed audio data

[0190] Step 3:

[0191] The terminal transmits the compressed audio data to the server.

[0192] Input: Compressed audio data

[0193] Data calculation: Data transmission

[0194] Output: Data transfer to the server

[0195] Step 4:

[0196] The server receives the voice data sent from the device and converts it into text data using voice recognition technology, using the Google Cloud Speech-to-Text API.

[0197] Input: Compressed audio data

[0198] Data processing: speech recognition and text conversion

[0199] Output: Text data

[0200] Step 5:

[0201] The server analyzes the text data using a natural language processing (NLP) engine, using the Google Cloud Natural Language API to extract specific keywords and contextual information from the text data.

[0202] Input: Text data

[0203] Data processing: Information extraction through natural language analysis

[0204] Output: Analysis results (specific keywords and contextual information)

[0205] Step 6:

[0206] The server evaluates the fraud risk based on the analysis results, checking whether specific keywords (e.g., "emergency," "transfer," "urgent") are included to determine the fraud risk.

[0207] Input: Analysis results

[0208] Data Computing: Applying Fraud Risk Assessment Algorithms

[0209] Output: Fraud risk rating (high, medium, low)

[0210] Step 7:

[0211] If the server determines that the risk of fraud is high, it will send a warning to the user's device.

[0212] Input: Fraud Risk Rating (High)

[0213] Data Calculation: Generating and Sending Warning Messages

[0214] Output: Warning notification to the user terminal

[0215] Step 8:

[0216] The user's device will display the received warning on the screen and also alert them with audio notifications and vibrations.

[0217] Input: Warning notification from the server

[0218] Data processing: Display warning message

[0219] Output: User attention (screen display and audio notification)

[0220] Step 9:

[0221] If the server determines there is a high risk of fraud, it will provide relevant information to authorities such as the police, sending data including the content of the conversation and the caller's number.

[0222] Input: Fraud Risk Rating (High)

[0223] Data Computing: Generating and transmitting relevant information

[0224] Output: Providing information to authorities such as police

[0225] Step 10:

[0226] If the server determines that the fraud risk is high, it sends an instruction to the terminal to activate an AI agent, which continues the conversation on behalf of the user.

[0227] Input: Fraud Risk Rating (High)

[0228] Data Calculation: Instructions for launching artificial intelligence agents

[0229] Output: Launch of artificial intelligence agent

[0230] Step 11:

[0231] The artificial intelligence agent mimics natural dialogue and continues the conversation on behalf of the user, maintaining a conversation with the fraudster to free up police response time.

[0232] Input: Conversational content and user instructions

[0233] Data processing: Natural conversation generation

[0234] Output: Ongoing dialogue with the fraudster

[0235] Furthermore, an emotion engine that estimates the user's emotion may be combined. That is, the identification processing unit 290 may estimate the user's emotion using the emotion identification model 59 and perform identification processing using the user's emotion.

[0236] The present invention is a system that monitors users' telephone conversations in real time and automatically detects fraud risks. Furthermore, by combining it with an emotion engine that recognizes the user's emotional state, it provides more advanced fraud detection and user protection functions. Detailed embodiments for specifically implementing this system are described below.

[0237] Voice Input Processing

[0238] When a user makes or receives a call, the terminal automatically captures the call's audio data in real time, which is then converted and compressed into an optimal format by the terminal and then sent to the server.

[0239] Converting audio data to text data

[0240] The server receives the voice data sent from the device and converts it into text data using voice recognition technology, which includes a highly accurate natural language processing (NLP) engine.

[0241] Conversation analysis

[0242] The server analyzes the text data and extracts contextual information, entities (such as names of people and places), sentiment, and intent from the conversation. This analysis process is performed using natural language processing techniques. The analyzed information is then passed to a fraud detection algorithm.

[0243] Emotion recognition by emotion engine

[0244] The server uses an emotion engine to further analyze the text data, which recognizes the user's emotional state (e.g., joy, sadness, anger, fear, anxiety, etc.) This data is provided as additional input to the fraud detection algorithm.

[0245] Fraud Detection Algorithms

[0246] The server evaluates the likelihood of fraud based on the analyzed text data and sentiment information. Fraud detection algorithms detect specific keywords and phrases (e.g., "emergency," "transfer," "urgent") and check whether they match a database of past frauds. If potential fraud is detected, it is flagged by the server.

[0247] User notification function

[0248] If the server determines that there is a high possibility of fraud, it immediately sends a warning signal to the device. The device then displays a warning message on the screen or makes an audio notification to the user. The strength and type of the warning are adjusted according to the emotional state detected by the emotion engine.

[0249] Police contact function

[0250] If the server determines that there is a possibility of fraud, it will collect relevant data (e.g., the content of the conversation, the caller's phone number, the date and time of the conversation, etc.) and provide it to the police promptly. This information will be automatically forwarded in an appropriate format to the police's receiving system.

[0251] AI conversational agent

[0252] If it determines that there is a high possibility of fraud, the server sends an instruction to the device to activate an AI conversation agent. The device then activates the AI ​​conversation agent based on this instruction and continues the conversation on the user's behalf. The AI ​​conversation agent generates natural conversation and continues to converse with the fraudster, buying time and supporting the police response.

[0253] Specific examples

[0254] For example, if a user answers a phone call and the caller says, "I'm your son, please send me money now," the device immediately sends the voice data to the server. The server converts the voice data into text data and analyzes it using an NLP engine to detect keywords such as "son" and "please send me money." The emotion engine also detects the user's feelings of fear or anxiety. If it determines that there is a high possibility of fraud, it immediately issues a warning to the user and notifies the police. Furthermore, an AI conversation agent is activated to continue the conversation on the user's behalf, allowing time for the police to respond.

[0255] As described above, the present invention is a system that protects users from telephone fraud in real time and further achieves more accurate fraud detection and warning by taking into account the user's emotional state.

[0256] The processing flow will be explained below.

[0257] Step 1:

[0258] A user makes or receives a call. The terminal captures the voice data of this call in real time. The captured voice data is compressed within the terminal and prepared for transmission to the server.

[0259] Step 2:

[0260] The device transmits the captured audio data to the server, and this communication uses a secure protocol to ensure data confidentiality.

[0261] Step 3:

[0262] The server processes the received voice data and converts it into text using speech recognition technology, which includes a highly accurate natural language processing (NLP) engine.

[0263] Step 4:

[0264] The server analyzes the text data, using an NLP engine to extract contextual information, entities (such as names of people or places), sentiment, and intent from the conversation, and passes the results of this analysis to a fraud detection algorithm.

[0265] Step 5:

[0266] The server then passes the text data to an emotion engine to evaluate the user's emotional state, which reads emotions from the user's tone of voice and word choice and adds them to the analysis results.

[0267] Step 6:

[0268] The server runs a fraud detection algorithm to assess the likelihood of fraud based on the analyzed text data and sentiment information. If certain keywords or phrases (e.g., "urgent," "transfer," "urgent") are detected, they are compared with a database of past frauds to determine the degree of match.

[0269] Step 7:

[0270] If the server determines that there is a high possibility of fraud, it will raise a warning flag and record the possibility of fraud. This flag will be immediately notified to the user.

[0271] Step 8:

[0272] When a warning flag is raised, the server immediately sends a warning signal to the device. The device receives this signal and displays a warning message on the screen or issues an audio notification to the user. The strength and type of the warning are adjusted according to the emotional state detected by the emotion engine.

[0273] Step 9:

[0274] If the server determines that there is a high possibility of fraud, it will collect detailed relevant information (e.g., the content of the conversation, the caller's phone number, the date and time of the conversation, etc.) and provide it to the police promptly. This information will be automatically forwarded in an appropriate format to the police's receiving system.

[0275] Step 10:

[0276] If the server determines that there is a high possibility of fraud, it sends an instruction to the device to activate an AI conversation agent. Based on this instruction, the device activates the AI ​​conversation agent and continues the conversation on behalf of the user.

[0277] Step 11:

[0278] The AI ​​conversational agent generates natural conversations on behalf of the user and continues the dialogue with the fraudster. For example, the AI ​​agent can ask, "Okay, where should I send the money?" to elicit additional information from the fraudster, buying time and assisting police in their response.

[0279] Step 12:

[0280] The server records all conversation data and stores it in a format that can be analyzed later. Newly detected fraud patterns and techniques are added to the database and used as training data to improve future detection accuracy.

[0281] Through these steps, the system can protect users from phone fraud in real time and assist police in responding quickly. Furthermore, by taking into account the user's emotional state, the accuracy of fraud detection and warning can be further improved.

[0282] Example 2

[0283] Next, a description will be given of Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0284] In recent years, telephone fraud has been on the rise, with many people falling victim to it. Furthermore, fraudulent activities are becoming more complex and sophisticated, making it difficult to detect and respond to them in real time using conventional methods. Furthermore, the victim's emotional state can be an important factor in conversations with fraudsters, but conventional systems do not take this into account. This creates a need for rapid and accurate fraud detection and response.

[0285] The identification process by the identification processing unit 290 of the data processing device 12 in Example 2 is realized by the following means. In this invention, the server includes means for acquiring a user's telephone conversation as voice data in real time, means for converting the acquired voice data into text data, means for analyzing the text data to determine the possibility of fraud, means for notifying the user when the possibility of fraud is determined, means for providing information to relevant authorities when the possibility of fraud is determined, an artificial intelligence agent means for continuing the conversation on behalf of the user, means for analyzing the user's emotional state, and means for reevaluating the possibility of fraud based on the emotional state. This makes it possible to quickly and accurately detect fraudulent acts and take appropriate measures taking into account the user's emotional state.

[0286] "User" means any person or entity that uses the System to conduct telephone conversations.

[0287] A "telephone conversation" refers to a voice-based communication that a user has with another person via telephone.

[0288] "Voice data" refers to voice information captured in real time as a digital representation of a user's telephone conversation.

[0289] "Text data" refers to text information converted from voice data using voice recognition technology.

[0290] "Analysis" refers to the process of extracting the content and meaning of text or audio data and converting it into an understandable format.

[0291] "Possibility of fraud" refers to the result of assessing the degree of likelihood of fraud based on the acquired conversation content.

[0292] "Notification" refers to the act of the system displaying or audibly conveying warnings or information to the user.

[0293] "Relevant authorities" refers to organisations and bodies that are provided with information to deal with fraudulent activity, including the police.

[0294] An "artificial intelligence agent" refers to a sophisticated automated response program that can continue a dialogue on behalf of a user.

[0295] "Emotional state" refers to a user's mental or emotional response or state (e.g., joy, sadness, anger, fear, anxiety, etc.).

[0296] "Reappraisal" refers to the process of reviewing and updating the initial appraisal results based on the acquired emotional state.

[0297] The present invention is a system that monitors users' telephone conversations in real time and automatically detects fraud risks. Furthermore, by combining it with an emotion engine that recognizes the user's emotional state, it provides more advanced fraud detection and user protection functions.

[0298] Voice Input Processing

[0299] The device automatically captures the call audio data when the user makes or receives a call. The audio data is collected in real time and temporarily stored in the internal memory. This process can be performed using a smartphone or a dedicated device.

[0300] Audio data transmission and format conversion

[0301] The device converts the captured audio data into the optimal format (e.g., WAV, MP3) and compresses it using commonly used data compression algorithms (e.g., G.711, Opus). After conversion, the audio data is sent from the device to the server.

[0302] Converting audio data to text

[0303] The server receives the voice data sent from the device and converts it into text data using speech recognition technologies such as Google Cloud Speech-to-Text and Amazon Transcribe. This process utilizes a highly accurate natural language processing (NLP) engine.

[0304] Text data analysis

[0305] The server then analyzes the converted text data using natural language processing tools such as SpaCy and BERT. During the analysis step, contextual information, entities (e.g., people's names, place names), sentiment, and intent are extracted.

[0306] Performing emotion recognition

[0307] The server uses an emotion recognition engine such as IBM Watson Tone Analyzer to analyze the user's emotional state (happiness, sadness, anger, fear, anxiety, etc.) from the text data. This emotional information is used as additional input for subsequent fraud risk assessment.

[0308] Fraud risk assessment

[0309] The server assesses the fraud risk based on the analyzed text data and sentiment information. The fraud detection algorithm looks for specific keywords and phrases (e.g., "emergency," "transfer," "urgent") and matches them with patterns in a fraud database. If a high fraud risk is detected, the server flags the system.

[0310] User warning notification

[0311] If the server determines that there is a high risk of fraud, it immediately sends a warning signal to the terminal. The terminal then receives this warning signal and displays a warning message on the screen and also makes an audio notification to the user. For example, a message such as "There is a possibility of fraud, please be careful" may be displayed.

[0312] Data sharing with the police

[0313] If the server determines that there is a possibility of fraud, it collects relevant data such as the content of the conversation, the originating phone number, the date and time of the conversation, etc. This data is organized in an appropriate format (e.g., CSV file, JSON data) and automatically forwarded to the receiving system of the relevant agency, such as the police.

[0314] Launching an AI conversation agent

[0315] If the risk of fraud is deemed high, the server sends instructions to the device to activate an AI conversation agent. The device then receives this instruction, activates the AI ​​conversation agent, and generates natural conversation on behalf of the user. The AI ​​conversation agent uses the generative AI model to continue dialogue with the fraudster, buying time and assisting the police in their response.

[0316] Specific examples

[0317] For example, if a user answers a phone call and says, "I'm your son, please send me money now," the device immediately sends the voice data to the server. The server then converts the voice data into text using Google Cloud Speech-to-Text and uses SpaCy to extract keywords such as "son" and "please send me money." The emotion engine then analyzes the user's fear and anxiety to detect a high risk of fraud. The server then immediately sends a warning signal to the device, displays a warning message to the user, and notifies the police. Furthermore, an AI conversation agent continues the conversation on behalf of the user, allowing time for the police to respond.

[0318] Prompt Sentence Examples

[0319] "If a user is having a potentially fraudulent conversation on the phone, capture the audio data in real time and assess the likelihood of it being fraudulent. If it is, create a system that alerts the user and activates an AI conversation agent to continue the conversation on the user's behalf."

[0320] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0321] Step 1:

[0322] When a user makes or receives a call, the device automatically captures the call's audio data. Specifically, the device uses the microphone on the smartphone or dedicated device to collect audio data in real time and temporarily stores it in its internal memory. The input is the user's voice during the call, and the output is digital audio data.

[0323] Step 2:

[0324] The device converts the captured audio data into an optimal format (e.g., WAV, MP3) and compresses it. This is done using audio file format conversion software and a data compression algorithm (e.g., G.711, Opus). The input is the audio data obtained in step 1, and the output is the compressed audio data.

[0325] Step 3:

[0326] The device sends the compressed audio data to the server using a data transfer protocol over the Internet (e.g., HTTPS). The input is the compressed audio data, and the output is the audio data transferred to the server.

[0327] Step 4:

[0328] The server receives the voice data sent from the device and converts it into text data using speech recognition technologies such as Google Cloud Speech-to-Text and Amazon Transcribe. The input is the voice data transferred to the server, and the output is text data.

[0329] Step 5:

[0330] The server analyzes the converted text data using natural language processing tools (e.g., SpaCy, BERT). The analysis step extracts contextual information, entities (e.g., people's names, place names), sentiment, and intent. The input is text data, and the output is analyzed information (contextual information, entities, sentiment, and intent).

[0331] Step 6:

[0332] The server inputs the analyzed text data into an emotion recognition engine (e.g., IBM Watson Tone Analyzer) to analyze the user's emotional state. The emotion recognition engine extracts emotions such as joy, sadness, anger, fear, and anxiety from the text data. The input is the analyzed text data, and the output is emotional data.

[0333] Step 7:

[0334] The server assesses the fraud risk based on the analyzed text data and sentiment information. The fraud detection algorithm detects specific keywords and phrases (e.g., "emergency," "transfer," "urgent") and checks whether they match patterns in a fraud database. The input is text data and sentiment data, and the output is a fraud risk assessment result.

[0335] Step 8:

[0336] If the server evaluates the fraud risk as high, it immediately sends a warning signal to the terminal. The terminal receives this warning signal and displays a warning message on the screen for the user, and if necessary, also makes an audio notification. The input is the fraud risk evaluation result, and the output is a warning message for the user.

[0337] Step 9:

[0338] If the server determines that there is a high risk of fraud, it collects relevant data (e.g., conversation content, caller phone number, conversation date and time, etc.) and automatically forwards it to the police receiving system. This information is organized in an appropriate format (e.g., CSV file, JSON data). The input is the conversation content and related data, and the output is the data to be sent to the police.

[0339] Step 10:

[0340] If the risk of fraud is determined to be high, the server sends instructions to the device to activate an AI conversation agent. The device then receives this instruction, activates the AI ​​conversation agent, and generates a natural conversation on behalf of the user. The AI ​​conversation agent uses a generative AI model to continue the dialogue with the fraudster, buying time and assisting the police in their response. The input is instructions from the server, and the output is the dialogue with the fraudster.

[0341] (Application example 2)

[0342] Next, a description will be given of Application Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0343] In modern society, telephone fraud is on the rise, posing a serious problem, especially for elderly users and those in situations where they are prone to anxiety. Conventional fraud prevention measures lack real-time detection of telephone fraud and warning functions that take into account the user's emotional state, making it difficult to prevent damage before it occurs. It is necessary for fraud detection systems to take the user's emotional state into account to achieve more accurate fraud detection and prompt warnings.

[0344] The identification processing by the identification processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means. In this invention, the server includes means for acquiring a user's telephone conversation as voice data in real time and converting the acquired voice data into text data, means for analyzing the text data to determine the possibility of fraud, means for notifying the user when the possibility of fraud is determined, means for recognizing the user's emotional state using an emotion recognition engine, and means for additionally providing the recognized emotional state to the fraud detection algorithm. This enables real-time fraud detection and warnings and responses based on the user's emotional state.

[0345] "User telephone conversation" refers to a user's verbal communication conducted over the telephone.

[0346] "Real-time" refers to processing or reaction occurring immediately, without delay.

[0347] "Audio data" refers to information that represents audio in digital form.

[0348] "Text data" refers to information obtained by converting voice data into a character string format.

[0349] "Means for converting" refers to a method or device for converting data of a particular format into another format.

[0350] "Means for analyzing to determine likelihood of fraud" refers to a method or device that analyzes input data to assess the risk of fraud.

[0351] "Means for notifying" refers to a method or device for notifying a user of specific information.

[0352] "Means of providing information to law enforcement" refers to a method or device that transmits specific information in an appropriate format to law enforcement agencies.

[0353] "Artificial intelligence agent means" refers to a method or device that uses artificial intelligence to act or interact on behalf of a user.

[0354] An "emotion recognition engine" refers to software or algorithms used to analyze and identify a user's emotional state.

[0355] "Fraud detection algorithm" refers to a computational procedure or program used to assess the risk of fraud based on specific patterns or keywords.

[0356] The present invention provides a system for monitoring users' telephone conversations in real time and automatically detecting fraud risks. Furthermore, by combining an emotion engine that recognizes the user's emotional state, it provides more advanced fraud detection and user protection capabilities. An embodiment of this system is described in detail below.

[0357] Voice Input Processing

[0358] When a user makes or receives a call, the device automatically captures the call's audio data in real time. The hardware used is a smartphone or tablet, and the software is a voice recognition library (e.g., speech_recognition). The captured audio data is converted and compressed into an optimal format by the device, and then sent to the server.

[0359] Converting audio data to text data

[0360] The server receives the voice data sent from the device and converts it into text using speech recognition technology, using a highly accurate natural language processing (NLP) engine (e.g., Google's speech recognition API).

[0361] Conversation analysis

[0362] The server analyzes the text data to extract contextual information, entities (such as names of people and places), sentiment, and intent from the conversation. This analysis process uses natural language processing techniques (e.g., NLP engines). The analyzed information is then passed to a fraud detection algorithm.

[0363] Emotion recognition by emotion engine

[0364] The server uses an emotion engine to further analyze the text data. The emotion engine recognizes the user's emotional state (e.g., joy, sadness, anger, fear, anxiety, etc.) This data is provided as additional input to the fraud detection algorithm.

[0365] Fraud Detection Algorithms

[0366] The server evaluates the likelihood of fraud based on the analyzed text data and sentiment information. Fraud detection algorithms detect specific keywords and phrases (e.g., "emergency," "transfer," "urgent") and check whether they match a database of past frauds. If potential fraud is detected, it is flagged by the server.

[0367] User notification function

[0368] If the server determines that there is a high possibility of fraud, it immediately sends a warning signal to the device. The device then displays a warning message on the screen or makes an audio notification to the user. The strength and type of the warning are adjusted according to the emotional state detected by the emotion engine.

[0369] Police contact function

[0370] If fraud is suspected, the server will collect relevant data (e.g., the content of the conversation, the caller's phone number, the date and time of the conversation, etc.) and provide it to the police promptly. This information will be automatically forwarded in an appropriate format to the police's receiving system.

[0371] AI conversational agent

[0372] If it determines that there is a high possibility of fraud, the server sends an instruction to the device to activate an AI conversation agent. The device then activates the AI ​​conversation agent based on this instruction and continues the conversation on the user's behalf. The AI ​​conversation agent generates natural conversation and continues to converse with the fraudster, buying time and supporting the police response.

[0373] Specific examples

[0374] For example, if a user answers a phone call and the caller says, "I'm your family, please send me money now," the device immediately sends the voice data to the server. The server converts the voice data into text data and analyzes it using an NLP engine to detect keywords such as "family" and "please send me money." The emotion engine also detects the user's feelings of fear or anxiety. If it determines that there is a high possibility of fraud, it immediately issues a warning to the user and notifies the police. Furthermore, an AI conversation agent is activated to continue the conversation on the user's behalf, allowing time for the police to respond.

[0375] Prompt Sentence Examples

[0376] Please provide sample code and an explanation for building a system that, when a user receives a phone call and the caller says, "I'm your family member, please send me money now," converts the voice data into text data, analyzes it using natural language processing, and determines whether it is likely to be a fraud.

[0377] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[0378] (System program processing flow)

[0379] Step 1: Audio Input Processing

[0380] The device captures the user's telephone conversation as voice data in real time. Specifically, it uses the smartphone's microphone to record the conversation during the call and temporarily stores the data. The input is the voice data captured in real time, and the output is the stored voice data.

[0381] Step 2: Converting audio data to text data

[0382] The device sends the captured voice data to the server, which uses voice recognition technology to convert the voice data into text data. Here, Google's voice recognition API is used to perform highly accurate text conversion. The input is voice data, and the output is text data.

[0383] Step 3: Analyzing the conversation

[0384] The server analyzes the text data to extract the conversation content, context information, entities, sentiment, and intent. This analysis uses natural language processing technology. The input is the text data, and the output is the analyzed context information and entity information.

[0385] Step 4: Emotion Recognition with the Emotion Engine

[0386] The server uses an emotion engine to recognize the user's emotional state from the text data. The emotion engine identifies emotions such as joy, sadness, anger, fear, and anxiety from the content and tone of the user's speech. The input is text data, and the output is emotional state data.

[0387] Step 5: Fraud Detection Algorithm

[0388] The server assesses the likelihood of fraud based on the analyzed text data and sentiment information. The fraud detection algorithm detects specific keywords and phrases and checks whether they match a historical fraud database. The input is contextual information, entity information, and sentiment data, and the output is a fraud risk assessment result.

[0389] Step 6: User Notification Function

[0390] If the server determines that there is a high possibility of fraud, it sends a warning signal to the terminal. The terminal receives this signal and displays a warning message on the screen or notifies the user by voice. The intensity and type of warning are adjusted according to the emotional state detected by the emotion engine. The input is the fraud evaluation result, and the output is a warning notification to the user.

[0391] Step 7: Contact the police

[0392] If the server determines that there is a possibility of fraud, it will collect relevant data (such as the content of the conversation, the caller's phone number, and the date and time of the conversation) and provide it promptly to the police. The information is automatically forwarded to the police's receiving system in an appropriate format. The input is the relevant data, and the output is a notification to the police.

[0393] Step 8: AI Conversational Agent

[0394] If the server determines that there is a high possibility of fraud, it sends an instruction to the terminal to activate an AI conversation agent. The terminal then activates the AI ​​conversation agent based on this instruction and continues the conversation on behalf of the user. The AI ​​conversation agent generates natural conversation and continues to converse with the fraudster, buying time and supporting police response. The input is the fraud assessment result, and the output is the activation of the AI ​​conversation agent.

[0395] The specific processing unit 290 transmits the result of the specific processing to the smart device 14. In the smart device 14, the control unit 46A causes the output device 40 to output the result of the specific processing. The microphone 38B acquires audio indicating a user input regarding the result of the specific processing. The control unit 46A transmits audio data indicating the user input acquired by the microphone 38B to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the audio data.

[0396] The data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of the data generation model 58 is ChatGPT (Internet Search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search <url: https: gemini.google.com ?hl="ja">) and other generation AIs. The data generation model 58 is obtained by performing deep learning on a neural network. A prompt including an instruction is input to the data generation model 58, and inference data such as voice data indicating voice, text data indicating text, and image data indicating an image is also input. The data generation model 58 performs inference on the input inference data in accordance with the instruction indicated by the prompt, and outputs the inference result in a data format such as voice data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[0397] In the above embodiment, an example in which the specific process is performed by the data processing device 12 has been given, but the technology of the present disclosure is not limited to this, and the specific process may be performed by the smart device 14.

[0398] [Second embodiment]

[0399] FIG. 3 shows an example of the configuration of a data processing system 210 according to the second embodiment.

[0400] 3, the data processing system 210 includes the data processing device 12 and smart glasses 214. An example of the data processing device 12 is a server.

[0401] The data processing device 12 includes a computer 22, a database 24, and a communication I / F 26. The computer 22 is an example of a "computer" according to the technology of the present disclosure. The computer 22 includes a processor 28, a RAM 30, and a storage 32. The processor 28, the RAM 30, and the storage 32 are connected to a bus 34. The database 24 and the communication I / F 26 are also connected to the bus 34. The communication I / F 26 is connected to a network 54. Examples of the network 54 include a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0402] The smart glasses 214 include a computer 36, a microphone 238, a speaker 240, a camera 42, and a communication I / F 44. The computer 36 includes a processor 46, a RAM 48, and a storage 50. The processor 46, the RAM 48, and the storage 50 are connected to a bus 52. The microphone 238, the speaker 240, and the camera 42 are also connected to the bus 52.

[0403] The microphone 238 receives instructions and the like from the user 20 by receiving voice uttered by the user 20. The microphone 238 captures the voice uttered by the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio in accordance with instructions from the processor 46.

[0404] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an imaging element such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the surroundings of user 20 (for example, an imaging range defined by an angle of view equivalent to the field of vision of a typical healthy person).

[0405] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 are responsible for the exchange of various information between the processor 46 and the processor 28 via the network 54. The exchange of various information between the processor 46 and the processor 28 using the communication I / Fs 44 and 26 is carried out in a secure state.

[0406] Fig. 4 shows an example of the main functions of the data processing device 12 and the smart glasses 214. As shown in Fig. 4, in the data processing device 12, a specific process is performed by the processor 28. A specific process program 56 is stored in the storage 32.

[0407] The specific processing program 56 is an example of a "program" according to the technology of the present disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific processing is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific processing program 56 executed on the RAM 30.

[0408] The storage 32 stores a data generation model 58 and an emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[0409] In the smart glasses 214, the reception output process is performed by the processor 46. A reception output program 60 is stored in the storage 50. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output process is realized by the processor 46 operating as the control unit 46A in accordance with the reception output program 60 executed on the RAM 48.

[0410] Next, a description will be given of the identification process performed by the identification processing unit 290 of the data processing device 12. In the following description, the data processing device 12 will be referred to as the "server" and the smart glasses 214 will be referred to as the "terminal."

[0411] The present invention is a system for monitoring users' telephone conversations in real time and automatically detecting fraud risks. Detailed embodiments for specifically implementing the system are described below.

[0412] Voice Input Processing

[0413] When a user makes or receives a call, the terminal automatically captures the call's audio data in real time, which is then converted and compressed into an optimal format by the terminal and then sent to the server.

[0414] Converting audio data to text data

[0415] The server receives the voice data sent from the device and converts it into text data using speech recognition technology, which includes a highly accurate natural language processing (NLP) engine.

[0416] Conversation analysis

[0417] The server analyzes the text data and extracts contextual information, entities (such as names of people and places), sentiment, and intent from the conversation. This analysis process is performed using natural language processing techniques. The analyzed information is then passed to a fraud detection algorithm.

[0418] Fraud Detection Algorithms

[0419] The server evaluates the likelihood of fraud based on the analyzed text data. Fraud detection algorithms look for specific keywords and phrases (e.g., "urgent," "transfer," "urgent") and check whether they match previous fraud cases. If potential fraud is detected, it is flagged by the server.

[0420] User notification function

[0421] If the server determines that there is a high possibility of fraud, it immediately sends a warning signal to the device, which then displays a warning message on the screen or makes an audio notification based on the received warning signal.

[0422] Police contact function

[0423] If the server determines that there is a possibility of fraud, it will promptly provide the police with relevant data (e.g., the content of the conversation, the caller's phone number, the date and time of the conversation, etc.) This information will be automatically forwarded to the police's receiving system in an appropriate format.

[0424] AI conversational agent

[0425] If it determines that there is a high possibility of fraud, the server sends an instruction to the device to activate an AI conversation agent. The device then activates the AI ​​conversation agent based on the received instruction and continues the conversation on behalf of the user. The AI ​​conversation agent generates natural conversation and continues to converse with the fraudster, buying time and supporting police response.

[0426] Specific examples

[0427] For example, if a user answers a phone call and the caller says, "I'm your son, please send me money now," the device immediately sends the voice data to the server. The server converts the voice data into text data and analyzes it using an NLP engine. If keywords such as "son" and "please send me money" are detected, it is determined to be a fraudulent call. A warning is immediately issued to the user and the police are notified. Furthermore, an AI conversation agent is activated to continue the conversation on the user's behalf, allowing time for the police to respond.

[0428] Thus, the present invention provides a system that keeps users safe from telephone fraud and allows for fast and effective police response.

[0429] The processing flow will be explained below.

[0430] Step 1:

[0431] A user makes or receives a call. The terminal captures the voice data of this phone conversation in real time. The voice data is compressed within the terminal and prepared to be sent to the server for further processing.

[0432] Step 2:

[0433] The device transmits the captured audio data to the server, and this communication uses a secure protocol to ensure data confidentiality.

[0434] Step 3:

[0435] The server receives the received voice data and converts it into text using speech recognition technology. The speech recognition model utilizes a highly accurate natural language processing (NLP) engine.

[0436] Step 4:

[0437] The server analyzes the text data and uses an NLP engine to extract contextual information, entities (people's names, places, etc.), sentiment, and intent from the conversation. The results of this analysis are passed to a fraud detection algorithm.

[0438] Step 5:

[0439] The server runs fraud detection algorithms to evaluate potential fraud patterns. If certain keywords or phrases (e.g., "urgent," "transfer," "urgent") are detected, they are compared against a database of past frauds to determine the likelihood of a match.

[0440] Step 6:

[0441] If the server determines that there is a high possibility of fraud, it will raise a warning flag and record the possibility of fraud. This flag will be immediately notified to the user.

[0442] Step 7:

[0443] When a warning flag is raised, the server immediately sends a warning signal to the terminal, which then displays a warning message on the screen or notifies the user by voice.

[0444] Step 8:

[0445] The server collects relevant information about the flagged call (e.g., the caller's phone number, the content of the conversation, the date and time, etc.) and prepares it for automatic provision to the police. This information is then forwarded in an appropriate format to the police's receiving system.

[0446] Step 9:

[0447] If the server determines that there is a high possibility of fraud, it sends an instruction to the device to activate an AI conversation agent. Based on this instruction, the device activates the AI ​​conversation agent and continues the conversation on behalf of the user.

[0448] Step 10:

[0449] The AI ​​conversational agent generates natural conversations on behalf of the user and continues the dialogue with the fraudster, buying time and assisting law enforcement. For example, the AI ​​agent can ask, "Okay, where should I send the money?" to elicit additional information from the fraudster.

[0450] Step 11:

[0451] All conversation data is recorded by the server and stored in a format that can be analyzed later. Newly detected fraud patterns and techniques are added to the database and used as training data to improve future detection accuracy.

[0452] Through these steps, the system can protect users from telephone fraud in real time and assist police in responding quickly.

[0453] Example 1

[0454] Next, a description will be given of Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[0455] Telephone scams are becoming more sophisticated every year, making vulnerable individuals such as the elderly particularly vulnerable to fraud. Current manual countermeasures often delay detection and response to scams, often failing to detect them until actual harm has occurred. Furthermore, continuing to talk to scammers is mentally stressful for users, so a fast and effective response method is needed.

[0456] The specific processing by the specific processing unit 290 of the data processing device 12 in the first embodiment is realized by the following means.

[0457] In this invention, the server includes: means for capturing a user's telephone conversation as voice data in real time; means for converting and compressing the captured voice data into an optimal format and transmitting it to the server; means for receiving the voice data and converting it into text data using voice recognition technology; means for analyzing the text data and extracting contextual information, entities, emotions, and intent; means for assessing the possibility of fraud based on the text data; means for notifying the user if a possibility of fraud is determined; means for providing information to the police if a possibility of fraud is determined; and an artificial intelligence agent for continuing the conversation on behalf of the user. This makes it possible to detect telephone fraud in real time, quickly warn the user, and contact the police if necessary, thereby preventing damage before it occurs. Furthermore, by having the artificial intelligence agent continue the conversation with the fraudster, it is possible to reduce the user's mental burden and ensure time for the police to respond.

[0458] "User" refers to a person who uses the system.

[0459] A "telephone conversation" refers to a voice communication that a user has over the telephone.

[0460] "Voice data" refers to the digital recording of a user's telephone conversation.

[0461] "Format" refers to the data structure and file format used to save audio data.

[0462] "Compression" refers to a process for reducing the data size of audio data.

[0463] "Server" refers to a computer system for processing and analyzing audio data.

[0464] "Speech recognition technology" refers to technology for analyzing voice data and converting its contents into text data.

[0465] "Text data" refers to character string data converted from audio data.

[0466] "Contextual information" refers to information about the background and situation of the conversation content.

[0467] "Entity" refers to proper nouns or specific information (e.g., people's names, place names, organization names, etc.) that appear in the conversation content.

[0468] "Emotion" refers to the speaker's emotional state (e.g., joy, anger, sadness, etc.) extracted from the conversation content.

[0469] "Intention" refers to the purpose or aim that a speaker is trying to convey through a conversation.

[0470] "Fraud Potential" refers to the result of assessing whether the content of a conversation poses a risk of fraudulent activity.

[0471] "Notification" refers to a message or signal that displays a warning or information to a user.

[0472] "Police" refers to a public security agency or its affiliates.

[0473] "Means of providing information" refers to the methods and functions for transmitting relevant information from the server to the police.

[0474] An "artificial intelligence agent" refers to an intelligent program that automatically converses on behalf of a user.

[0475] The present invention is a real-time monitoring system for protecting users from telephone fraud. This system monitors users' telephone conversations in real time and automatically detects and responds to fraud risks. Detailed embodiments for specifically implementing the present invention are described below.

[0476] Voice Input Processing

[0477] When a user makes or receives a call, the device automatically captures the call's audio data in real time, converts the captured audio data into an optimal format (e.g., from WAV to MP3), compresses it, and then sends it to the server.

[0478] Converting audio data to text data

[0479] The server receives the voice data sent from the device and converts it into text data using voice recognition technology (e.g., a high-precision voice recognition engine). Software used here includes the Google Cloud Speech-to-Text API.

[0480] Conversation analysis

[0481] The server analyzes the text data. Natural language processing techniques (e.g., SpaCy) are used to extract contextual information, entities (people's names, places, etc.), sentiment, and intent. The analyzed information is then passed to a fraud detection algorithm for evaluation.

[0482] Fraud Detection Algorithms

[0483] The server evaluates the likelihood of fraud based on the analyzed text data. Fraud detection algorithms look for specific keywords and phrases (e.g., "urgent," "transfer," "urgent") and check whether they match previous fraud cases. Potential fraud is detected and flagged.

[0484] User notification function

[0485] If the server determines that there is a high possibility of fraud, it immediately sends a warning signal to the device, which then displays a warning message on the screen or makes an audio notification based on the received warning signal.

[0486] Police contact function

[0487] If a possible fraud is identified, the server will promptly provide the police with relevant data (e.g., the content of the conversation, the caller's phone number, the date and time of the conversation, etc.) This information will be automatically forwarded to the police's receiving system in an appropriate format.

[0488] AI conversational agent

[0489] If it determines that there is a high possibility of fraud, the server sends an instruction to the device to launch an AI conversation agent. The device then launches the AI ​​conversation agent based on this instruction and continues the conversation on the user's behalf. The AI ​​conversation agent generates natural conversation and continues to converse with the fraudster, buying time and supporting police response.

[0490] Specific examples

[0491] For example, if a user answers a phone call and the caller says, "I'm your son, please send me money now," the device immediately sends the voice data to the server. The server converts the voice data into text data and analyzes it using an NLP engine. If keywords such as "son" and "please send me money" are detected, it is determined to be a fraudulent call. A warning is immediately issued to the user and the police are notified. Furthermore, an AI conversation agent is activated to continue the conversation on the user's behalf, allowing time for the police to respond.

[0492] Prompt Sentence Examples

[0493] By inputting the prompt "Please explain in detail how the system will respond if a user encounters a phone scam," the generative AI model will explain the detailed operation of each processing step. Furthermore, by using the prompt "Please tell me how to convert voice data into text and assess the risk of fraud," a detailed explanation of the speech recognition and fraud detection algorithms will be obtained.

[0494] The present invention provides a system that keeps users safe from telephone fraud and enables fast and effective police response.

[0495] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0496] Program processing flow

[0497] Step 1: Getting voice input

[0498] When the device detects a user's phone call, it captures audio data in real time through the microphone. The input is the user's voice, and the output is digital audio data. The device converts this captured audio data into an optimal format (e.g., WAV format). After conversion, the data is compressed (e.g., converted to MP3 format).

[0499] Step 2: Sending audio data to the server

[0500] The device sends the compressed audio data to the server. The input is the compressed audio data obtained in step 1, and the output is the data to be sent to the server. This data is sent using a secure protocol (e.g., HTTPS).

[0501] Step 3: Converting audio data to text data

[0502] The server receives the voice data sent from the device and converts it into text data using a speech recognition engine (e.g., Google Cloud Speech-to-Text API). The input is compressed voice data, and the output is text data. Specifically, the server analyzes the voice file and outputs the spoken content as a string of characters.

[0503] Step 4: Analyzing the conversation

[0504] The server analyzes the text data. It uses natural language processing techniques (e.g., SpaCy) to extract contextual information, entities, sentiment, and intent. The input is text data, and the output is analyzed data. Specifically, the server tokenizes the text data and analyzes the meaning of each token.

[0505] Step 5: Fraud Potential Assessment

[0506] The server evaluates the likelihood of fraud based on the analysis results. It checks whether certain keywords or phrases (e.g., "urgent," "transfer," "urgent") match a database of past frauds. The input is the analyzed data, and the output is an assessment of the likelihood of fraud. Based on the assessment, if there is a high likelihood of fraud, it is flagged.

[0507] Step 6: User Notification

[0508] If it is determined that there is a high possibility of fraud, the server immediately sends a warning signal to the terminal. The input is the fraud evaluation result, and the output is the warning signal to the terminal. Based on the received warning signal, the terminal displays a warning message on the screen or issues an audio notification.

[0509] Step 7: Contact the police

[0510] If it is determined that there is a high possibility of fraud, the server will provide the police with relevant data (e.g., the content of the conversation, the caller's phone number, the date and time of the conversation, etc.). The input is the fraud assessment result and relevant data, and the output is the data to be sent to the police. The server formats the data in an appropriate format and forwards it to the police's receiving system.

[0511] Step 8: Launching the AI ​​Conversation Agent

[0512] The server sends an instruction to launch the AI ​​conversation agent to the device. The input is the fraud evaluation result, and the output is the startup instruction to the device. The device then launches the AI ​​conversation agent based on the instruction and continues the conversation on behalf of the user. Specifically, a natural conversation is conducted based on the prompt sentences generated by the AI ​​conversation agent.

[0513] Through the above processing steps, the system of the present invention monitors telephone fraud in real time and responds quickly and efficiently.

[0514] (Application example 1)

[0515] Next, a description will be given of Application Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[0516] In recent years, telephone fraud has been on the rise, and countermeasures are urgently needed. However, with current systems, it takes time for users to realize they have been scammed, which often results in greater damage. Furthermore, delays in contacting authorities such as the police make it difficult to take prompt action. Furthermore, the pressure of continuing a conversation with the fraudster places a heavy psychological burden on users. Therefore, there is a need for a system that can detect telephone fraud in real time, immediately warn users, and, if necessary, contact authorities such as the police and take action on their behalf.

[0517] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 1 is realized by the following means.

[0518] In this invention, the server includes: means for acquiring a user's telephone conversation as voice data in real time; means for converting the acquired voice data into text data; means for analyzing the text data to determine the possibility of fraud; means for monitoring the voice data in real time and assessing the risk of fraud; means for determining the risk of fraud based on the detection of specific keywords and contextual information and displaying a warning; means for notifying the user when a possibility of fraud is determined; means for providing information to authorities such as the police when a possibility of fraud is determined; and artificial intelligence agent means for continuing the conversation on behalf of the user. This allows the user to know the risk of telephone fraud in real time and respond quickly. Furthermore, since authorities such as the police can be contacted quickly, a quick response can be expected. Furthermore, by having the artificial intelligence agent continue the conversation on behalf of the user, the user's mental burden can be reduced.

[0519] A "user" is someone who uses the system.

[0520] A "telephone conversation" is a dialogue conducted through voice communication.

[0521] "Voice data" refers to information that has been digitized and recorded from a user's telephone conversation.

[0522] "Text data" is voice data converted into character information.

[0523] "Analyzing text data" means using natural language processing technology to understand the content of the text data and extract specific meanings and information.

[0524] "Determining the likelihood of fraud" means that the system evaluates whether there is a risk of fraud based on the analyzed text data.

[0525] "Notifying the user" means that if the system determines that there is a possibility of fraud, it will provide the user with information to warn or caution them.

[0526] "Providing information to police or other authorities" means transmitting the content of the conversation and related information to police or other law enforcement authorities if it is determined that there is a possibility of fraud.

[0527] An "artificial intelligence agent" is a program that carries on a telephone conversation on behalf of a user and mimics natural dialogue.

[0528] "Monitoring voice data in real time" means instantly monitoring a user's telephone conversation and analyzing the voice data as needed.

[0529] "Assessing the risk of fraud" means determining the possibility of fraud based on the content of the text data.

[0530] "Detecting specific keywords and contextual information" means finding important words and context related to fraudulent activity from the analyzed text data.

[0531] "Display a warning" means that the system will display a warning message on the user's terminal if a fraud risk is detected.

[0532] The present invention relates to a system for monitoring a user's telephone conversations in real time and assessing the risk of fraud. Detailed embodiments for specifically implementing this system are described below.

[0533] System Configuration

[0534] The system mainly consists of the following components:

[0535] 1. User's device

[0536] 2. Server

[0537] 3. Artificial Intelligence Agents

[0538] User's device

[0539] The user's device is a communication device such as a smartphone, and captures the user's telephone conversation as voice data in real time. The device converts the captured voice data into an optimal format, compresses it, and sends it to the server. As a specific example, the device uses the "Google Cloud Speech-to-Text API" as voice recognition technology to convert the voice data into text data.

[0540] server

[0541] The server receives the voice data sent from the device and analyzes it in real time. The voice data is converted into text data using speech recognition technology, and then analyzed using a natural language processing (NLP) engine. The server performs the following processes based on the analyzed text data:

[0542] Detect specific keywords and contextual information to assess fraud risk.

[0543] If the risk is deemed high, a warning will be sent to the user.

[0544] Providing information to police and other authorities as necessary.

[0545] The server detects specific keywords (e.g., "urgent," "transfer," "urgent") and contextual information based on an algorithm to assess the risk of fraud. This determines whether there is a risk of fraud. Natural language processing technologies used in this process include Google Cloud Natural Language API.

[0546] Artificial Intelligence Agent

[0547] An AI agent is activated when a high risk of fraud is detected and continues the conversation on behalf of the user. The AI ​​agent mimics natural dialogue and maintains a conversation with the fraudster, freeing up police response time.

[0548] Specific examples

[0549] For example, if a user answers a phone call and the caller says, "I'm your son, please send me money now," the device immediately sends the voice data to the server. The server converts the voice data into text data and analyzes it using an NLP engine. If keywords such as "son" and "please send me money" are detected, it is determined to be a fraudulent call. A warning is immediately issued to the user, and the police are notified. An artificial intelligence agent is then activated to continue the conversation on the user's behalf.

[0550] Prompt Sentence Examples

[0551] Here is an example prompt for creating an AI model:

[0552] "Create an NLP model that converts phone voice data into text in real time and detects fraud risk."

[0553] "Convert the audio data into text data using the Google Cloud Speech-to-Text API."

[0554] As can be seen, the present invention provides a system that keeps users safe from telephone fraud and allows for a fast and effective police response.

[0555] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0556] Step 1:

[0557] The device receives the user's phone conversation in real time, capturing voice data through a microphone and converting it into a digital format.

[0558] Input: User's phone conversation (voice data)

[0559] Data processing: capturing audio data and converting it to a digital format

[0560] Output: Digital audio data

[0561] Step 2:

[0562] The device converts the acquired audio data into the optimal format and compresses it.

[0563] Input: Digital audio data

[0564] Data processing: format conversion and compression

[0565] Output: Compressed audio data

[0566] Step 3:

[0567] The terminal transmits the compressed audio data to the server.

[0568] Input: Compressed audio data

[0569] Data calculation: Data transmission

[0570] Output: Data transfer to the server

[0571] Step 4:

[0572] The server receives the voice data sent from the device and converts it into text data using voice recognition technology, using the Google Cloud Speech-to-Text API.

[0573] Input: Compressed audio data

[0574] Data processing: speech recognition and text conversion

[0575] Output: Text data

[0576] Step 5:

[0577] The server analyzes the text data using a natural language processing (NLP) engine, using the Google Cloud Natural Language API to extract specific keywords and contextual information from the text data.

[0578] Input: Text data

[0579] Data processing: Information extraction through natural language analysis

[0580] Output: Analysis results (specific keywords and contextual information)

[0581] Step 6:

[0582] The server evaluates the fraud risk based on the analysis results, checking whether specific keywords (e.g., "emergency," "transfer," "urgent") are included to determine the fraud risk.

[0583] Input: Analysis results

[0584] Data Computing: Applying Fraud Risk Assessment Algorithms

[0585] Output: Fraud risk rating (high, medium, low)

[0586] Step 7:

[0587] If the server determines that the risk of fraud is high, it will send a warning to the user's device.

[0588] Input: Fraud Risk Rating (High)

[0589] Data Calculation: Generating and Sending Warning Messages

[0590] Output: Warning notification to the user terminal

[0591] Step 8:

[0592] The user's device will display the received warning on the screen and also alert them with audio notifications and vibrations.

[0593] Input: Warning notification from the server

[0594] Data processing: Display warning message

[0595] Output: User attention (screen display and audio notification)

[0596] Step 9:

[0597] If the server determines there is a high risk of fraud, it will provide relevant information to authorities such as the police, sending data including the content of the conversation and the caller's number.

[0598] Input: Fraud Risk Rating (High)

[0599] Data Computing: Generating and transmitting relevant information

[0600] Output: Providing information to authorities such as police

[0601] Step 10:

[0602] If the server determines that the fraud risk is high, it sends an instruction to the terminal to activate an AI agent, which continues the conversation on behalf of the user.

[0603] Input: Fraud Risk Rating (High)

[0604] Data Calculation: Instructions for launching artificial intelligence agents

[0605] Output: Launch of artificial intelligence agent

[0606] Step 11:

[0607] The artificial intelligence agent mimics natural dialogue and continues the conversation on behalf of the user, maintaining a conversation with the fraudster to free up police response time.

[0608] Input: Conversational content and user instructions

[0609] Data processing: Natural conversation generation

[0610] Output: Ongoing dialogue with the fraudster

[0611] Furthermore, an emotion engine that estimates the user's emotion may be further combined. That is, the identification processing unit 290 may estimate the user's emotion using the emotion identification model 59, and perform identification processing using the user's emotion.

[0612] The present invention is a system that monitors users' telephone conversations in real time and automatically detects fraud risks. Furthermore, by combining it with an emotion engine that recognizes the user's emotional state, it provides more advanced fraud detection and user protection functions. Detailed embodiments for specifically implementing this system are described below.

[0613] Voice Input Processing

[0614] When a user makes or receives a call, the terminal automatically captures the call's audio data in real time, which is then converted and compressed into an optimal format by the terminal and then sent to the server.

[0615] Converting audio data to text data

[0616] The server receives the voice data sent from the device and converts it into text data using voice recognition technology, which includes a highly accurate natural language processing (NLP) engine.

[0617] Conversation analysis

[0618] The server analyzes the text data and extracts contextual information, entities (such as names of people and places), sentiment, and intent from the conversation. This analysis process is performed using natural language processing techniques. The analyzed information is then passed to a fraud detection algorithm.

[0619] Emotion recognition by emotion engine

[0620] The server uses an emotion engine to further analyze the text data, which recognizes the user's emotional state (e.g., joy, sadness, anger, fear, anxiety, etc.) This data is provided as additional input to the fraud detection algorithm.

[0621] Fraud Detection Algorithms

[0622] The server evaluates the likelihood of fraud based on the analyzed text data and sentiment information. Fraud detection algorithms detect specific keywords and phrases (e.g., "emergency," "transfer," "urgent") and check whether they match a database of past frauds. If potential fraud is detected, it is flagged by the server.

[0623] User notification function

[0624] If the server determines that there is a high possibility of fraud, it immediately sends a warning signal to the device. The device then displays a warning message on the screen or makes an audio notification to the user. The strength and type of the warning are adjusted according to the emotional state detected by the emotion engine.

[0625] Police contact function

[0626] If the server determines that there is a possibility of fraud, it will collect relevant data (e.g., the content of the conversation, the caller's phone number, the date and time of the conversation, etc.) and provide it to the police promptly. This information will be automatically forwarded in an appropriate format to the police's receiving system.

[0627] AI conversational agent

[0628] If it determines that there is a high possibility of fraud, the server sends an instruction to the device to activate an AI conversation agent. The device then activates the AI ​​conversation agent based on this instruction and continues the conversation on the user's behalf. The AI ​​conversation agent generates natural conversation and continues to converse with the fraudster, buying time and supporting the police response.

[0629] Specific examples

[0630] For example, if a user answers a phone call and the caller says, "I'm your son, please send me money now," the device immediately sends the voice data to the server. The server converts the voice data into text data and analyzes it using an NLP engine to detect keywords such as "son" and "please send me money." The emotion engine also detects the user's feelings of fear or anxiety. If it determines that there is a high possibility of fraud, it immediately issues a warning to the user and notifies the police. Furthermore, an AI conversation agent is activated to continue the conversation on the user's behalf, allowing time for the police to respond.

[0631] As described above, the present invention is a system that protects users from telephone fraud in real time and further achieves more accurate fraud detection and warning by taking into account the user's emotional state.

[0632] The processing flow will be explained below.

[0633] Step 1:

[0634] A user makes or receives a call. The terminal captures the voice data of this call in real time. The captured voice data is compressed within the terminal and prepared for transmission to the server.

[0635] Step 2:

[0636] The device transmits the captured audio data to the server, and this communication uses a secure protocol to ensure data confidentiality.

[0637] Step 3:

[0638] The server processes the received voice data and converts it into text using speech recognition technology, which includes a highly accurate natural language processing (NLP) engine.

[0639] Step 4:

[0640] The server analyzes the text data, using an NLP engine to extract contextual information, entities (such as names of people or places), sentiment, and intent from the conversation, and passes the results of this analysis to a fraud detection algorithm.

[0641] Step 5:

[0642] The server then passes the text data to an emotion engine to evaluate the user's emotional state, which reads emotions from the user's tone of voice and word choice and adds them to the analysis results.

[0643] Step 6:

[0644] The server runs a fraud detection algorithm to assess the likelihood of fraud based on the analyzed text data and sentiment information. If certain keywords or phrases (e.g., "urgent," "transfer," "urgent") are detected, they are compared with a database of past frauds to determine the degree of match.

[0645] Step 7:

[0646] If the server determines that there is a high possibility of fraud, it will raise a warning flag and record the possibility of fraud. This flag will be immediately notified to the user.

[0647] Step 8:

[0648] When a warning flag is raised, the server immediately sends a warning signal to the device. The device receives this signal and displays a warning message on the screen or issues an audio notification to the user. The strength and type of the warning are adjusted according to the emotional state detected by the emotion engine.

[0649] Step 9:

[0650] If the server determines that there is a high possibility of fraud, it will collect detailed relevant information (e.g., the content of the conversation, the caller's phone number, the date and time of the conversation, etc.) and provide it to the police promptly. This information will be automatically forwarded in an appropriate format to the police's receiving system.

[0651] Step 10:

[0652] If the server determines that there is a high possibility of fraud, it sends an instruction to the device to activate an AI conversation agent. Based on this instruction, the device activates the AI ​​conversation agent and continues the conversation on behalf of the user.

[0653] Step 11:

[0654] The AI ​​conversational agent generates natural conversations on behalf of the user and continues the dialogue with the fraudster. For example, the AI ​​agent can ask, "Okay, where should I send the money?" to elicit additional information from the fraudster, buying time and assisting police in their response.

[0655] Step 12:

[0656] The server records all conversation data and stores it in a format that can be analyzed later. Newly detected fraud patterns and techniques are added to the database and used as training data to improve future detection accuracy.

[0657] Through these steps, the system can protect users from phone fraud in real time and assist police in responding quickly. Furthermore, by taking into account the user's emotional state, the accuracy of fraud detection and warning can be further improved.

[0658] Example 2

[0659] Next, a description will be given of Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[0660] In recent years, telephone fraud has been on the rise, with many people falling victim to it. Furthermore, fraudulent activities are becoming more complex and sophisticated, making it difficult to detect and respond to them in real time using conventional methods. Furthermore, the victim's emotional state can be an important factor in conversations with fraudsters, but conventional systems do not take this into account. This creates a need for rapid and accurate fraud detection and response.

[0661] The identification process by the identification processing unit 290 of the data processing device 12 in Example 2 is realized by the following means. In this invention, the server includes means for acquiring a user's telephone conversation as voice data in real time, means for converting the acquired voice data into text data, means for analyzing the text data to determine the possibility of fraud, means for notifying the user when the possibility of fraud is determined, means for providing information to relevant authorities when the possibility of fraud is determined, an artificial intelligence agent means for continuing the conversation on behalf of the user, means for analyzing the user's emotional state, and means for reevaluating the possibility of fraud based on the emotional state. This makes it possible to quickly and accurately detect fraudulent acts and take appropriate measures taking into account the user's emotional state.

[0662] "User" means any person or entity that uses the System to conduct telephone conversations.

[0663] A "telephone conversation" refers to a voice-based communication that a user has with another person via telephone.

[0664] "Voice data" refers to voice information captured in real time as a digital representation of a user's telephone conversation.

[0665] "Text data" refers to text information converted from voice data using voice recognition technology.

[0666] "Analysis" refers to the process of extracting the content and meaning of text or audio data and converting it into an understandable format.

[0667] "Possibility of fraud" refers to the result of assessing the degree of likelihood of fraud based on the acquired conversation content.

[0668] "Notification" refers to the act of the system displaying or audibly conveying warnings or information to the user.

[0669] "Relevant authorities" refers to organisations and bodies that are provided with information to deal with fraudulent activity, including the police.

[0670] An "artificial intelligence agent" refers to a sophisticated automated response program that can continue a dialogue on behalf of a user.

[0671] "Emotional state" refers to a user's mental or emotional response or state (e.g., joy, sadness, anger, fear, anxiety, etc.).

[0672] "Reappraisal" refers to the process of reviewing and updating the initial appraisal results based on the acquired emotional state.

[0673] The present invention is a system that monitors users' telephone conversations in real time and automatically detects fraud risks. Furthermore, by combining it with an emotion engine that recognizes the user's emotional state, it provides more advanced fraud detection and user protection functions.

[0674] Voice Input Processing

[0675] The device automatically captures the call audio data when the user makes or receives a call. The audio data is collected in real time and temporarily stored in the internal memory. This process can be performed using a smartphone or a dedicated device.

[0676] Audio data transmission and format conversion

[0677] The device converts the captured audio data into the optimal format (e.g., WAV, MP3) and compresses it using commonly used data compression algorithms (e.g., G.711, Opus). After conversion, the audio data is sent from the device to the server.

[0678] Converting audio data to text

[0679] The server receives the voice data sent from the device and converts it into text data using speech recognition technologies such as Google Cloud Speech-to-Text and Amazon Transcribe. This process utilizes a highly accurate natural language processing (NLP) engine.

[0680] Text data analysis

[0681] The server then analyzes the converted text data using natural language processing tools such as SpaCy and BERT. During the analysis step, contextual information, entities (e.g., people's names, place names), sentiment, and intent are extracted.

[0682] Performing emotion recognition

[0683] The server uses an emotion recognition engine such as IBM Watson Tone Analyzer to analyze the user's emotional state (happiness, sadness, anger, fear, anxiety, etc.) from the text data. This emotional information is used as additional input for subsequent fraud risk assessment.

[0684] Fraud risk assessment

[0685] The server assesses the fraud risk based on the analyzed text data and sentiment information. The fraud detection algorithm looks for specific keywords and phrases (e.g., "emergency," "transfer," "urgent") and matches them with patterns in a fraud database. If a high fraud risk is detected, the server flags the system.

[0686] User warning notification

[0687] If the server determines that there is a high risk of fraud, it immediately sends a warning signal to the terminal. The terminal then receives this warning signal and displays a warning message on the screen and also makes an audio notification to the user. For example, a message such as "There is a possibility of fraud, please be careful" may be displayed.

[0688] Data sharing with the police

[0689] If the server determines that there is a possibility of fraud, it collects relevant data such as the content of the conversation, the originating phone number, the date and time of the conversation, etc. This data is organized in an appropriate format (e.g., CSV file, JSON data) and automatically forwarded to the receiving system of the relevant agency, such as the police.

[0690] Launching an AI conversation agent

[0691] If the risk of fraud is deemed high, the server sends instructions to the device to activate an AI conversation agent. The device then receives this instruction, activates the AI ​​conversation agent, and generates natural conversation on behalf of the user. The AI ​​conversation agent uses the generative AI model to continue dialogue with the fraudster, buying time and assisting the police in their response.

[0692] Specific examples

[0693] For example, if a user answers a phone call and says, "I'm your son, please send me money now," the device immediately sends the voice data to the server. The server then converts the voice data into text using Google Cloud Speech-to-Text and uses SpaCy to extract keywords such as "son" and "please send me money." The emotion engine then analyzes the user's fear and anxiety to detect a high risk of fraud. The server then immediately sends a warning signal to the device, displays a warning message to the user, and notifies the police. Furthermore, an AI conversation agent continues the conversation on behalf of the user, allowing time for the police to respond.

[0694] Prompt Sentence Examples

[0695] "If a user is having a potentially fraudulent conversation on the phone, capture the audio data in real time and assess the likelihood of it being fraudulent. If it is, create a system that alerts the user and activates an AI conversation agent to continue the conversation on the user's behalf."

[0696] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0697] Step 1:

[0698] When a user makes or receives a call, the device automatically captures the call's audio data. Specifically, the device uses the microphone on the smartphone or dedicated device to collect audio data in real time and temporarily stores it in its internal memory. The input is the user's voice during the call, and the output is digital audio data.

[0699] Step 2:

[0700] The device converts the captured audio data into an optimal format (e.g., WAV, MP3) and compresses it. This is done using audio file format conversion software and a data compression algorithm (e.g., G.711, Opus). The input is the audio data obtained in step 1, and the output is the compressed audio data.

[0701] Step 3:

[0702] The device sends the compressed audio data to the server using a data transfer protocol over the Internet (e.g., HTTPS). The input is the compressed audio data, and the output is the audio data transferred to the server.

[0703] Step 4:

[0704] The server receives the voice data sent from the device and converts it into text data using speech recognition technologies such as Google Cloud Speech-to-Text and Amazon Transcribe. The input is the voice data transferred to the server, and the output is text data.

[0705] Step 5:

[0706] The server analyzes the converted text data using natural language processing tools (e.g., SpaCy, BERT). The analysis step extracts contextual information, entities (e.g., people's names, place names), sentiment, and intent. The input is text data, and the output is analyzed information (contextual information, entities, sentiment, and intent).

[0707] Step 6:

[0708] The server inputs the analyzed text data into an emotion recognition engine (e.g., IBM Watson Tone Analyzer) to analyze the user's emotional state. The emotion recognition engine extracts emotions such as joy, sadness, anger, fear, and anxiety from the text data. The input is the analyzed text data, and the output is emotional data.

[0709] Step 7:

[0710] The server assesses the fraud risk based on the analyzed text data and sentiment information. The fraud detection algorithm detects specific keywords and phrases (e.g., "emergency," "transfer," "urgent") and checks whether they match patterns in a fraud database. The input is text data and sentiment data, and the output is a fraud risk assessment result.

[0711] Step 8:

[0712] If the server evaluates the fraud risk as high, it immediately sends a warning signal to the terminal. The terminal receives this warning signal and displays a warning message on the screen for the user, and if necessary, also makes an audio notification. The input is the fraud risk evaluation result, and the output is a warning message for the user.

[0713] Step 9:

[0714] If the server determines that there is a high risk of fraud, it collects relevant data (e.g., conversation content, caller phone number, conversation date and time, etc.) and automatically forwards it to the police receiving system. This information is organized in an appropriate format (e.g., CSV file, JSON data). The input is the conversation content and related data, and the output is the data to be sent to the police.

[0715] Step 10:

[0716] If the risk of fraud is determined to be high, the server sends instructions to the device to activate an AI conversation agent. The device then receives this instruction, activates the AI ​​conversation agent, and generates a natural conversation on behalf of the user. The AI ​​conversation agent uses a generative AI model to continue the dialogue with the fraudster, buying time and assisting the police in their response. The input is instructions from the server, and the output is the dialogue with the fraudster.

[0717] (Application example 2)

[0718] Next, a description will be given of Application Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[0719] In modern society, telephone fraud is on the rise, posing a serious problem, especially for elderly users and those in situations where they are prone to anxiety. Conventional fraud prevention measures lack real-time detection of telephone fraud and warning functions that take into account the user's emotional state, making it difficult to prevent damage before it occurs. It is necessary for fraud detection systems to take the user's emotional state into account to achieve more accurate fraud detection and prompt warnings.

[0720] The identification processing by the identification processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means. In this invention, the server includes means for acquiring a user's telephone conversation as voice data in real time and converting the acquired voice data into text data, means for analyzing the text data to determine the possibility of fraud, means for notifying the user when the possibility of fraud is determined, means for recognizing the user's emotional state using an emotion recognition engine, and means for additionally providing the recognized emotional state to the fraud detection algorithm. This enables real-time fraud detection and warnings and responses based on the user's emotional state.

[0721] "User telephone conversation" refers to a user's verbal communication conducted over the telephone.

[0722] "Real-time" refers to processing or reaction occurring immediately, without delay.

[0723] "Audio data" refers to information that represents audio in digital form.

[0724] "Text data" refers to information obtained by converting voice data into a character string format.

[0725] "Means for converting" refers to a method or device for converting data of a particular format into another format.

[0726] "Means for analyzing to determine likelihood of fraud" refers to a method or device that analyzes input data to assess the risk of fraud.

[0727] "Means for notifying" refers to a method or device for notifying a user of specific information.

[0728] "Means of providing information to law enforcement" refers to a method or device that transmits specific information in an appropriate format to law enforcement agencies.

[0729] "Artificial intelligence agent means" refers to a method or device that uses artificial intelligence to act or interact on behalf of a user.

[0730] An "emotion recognition engine" refers to software or algorithms used to analyze and identify a user's emotional state.

[0731] "Fraud detection algorithm" refers to a computational procedure or program used to assess the risk of fraud based on specific patterns or keywords.

[0732] The present invention provides a system for automatically detecting fraud risks by monitoring users' telephone conversations in real time. Furthermore, by combining an emotion engine that recognizes the user's emotional state, it provides more advanced fraud detection and user protection capabilities. An embodiment of this system is described in detail below.

[0733] Voice Input Processing

[0734] When a user makes or receives a call, the device automatically captures the call's audio data in real time. The hardware used is a smartphone or tablet, and the software is a voice recognition library (e.g., speech_recognition). The captured audio data is converted and compressed into an optimal format by the device, and then sent to the server.

[0735] Converting audio data to text data

[0736] The server receives the voice data sent from the device and converts it into text using speech recognition technology, using a highly accurate natural language processing (NLP) engine (e.g., Google's speech recognition API).

[0737] Conversation analysis

[0738] The server analyzes the text data to extract contextual information, entities (such as names of people and places), sentiment, and intent from the conversation. This analysis process uses natural language processing techniques (e.g., NLP engines). The analyzed information is then passed to a fraud detection algorithm.

[0739] Emotion recognition by emotion engine

[0740] The server uses an emotion engine to further analyze the text data. The emotion engine recognizes the user's emotional state (e.g., joy, sadness, anger, fear, anxiety, etc.) This data is provided as additional input to the fraud detection algorithm.

[0741] Fraud Detection Algorithms

[0742] The server evaluates the likelihood of fraud based on the analyzed text data and sentiment information. Fraud detection algorithms detect specific keywords and phrases (e.g., "emergency," "transfer," "urgent") and check whether they match a database of past frauds. If potential fraud is detected, it is flagged by the server.

[0743] User notification function

[0744] If the server determines that there is a high possibility of fraud, it immediately sends a warning signal to the device. The device then displays a warning message on the screen or makes an audio notification to the user. The strength and type of the warning are adjusted according to the emotional state detected by the emotion engine.

[0745] Police contact function

[0746] If fraud is suspected, the server will collect relevant data (e.g., the content of the conversation, the caller's phone number, the date and time of the conversation, etc.) and provide it to the police promptly. This information will be automatically forwarded in an appropriate format to the police's receiving system.

[0747] AI conversational agent

[0748] If it determines that there is a high possibility of fraud, the server sends an instruction to the device to activate an AI conversation agent. The device then activates the AI ​​conversation agent based on this instruction and continues the conversation on the user's behalf. The AI ​​conversation agent generates natural conversation and continues to converse with the fraudster, buying time and supporting the police response.

[0749] Specific examples

[0750] For example, if a user answers a phone call and the caller says, "I'm your family, please send me money now," the device immediately sends the voice data to the server. The server converts the voice data into text data and analyzes it using an NLP engine to detect keywords such as "family" and "please send me money." The emotion engine also detects the user's feelings of fear or anxiety. If it determines that there is a high possibility of fraud, it immediately issues a warning to the user and notifies the police. Furthermore, an AI conversation agent is activated to continue the conversation on the user's behalf, allowing time for the police to respond.

[0751] Prompt Sentence Examples

[0752] Please provide sample code and an explanation for building a system that, when a user receives a phone call and the caller says, "I'm your family member, please send me money now," converts the voice data into text data, analyzes it using natural language processing, and determines whether it is likely to be a fraud.

[0753] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[0754] (System program processing flow)

[0755] Step 1: Audio Input Processing

[0756] The device captures the user's telephone conversation as voice data in real time. Specifically, it uses the smartphone's microphone to record the conversation during the call and temporarily stores the data. The input is the voice data captured in real time, and the output is the stored voice data.

[0757] Step 2: Converting audio data to text data

[0758] The device sends the captured voice data to the server, which uses voice recognition technology to convert the voice data into text data. Here, Google's voice recognition API is used to perform highly accurate text conversion. The input is voice data, and the output is text data.

[0759] Step 3: Analyzing the conversation

[0760] The server analyzes the text data to extract the conversation content, context information, entities, sentiment, and intent. This analysis uses natural language processing technology. The input is the text data, and the output is the analyzed context information and entity information.

[0761] Step 4: Emotion Recognition with the Emotion Engine

[0762] The server uses an emotion engine to recognize the user's emotional state from the text data. The emotion engine identifies emotions such as joy, sadness, anger, fear, and anxiety from the content and tone of the user's speech. The input is text data, and the output is emotional state data.

[0763] Step 5: Fraud Detection Algorithm

[0764] The server assesses the likelihood of fraud based on the analyzed text data and sentiment information. The fraud detection algorithm detects specific keywords and phrases and checks whether they match a historical fraud database. The input is contextual information, entity information, and sentiment data, and the output is a fraud risk assessment result.

[0765] Step 6: User Notification Function

[0766] If the server determines that there is a high possibility of fraud, it sends a warning signal to the terminal. The terminal receives this signal and displays a warning message on the screen or notifies the user by voice. The intensity and type of warning are adjusted according to the emotional state detected by the emotion engine. The input is the fraud evaluation result, and the output is a warning notification to the user.

[0767] Step 7: Contact the police

[0768] If the server determines that there is a possibility of fraud, it will collect relevant data (such as the content of the conversation, the caller's phone number, and the date and time of the conversation) and provide it to the police promptly. The information is automatically forwarded to the police's receiving system in an appropriate format. The input is the relevant data, and the output is a notification to the police.

[0769] Step 8: AI Conversational Agent

[0770] If the server determines that there is a high possibility of fraud, it sends an instruction to the terminal to activate an AI conversation agent. The terminal then activates the AI ​​conversation agent based on this instruction and continues the conversation on behalf of the user. The AI ​​conversation agent generates natural conversation and continues to converse with the fraudster, buying time and supporting police response. The input is the fraud assessment result, and the output is the activation of the AI ​​conversation agent.

[0771] The specific processing unit 290 transmits the result of the specific processing to the smart glasses 214. In the smart glasses 214, the control unit 46A causes the speaker 240 to output the result of the specific processing. The microphone 238 acquires audio indicating a user input regarding the result of the specific processing. The control unit 46A transmits audio data indicating the user input acquired by the microphone 238 to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the audio data.

[0772] The data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of the data generation model 58 is ChatGPT (Internet Search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search <url: https: gemini.google.com ?hl="ja">) and other generation AIs. The data generation model 58 is obtained by performing deep learning on a neural network. A prompt including an instruction is input to the data generation model 58, and inference data such as voice data indicating voice, text data indicating text, and image data indicating an image is also input. The data generation model 58 performs inference on the input inference data in accordance with the instruction indicated by the prompt, and outputs the inference result in a data format such as voice data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[0773] In the above embodiment, an example in which the specific processing is performed by the data processing device 12 has been given, but the technology of the present disclosure is not limited to this, and the specific processing may be performed by the smart glasses 214.

[0774] [Third embodiment]

[0775] FIG. 5 shows an example of the configuration of a data processing system 310 according to the third embodiment.

[0776] 5, the data processing system 310 includes the data processing device 12 and a headset terminal 314. An example of the data processing device 12 is a server.

[0777] The data processing device 12 includes a computer 22, a database 24, and a communication I / F 26. The computer 22 is an example of a "computer" according to the technology of the present disclosure. The computer 22 includes a processor 28, a RAM 30, and a storage 32. The processor 28, the RAM 30, and the storage 32 are connected to a bus 34. The database 24 and the communication I / F 26 are also connected to the bus 34. The communication I / F 26 is connected to a network 54. Examples of the network 54 include a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0778] The headset type terminal 314 includes a computer 36, a microphone 238, a speaker 240, a camera 42, a communication I / F 44, and a display 343. The computer 36 includes a processor 46, a RAM 48, and a storage 50. The processor 46, the RAM 48, and the storage 50 are connected to a bus 52. The microphone 238, the speaker 240, the camera 42, and the display 343 are also connected to the bus 52.

[0779] The microphone 238 receives instructions and the like from the user 20 by receiving voice uttered by the user 20. The microphone 238 captures the voice uttered by the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio in accordance with instructions from the processor 46.

[0780] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an imaging element such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the surroundings of user 20 (for example, an imaging range defined by an angle of view equivalent to the field of vision of a typical healthy person).

[0781] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 are responsible for the exchange of various information between the processor 46 and the processor 28 via the network 54. The exchange of various information between the processor 46 and the processor 28 using the communication I / Fs 44 and 26 is carried out in a secure state.

[0782] Fig. 6 shows an example of the main functions of the data processing device 12 and the headset type terminal 314. As shown in Fig. 6, in the data processing device 12, a specific process is performed by the processor 28. A specific process program 56 is stored in the storage 32.

[0783] The specific processing program 56 is an example of a "program" according to the technology of the present disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific processing is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific processing program 56 executed on the RAM 30.

[0784] The storage 32 stores a data generation model 58 and an emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[0785] In the headset type terminal 314, a reception output process is performed by the processor 46. A reception output program 60 is stored in the storage 50. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output process is realized by the processor 46 operating as the control unit 46A in accordance with the reception output program 60 executed on the RAM 48.

[0786] Next, a description will be given of the identification process performed by the identification processing unit 290 of the data processing device 12. In the following description, the data processing device 12 will be referred to as the "server" and the headset type terminal 314 will be referred to as the "terminal."

[0787] The present invention is a system for monitoring users' telephone conversations in real time and automatically detecting fraud risks. Detailed embodiments for specifically implementing the system are described below.

[0788] Voice Input Processing

[0789] When a user makes or receives a call, the terminal automatically captures the call's audio data in real time, which is then converted and compressed into an optimal format by the terminal and then sent to the server.

[0790] Converting audio data to text data

[0791] The server receives the voice data sent from the device and converts it into text data using speech recognition technology, which includes a highly accurate natural language processing (NLP) engine.

[0792] Conversation analysis

[0793] The server analyzes the text data and extracts contextual information, entities (such as names of people and places), sentiment, and intent from the conversation. This analysis process is performed using natural language processing techniques. The analyzed information is then passed to a fraud detection algorithm.

[0794] Fraud Detection Algorithms

[0795] The server evaluates the likelihood of fraud based on the analyzed text data. Fraud detection algorithms look for specific keywords and phrases (e.g., "urgent," "transfer," "urgent") and check whether they match previous fraud cases. If potential fraud is detected, it is flagged by the server.

[0796] User notification function

[0797] If the server determines that there is a high possibility of fraud, it immediately sends a warning signal to the device, which then displays a warning message on the screen or makes an audio notification based on the received warning signal.

[0798] Police contact function

[0799] If the server determines that there is a possibility of fraud, it will promptly provide the police with relevant data (e.g., the content of the conversation, the caller's phone number, the date and time of the conversation, etc.) This information will be automatically forwarded to the police's receiving system in an appropriate format.

[0800] AI conversational agent

[0801] If it determines that there is a high possibility of fraud, the server sends an instruction to the device to activate an AI conversation agent. The device then activates the AI ​​conversation agent based on the received instruction and continues the conversation on behalf of the user. The AI ​​conversation agent generates natural conversation and continues to converse with the fraudster, buying time and supporting police response.

[0802] Specific examples

[0803] For example, if a user answers a phone call and the caller says, "I'm your son, please send me money now," the device immediately sends the voice data to the server. The server converts the voice data into text data and analyzes it using an NLP engine. If keywords such as "son" and "please send me money" are detected, it is determined to be a fraudulent call. A warning is immediately issued to the user and the police are notified. Furthermore, an AI conversation agent is activated to continue the conversation on the user's behalf, allowing time for the police to respond.

[0804] Thus, the present invention provides a system that keeps users safe from telephone fraud and allows for fast and effective police response.

[0805] The processing flow will be explained below.

[0806] Step 1:

[0807] A user makes or receives a call. The terminal captures the voice data of this phone conversation in real time. The voice data is compressed within the terminal and prepared to be sent to the server for further processing.

[0808] Step 2:

[0809] The device transmits the captured audio data to the server, and this communication uses a secure protocol to ensure data confidentiality.

[0810] Step 3:

[0811] The server receives the received voice data and converts it into text using speech recognition technology. The speech recognition model utilizes a highly accurate natural language processing (NLP) engine.

[0812] Step 4:

[0813] The server analyzes the text data and uses an NLP engine to extract contextual information, entities (people's names, places, etc.), sentiment, and intent from the conversation. The results of this analysis are passed to a fraud detection algorithm.

[0814] Step 5:

[0815] The server runs fraud detection algorithms to evaluate potential fraud patterns. If certain keywords or phrases (e.g., "urgent," "transfer," "urgent") are detected, they are compared against a database of past frauds to determine the likelihood of a match.

[0816] Step 6:

[0817] If the server determines that there is a high possibility of fraud, it will raise a warning flag and record the possibility of fraud. This flag will be immediately notified to the user.

[0818] Step 7:

[0819] When a warning flag is raised, the server immediately sends a warning signal to the terminal, which then displays a warning message on the screen or notifies the user by voice.

[0820] Step 8:

[0821] The server collects relevant information about the flagged call (e.g., the caller's phone number, the content of the conversation, the date and time, etc.) and prepares it for automatic provision to the police. This information is then forwarded in an appropriate format to the police's receiving system.

[0822] Step 9:

[0823] If the server determines that there is a high possibility of fraud, it sends an instruction to the device to activate an AI conversation agent. Based on this instruction, the device activates the AI ​​conversation agent and continues the conversation on behalf of the user.

[0824] Step 10:

[0825] The AI ​​conversational agent generates natural conversations on behalf of the user and continues the dialogue with the fraudster, buying time and assisting law enforcement. For example, the AI ​​agent can ask, "Okay, where should I send the money?" to elicit additional information from the fraudster.

[0826] Step 11:

[0827] All conversation data is recorded by the server and stored in a format that can be analyzed later. Newly detected fraud patterns and techniques are added to the database and used as training data to improve future detection accuracy.

[0828] Through these steps, the system can protect users from telephone fraud in real time and assist police in responding quickly.

[0829] Example 1

[0830] Next, a description will be given of Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal."

[0831] Telephone scams are becoming more sophisticated every year, making vulnerable individuals such as the elderly particularly vulnerable to fraud. Current manual countermeasures often delay detection and response to scams, often failing to detect them until actual harm has occurred. Furthermore, continuing to talk to scammers is mentally stressful for users, so a fast and effective response method is needed.

[0832] The specific processing by the specific processing unit 290 of the data processing device 12 in the first embodiment is realized by the following means.

[0833] In this invention, the server includes: means for capturing a user's telephone conversation as voice data in real time; means for converting and compressing the captured voice data into an optimal format and transmitting it to the server; means for receiving the voice data and converting it into text data using voice recognition technology; means for analyzing the text data and extracting contextual information, entities, emotions, and intent; means for assessing the possibility of fraud based on the text data; means for notifying the user if a possibility of fraud is determined; means for providing information to the police if a possibility of fraud is determined; and an artificial intelligence agent for continuing the conversation on behalf of the user. This makes it possible to detect telephone fraud in real time, quickly warn the user, and contact the police if necessary, thereby preventing damage before it occurs. Furthermore, by having the artificial intelligence agent continue the conversation with the fraudster, it is possible to reduce the user's mental burden and ensure time for the police to respond.

[0834] "User" refers to a person who uses the system.

[0835] A "telephone conversation" refers to a voice communication that a user has over the telephone.

[0836] "Voice data" refers to the digital recording of a user's telephone conversation.

[0837] "Format" refers to the data structure and file format used to save audio data.

[0838] "Compression" refers to a process for reducing the data size of audio data.

[0839] "Server" refers to a computer system for processing and analyzing audio data.

[0840] "Speech recognition technology" refers to technology for analyzing voice data and converting its contents into text data.

[0841] "Text data" refers to character string data converted from audio data.

[0842] "Contextual information" refers to information about the background and situation of the conversation content.

[0843] "Entity" refers to proper nouns or specific information (e.g., people's names, place names, organization names, etc.) that appear in the conversation content.

[0844] "Emotion" refers to the speaker's emotional state (e.g., joy, anger, sadness, etc.) extracted from the conversation content.

[0845] "Intention" refers to the purpose or aim that a speaker is trying to convey through a conversation.

[0846] "Fraud Potential" refers to the result of assessing whether the content of a conversation poses a risk of fraudulent activity.

[0847] "Notification" refers to a message or signal that displays a warning or information to a user.

[0848] "Police" refers to a public security agency or its affiliates.

[0849] "Means of providing information" refers to the methods and functions for transmitting relevant information from the server to the police.

[0850] An "artificial intelligence agent" refers to an intelligent program that automatically converses on behalf of a user.

[0851] The present invention is a real-time monitoring system for protecting users from telephone fraud. This system monitors users' telephone conversations in real time and automatically detects and responds to fraud risks. Detailed embodiments for specifically implementing the present invention are described below.

[0852] Voice Input Processing

[0853] When a user makes or receives a call, the device automatically captures the call's audio data in real time, converts the captured audio data into an optimal format (e.g., from WAV to MP3), compresses it, and then sends it to the server.

[0854] Converting audio data to text data

[0855] The server receives the voice data sent from the device and converts it into text data using voice recognition technology (e.g., a high-precision voice recognition engine). Software used here includes the Google Cloud Speech-to-Text API.

[0856] Conversation analysis

[0857] The server analyzes the text data. Natural language processing techniques (e.g., SpaCy) are used to extract contextual information, entities (people's names, places, etc.), sentiment, and intent. The analyzed information is then passed to a fraud detection algorithm for evaluation.

[0858] Fraud Detection Algorithms

[0859] The server evaluates the likelihood of fraud based on the analyzed text data. Fraud detection algorithms look for specific keywords and phrases (e.g., "urgent," "transfer," "urgent") and check whether they match previous fraud cases. Potential fraud is detected and flagged.

[0860] User notification function

[0861] If the server determines that there is a high possibility of fraud, it immediately sends a warning signal to the device, which then displays a warning message on the screen or makes an audio notification based on the received warning signal.

[0862] Police contact function

[0863] If a possible fraud is identified, the server will promptly provide the police with relevant data (e.g., the content of the conversation, the caller's phone number, the date and time of the conversation, etc.) This information will be automatically forwarded to the police's receiving system in an appropriate format.

[0864] AI conversational agent

[0865] If it determines that there is a high possibility of fraud, the server sends an instruction to the device to launch an AI conversation agent. The device then launches the AI ​​conversation agent based on this instruction and continues the conversation on the user's behalf. The AI ​​conversation agent generates natural conversation and continues to converse with the fraudster, buying time and supporting police response.

[0866] Specific examples

[0867] For example, if a user answers a phone call and the caller says, "I'm your son, please send me money now," the device immediately sends the voice data to the server. The server converts the voice data into text data and analyzes it using an NLP engine. If keywords such as "son" and "please send me money" are detected, it is determined to be a fraudulent call. A warning is immediately issued to the user and the police are notified. Furthermore, an AI conversation agent is activated to continue the conversation on the user's behalf, allowing time for the police to respond.

[0868] Prompt Sentence Examples

[0869] By inputting the prompt "Please explain in detail how the system will respond if a user encounters a phone scam," the generative AI model will explain the detailed operation of each processing step. Furthermore, by using the prompt "Please tell me how to convert voice data into text and assess the risk of fraud," a detailed explanation of the speech recognition and fraud detection algorithms will be obtained.

[0870] The present invention provides a system that keeps users safe from telephone fraud and enables fast and effective police response.

[0871] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0872] Program processing flow

[0873] Step 1: Getting voice input

[0874] When the device detects a user's phone call, it captures audio data in real time through the microphone. The input is the user's voice, and the output is digital audio data. The device converts this captured audio data into an optimal format (e.g., WAV format). After conversion, the data is compressed (e.g., converted to MP3 format).

[0875] Step 2: Sending audio data to the server

[0876] The device sends the compressed audio data to the server. The input is the compressed audio data obtained in step 1, and the output is the data to be sent to the server. This data is sent using a secure protocol (e.g., HTTPS).

[0877] Step 3: Converting audio data to text data

[0878] The server receives the voice data sent from the device and converts it into text data using a speech recognition engine (e.g., Google Cloud Speech-to-Text API). The input is compressed voice data, and the output is text data. Specifically, the server analyzes the voice file and outputs the spoken content as a string of characters.

[0879] Step 4: Analyzing the conversation

[0880] The server analyzes the text data. It uses natural language processing techniques (e.g., SpaCy) to extract contextual information, entities, sentiment, and intent. The input is text data, and the output is analyzed data. Specifically, the server tokenizes the text data and analyzes the meaning of each token.

[0881] Step 5: Fraud Potential Assessment

[0882] The server evaluates the likelihood of fraud based on the analysis results. It checks whether certain keywords or phrases (e.g., "urgent," "transfer," "urgent") match a database of past frauds. The input is the analyzed data, and the output is an assessment of the likelihood of fraud. Based on the assessment, if there is a high likelihood of fraud, it is flagged.

[0883] Step 6: User Notification

[0884] If it is determined that there is a high possibility of fraud, the server immediately sends a warning signal to the terminal. The input is the fraud evaluation result, and the output is the warning signal to the terminal. Based on the received warning signal, the terminal displays a warning message on the screen or issues an audio notification.

[0885] Step 7: Contact the police

[0886] If it is determined that there is a high possibility of fraud, the server will provide the police with relevant data (e.g., the content of the conversation, the caller's phone number, the date and time of the conversation, etc.). The input is the fraud assessment result and relevant data, and the output is the data to be sent to the police. The server formats the data in an appropriate format and forwards it to the police's receiving system.

[0887] Step 8: Launching the AI ​​Conversation Agent

[0888] The server sends an instruction to launch the AI ​​conversation agent to the device. The input is the fraud evaluation result, and the output is the startup instruction to the device. The device then launches the AI ​​conversation agent based on the instruction and continues the conversation on behalf of the user. Specifically, a natural conversation is conducted based on the prompt sentences generated by the AI ​​conversation agent.

[0889] Through the above processing steps, the system of the present invention monitors telephone fraud in real time and responds quickly and efficiently.

[0890] (Application example 1)

[0891] Next, a description will be given of Application Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal."

[0892] In recent years, telephone fraud has been on the rise, and countermeasures are urgently needed. However, with current systems, it takes time for users to realize they have been scammed, which often results in greater damage. Furthermore, delays in contacting authorities such as the police make it difficult to take prompt action. Furthermore, the pressure of continuing a conversation with the fraudster places a heavy psychological burden on users. Therefore, there is a need for a system that can detect telephone fraud in real time, immediately warn users, and, if necessary, contact authorities such as the police and take action on their behalf.

[0893] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 1 is realized by the following means.

[0894] In this invention, the server includes: means for acquiring a user's telephone conversation as voice data in real time; means for converting the acquired voice data into text data; means for analyzing the text data to determine the possibility of fraud; means for monitoring the voice data in real time and assessing the risk of fraud; means for determining the risk of fraud based on the detection of specific keywords and contextual information and displaying a warning; means for notifying the user when a possibility of fraud is determined; means for providing information to authorities such as the police when a possibility of fraud is determined; and artificial intelligence agent means for continuing the conversation on behalf of the user. This allows the user to know the risk of telephone fraud in real time and respond quickly. Furthermore, since authorities such as the police can be contacted quickly, a quick response can be expected. Furthermore, by having the artificial intelligence agent continue the conversation on behalf of the user, the user's mental burden can be reduced.

[0895] A "user" is someone who uses the system.

[0896] A "telephone conversation" is a dialogue conducted through voice communication.

[0897] "Voice data" refers to information that has been digitized and recorded from a user's telephone conversation.

[0898] "Text data" is voice data converted into character information.

[0899] "Analyzing text data" means using natural language processing technology to understand the content of the text data and extract specific meanings and information.

[0900] "Determining the likelihood of fraud" means that the system evaluates whether there is a risk of fraud based on the analyzed text data.

[0901] "Notifying the user" means that if the system determines that there is a possibility of fraud, it will provide the user with information to warn or caution them.

[0902] "Providing information to police or other authorities" means transmitting the content of the conversation and related information to police or other law enforcement authorities if it is determined that there is a possibility of fraud.

[0903] An "artificial intelligence agent" is a program that carries on a telephone conversation on behalf of a user and mimics natural dialogue.

[0904] "Monitoring voice data in real time" means instantly monitoring a user's telephone conversation and analyzing the voice data as needed.

[0905] "Assessing the risk of fraud" means determining the possibility of fraud based on the content of the text data.

[0906] "Detecting specific keywords and contextual information" means finding important words and context related to fraudulent activity from the analyzed text data.

[0907] "Display a warning" means that the system will display a warning message on the user's terminal if a fraud risk is detected.

[0908] The present invention relates to a system for monitoring a user's telephone conversations in real time and assessing the risk of fraud. Detailed embodiments for specifically implementing this system are described below.

[0909] System Configuration

[0910] The system mainly consists of the following components:

[0911] 1. User's device

[0912] 2. Server

[0913] 3. Artificial Intelligence Agents

[0914] User's device

[0915] The user's device is a communication device such as a smartphone, and captures the user's telephone conversation as voice data in real time. The device converts the captured voice data into an optimal format, compresses it, and sends it to the server. As a specific example, the device uses the "Google Cloud Speech-to-Text API" as voice recognition technology to convert the voice data into text data.

[0916] server

[0917] The server receives the voice data sent from the device and analyzes it in real time. The voice data is converted into text data using speech recognition technology, and then analyzed using a natural language processing (NLP) engine. The server performs the following processes based on the analyzed text data:

[0918] Detect specific keywords and contextual information to assess fraud risk.

[0919] If the risk is deemed high, a warning will be sent to the user.

[0920] Providing information to police and other authorities as necessary.

[0921] The server detects specific keywords (e.g., "urgent," "transfer," "urgent") and contextual information based on an algorithm to assess the risk of fraud. This determines whether there is a risk of fraud. Natural language processing technologies used in this process include Google Cloud Natural Language API.

[0922] Artificial Intelligence Agent

[0923] An AI agent is activated when a high risk of fraud is detected and continues the conversation on behalf of the user. The AI ​​agent mimics natural dialogue and maintains a conversation with the fraudster, freeing up police response time.

[0924] Specific examples

[0925] For example, if a user answers a phone call and the caller says, "I'm your son, please send me money now," the device immediately sends the voice data to the server. The server converts the voice data into text data and analyzes it using an NLP engine. If keywords such as "son" and "please send me money" are detected, it is determined to be a fraudulent call. A warning is immediately issued to the user, and the police are notified. An artificial intelligence agent is then activated to continue the conversation on the user's behalf.

[0926] Prompt Sentence Examples

[0927] Here is an example prompt for creating an AI model:

[0928] "Create an NLP model that converts phone voice data into text in real time and detects fraud risk."

[0929] "Convert the audio data into text data using the Google Cloud Speech-to-Text API."

[0930] As can be seen, the present invention provides a system that keeps users safe from telephone fraud and allows for a fast and effective police response.

[0931] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0932] Step 1:

[0933] The device receives the user's phone conversation in real time, capturing voice data through a microphone and converting it into a digital format.

[0934] Input: User's phone conversation (voice data)

[0935] Data processing: capturing audio data and converting it to a digital format

[0936] Output: Digital audio data

[0937] Step 2:

[0938] The device converts the acquired audio data into the optimal format and compresses it.

[0939] Input: Digital audio data

[0940] Data processing: format conversion and compression

[0941] Output: Compressed audio data

[0942] Step 3:

[0943] The terminal transmits the compressed audio data to the server.

[0944] Input: Compressed audio data

[0945] Data calculation: Data transmission

[0946] Output: Data transfer to the server

[0947] Step 4:

[0948] The server receives the voice data sent from the device and converts it into text data using voice recognition technology, using the Google Cloud Speech-to-Text API.

[0949] Input: Compressed audio data

[0950] Data processing: speech recognition and text conversion

[0951] Output: Text data

[0952] Step 5:

[0953] The server analyzes the text data using a natural language processing (NLP) engine, using the Google Cloud Natural Language API to extract specific keywords and contextual information from the text data.

[0954] Input: Text data

[0955] Data processing: Information extraction through natural language analysis

[0956] Output: Analysis results (specific keywords and contextual information)

[0957] Step 6:

[0958] The server evaluates the fraud risk based on the analysis results, checking whether specific keywords (e.g., "emergency," "transfer," "urgent") are included to determine the fraud risk.

[0959] Input: Analysis results

[0960] Data Computing: Applying Fraud Risk Assessment Algorithms

[0961] Output: Fraud risk rating (high, medium, low)

[0962] Step 7:

[0963] If the server determines that the risk of fraud is high, it will send a warning to the user's device.

[0964] Input: Fraud Risk Rating (High)

[0965] Data Calculation: Generating and Sending Warning Messages

[0966] Output: Warning notification to the user terminal

[0967] Step 8:

[0968] The user's device will display the received warning on the screen and also alert them with audio notifications and vibrations.

[0969] Input: Warning notification from the server

[0970] Data processing: Display warning message

[0971] Output: User attention (screen display and audio notification)

[0972] Step 9:

[0973] If the server determines there is a high risk of fraud, it will provide relevant information to authorities such as the police, sending data including the content of the conversation and the caller's number.

[0974] Input: Fraud Risk Rating (High)

[0975] Data Computing: Generating and transmitting relevant information

[0976] Output: Providing information to authorities such as police

[0977] Step 10:

[0978] If the server determines that the fraud risk is high, it sends an instruction to the terminal to activate an AI agent, which continues the conversation on behalf of the user.

[0979] Input: Fraud Risk Rating (High)

[0980] Data Calculation: Instructions for launching artificial intelligence agents

[0981] Output: Launch of artificial intelligence agent

[0982] Step 11:

[0983] The artificial intelligence agent mimics natural dialogue and continues the conversation on behalf of the user, maintaining a conversation with the fraudster to free up police response time.

[0984] Input: Conversational content and user instructions

[0985] Data processing: Natural conversation generation

[0986] Output: Ongoing dialogue with the fraudster

[0987] Furthermore, an emotion engine that estimates the user's emotion may be further combined. That is, the identification processing unit 290 may estimate the user's emotion using the emotion identification model 59, and perform identification processing using the user's emotion.

[0988] The present invention is a system that monitors users' telephone conversations in real time and automatically detects fraud risks. Furthermore, by combining it with an emotion engine that recognizes the user's emotional state, it provides more advanced fraud detection and user protection functions. Detailed embodiments for specifically implementing this system are described below.

[0989] Voice Input Processing

[0990] When a user makes or receives a call, the terminal automatically captures the call's audio data in real time, which is then converted and compressed into an optimal format by the terminal and then sent to the server.

[0991] Converting audio data to text data

[0992] The server receives the voice data sent from the device and converts it into text data using voice recognition technology, which includes a highly accurate natural language processing (NLP) engine.

[0993] Conversation analysis

[0994] The server analyzes the text data and extracts contextual information, entities (such as names of people and places), sentiment, and intent from the conversation. This analysis process is performed using natural language processing techniques. The analyzed information is then passed to a fraud detection algorithm.

[0995] Emotion recognition by emotion engine

[0996] The server uses an emotion engine to further analyze the text data, which recognizes the user's emotional state (e.g., joy, sadness, anger, fear, anxiety, etc.) This data is provided as additional input to the fraud detection algorithm.

[0997] Fraud Detection Algorithms

[0998] The server evaluates the likelihood of fraud based on the analyzed text data and sentiment information. Fraud detection algorithms detect specific keywords and phrases (e.g., "emergency," "transfer," "urgent") and check whether they match a database of past frauds. If potential fraud is detected, it is flagged by the server.

[0999] User notification function

[1000] If the server determines that there is a high possibility of fraud, it immediately sends a warning signal to the device. The device then displays a warning message on the screen or makes an audio notification to the user. The strength and type of the warning are adjusted according to the emotional state detected by the emotion engine.

[1001] Police contact function

[1002] If the server determines that there is a possibility of fraud, it will collect relevant data (e.g., the content of the conversation, the caller's phone number, the date and time of the conversation, etc.) and provide it to the police promptly. This information will be automatically forwarded in an appropriate format to the police's receiving system.

[1003] AI conversational agent

[1004] If it determines that there is a high possibility of fraud, the server sends an instruction to the device to activate an AI conversation agent. The device then activates the AI ​​conversation agent based on this instruction and continues the conversation on the user's behalf. The AI ​​conversation agent generates natural conversation and continues to converse with the fraudster, buying time and supporting the police response.

[1005] Specific examples

[1006] For example, if a user answers a phone call and the caller says, "I'm your son, please send me money now," the device immediately sends the voice data to the server. The server converts the voice data into text data and analyzes it using an NLP engine to detect keywords such as "son" and "please send me money." The emotion engine also detects the user's feelings of fear or anxiety. If it determines that there is a high possibility of fraud, it immediately issues a warning to the user and notifies the police. Furthermore, an AI conversation agent is activated to continue the conversation on the user's behalf, allowing time for the police to respond.

[1007] As described above, the present invention is a system that protects users from telephone fraud in real time and further achieves more accurate fraud detection and warning by taking into account the user's emotional state.

[1008] The processing flow will be explained below.

[1009] Step 1:

[1010] A user makes or receives a call. The terminal captures the voice data of this call in real time. The captured voice data is compressed within the terminal and prepared for transmission to the server.

[1011] Step 2:

[1012] The device transmits the captured audio data to the server, and this communication uses a secure protocol to ensure data confidentiality.

[1013] Step 3:

[1014] The server processes the received voice data and converts it into text using speech recognition technology, which includes a highly accurate natural language processing (NLP) engine.

[1015] Step 4:

[1016] The server analyzes the text data, using an NLP engine to extract contextual information, entities (such as names of people or places), sentiment, and intent from the conversation, and passes the results of this analysis to a fraud detection algorithm.

[1017] Step 5:

[1018] The server then passes the text data to an emotion engine to evaluate the user's emotional state, which reads emotions from the user's tone of voice and word choice and adds them to the analysis results.

[1019] Step 6:

[1020] The server runs a fraud detection algorithm to assess the likelihood of fraud based on the analyzed text data and sentiment information. If certain keywords or phrases (e.g., "urgent," "transfer," "urgent") are detected, they are compared with a database of past frauds to determine the degree of match.

[1021] Step 7:

[1022] If the server determines that there is a high possibility of fraud, it will raise a warning flag and record the possibility of fraud. This flag will be immediately notified to the user.

[1023] Step 8:

[1024] When a warning flag is raised, the server immediately sends a warning signal to the device. The device receives this signal and displays a warning message on the screen or issues an audio notification to the user. The strength and type of the warning are adjusted according to the emotional state detected by the emotion engine.

[1025] Step 9:

[1026] If the server determines that there is a high possibility of fraud, it will collect detailed relevant information (e.g., the content of the conversation, the caller's phone number, the date and time of the conversation, etc.) and provide it to the police promptly. This information will be automatically forwarded in an appropriate format to the police's receiving system.

[1027] Step 10:

[1028] If the server determines that there is a high possibility of fraud, it sends an instruction to the device to activate an AI conversation agent. Based on this instruction, the device activates the AI ​​conversation agent and continues the conversation on behalf of the user.

[1029] Step 11:

[1030] The AI ​​conversational agent generates natural conversations on behalf of the user and continues the dialogue with the fraudster. For example, the AI ​​agent can ask, "Okay, where should I send the money?" to elicit additional information from the fraudster, buying time and assisting police in their response.

[1031] Step 12:

[1032] The server records all conversation data and stores it in a format that can be analyzed later. Newly detected fraud patterns and techniques are added to the database and used as training data to improve future detection accuracy.

[1033] Through these steps, the system can protect users from phone fraud in real time and assist police in responding quickly. Furthermore, by taking into account the user's emotional state, the accuracy of fraud detection and warning can be further improved.

[1034] Example 2

[1035] Next, a description will be given of Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal."

[1036] In recent years, telephone fraud has been on the rise, with many people falling victim to it. Furthermore, fraudulent activities are becoming more complex and sophisticated, making it difficult to detect and respond to them in real time using conventional methods. Furthermore, the victim's emotional state can be an important factor in conversations with fraudsters, but conventional systems do not take this into account. This creates a need for rapid and accurate fraud detection and response.

[1037] The identification process by the identification processing unit 290 of the data processing device 12 in Example 2 is realized by the following means. In this invention, the server includes means for acquiring a user's telephone conversation as voice data in real time, means for converting the acquired voice data into text data, means for analyzing the text data to determine the possibility of fraud, means for notifying the user when the possibility of fraud is determined, means for providing information to relevant authorities when the possibility of fraud is determined, an artificial intelligence agent means for continuing the conversation on behalf of the user, means for analyzing the user's emotional state, and means for reevaluating the possibility of fraud based on the emotional state. This makes it possible to quickly and accurately detect fraudulent acts and take appropriate measures taking into account the user's emotional state.

[1038] "User" means any person or entity that uses the System to conduct telephone conversations.

[1039] A "telephone conversation" refers to a voice-based communication that a user has with another person via telephone.

[1040] "Voice data" refers to voice information captured in real time as a digital representation of a user's telephone conversation.

[1041] "Text data" refers to text information converted from voice data using voice recognition technology.

[1042] "Analysis" refers to the process of extracting the content and meaning of text or audio data and converting it into an understandable format.

[1043] "Possibility of fraud" refers to the result of assessing the degree of likelihood of fraud based on the acquired conversation content.

[1044] "Notification" refers to the act of the system displaying or audibly conveying warnings or information to the user.

[1045] "Relevant authorities" refers to organisations and bodies that are provided with information to deal with fraudulent activity, including the police.

[1046] An "artificial intelligence agent" refers to a sophisticated automated response program that can continue a dialogue on behalf of a user.

[1047] "Emotional state" refers to a user's mental or emotional response or state (e.g., joy, sadness, anger, fear, anxiety, etc.).

[1048] "Reappraisal" refers to the process of reviewing and updating the initial appraisal results based on the acquired emotional state.

[1049] The present invention is a system that monitors users' telephone conversations in real time and automatically detects fraud risks. Furthermore, by combining it with an emotion engine that recognizes the user's emotional state, it provides more advanced fraud detection and user protection functions.

[1050] Voice Input Processing

[1051] The device automatically captures the call audio data when the user makes or receives a call. The audio data is collected in real time and temporarily stored in the internal memory. This process can be performed using a smartphone or a dedicated device.

[1052] Audio data transmission and format conversion

[1053] The device converts the captured audio data into the optimal format (e.g., WAV, MP3) and compresses it using commonly used data compression algorithms (e.g., G.711, Opus). After conversion, the audio data is sent from the device to the server.

[1054] Converting audio data to text

[1055] The server receives the voice data sent from the device and converts it into text data using speech recognition technologies such as Google Cloud Speech-to-Text and Amazon Transcribe. This process utilizes a highly accurate natural language processing (NLP) engine.

[1056] Text data analysis

[1057] The server then analyzes the converted text data using natural language processing tools such as SpaCy and BERT. During the analysis step, contextual information, entities (e.g., people's names, place names), sentiment, and intent are extracted.

[1058] Performing emotion recognition

[1059] The server uses an emotion recognition engine such as IBM Watson Tone Analyzer to analyze the user's emotional state (happiness, sadness, anger, fear, anxiety, etc.) from the text data. This emotional information is used as additional input for subsequent fraud risk assessment.

[1060] Fraud risk assessment

[1061] The server assesses the fraud risk based on the analyzed text data and sentiment information. The fraud detection algorithm looks for specific keywords and phrases (e.g., "emergency," "transfer," "urgent") and matches them with patterns in a fraud database. If a high fraud risk is detected, the server flags the system.

[1062] User warning notification

[1063] If the server determines that there is a high risk of fraud, it immediately sends a warning signal to the terminal. The terminal then receives this warning signal and displays a warning message on the screen and also makes an audio notification to the user. For example, a message such as "There is a possibility of fraud, please be careful" may be displayed.

[1064] Data sharing with the police

[1065] If the server determines that there is a possibility of fraud, it collects relevant data such as the content of the conversation, the originating phone number, the date and time of the conversation, etc. This data is organized in an appropriate format (e.g., CSV file, JSON data) and automatically forwarded to the receiving system of the relevant agency, such as the police.

[1066] Launching an AI conversation agent

[1067] If the risk of fraud is deemed high, the server sends instructions to the device to activate an AI conversation agent. The device then receives this instruction, activates the AI ​​conversation agent, and generates natural conversation on behalf of the user. The AI ​​conversation agent uses the generative AI model to continue dialogue with the fraudster, buying time and assisting the police in their response.

[1068] Specific examples

[1069] For example, if a user answers a phone call and says, "I'm your son, please send me money now," the device immediately sends the voice data to the server. The server then converts the voice data into text using Google Cloud Speech-to-Text and uses SpaCy to extract keywords such as "son" and "please send me money." The emotion engine then analyzes the user's fear and anxiety to detect a high risk of fraud. The server then immediately sends a warning signal to the device, displays a warning message to the user, and notifies the police. Furthermore, an AI conversation agent continues the conversation on behalf of the user, allowing time for the police to respond.

[1070] Prompt Sentence Examples

[1071] "If a user is having a potentially fraudulent conversation on the phone, capture the audio data in real time and assess the likelihood of it being fraudulent. If it is, create a system that alerts the user and activates an AI conversation agent to continue the conversation on the user's behalf."

[1072] The flow of the identification process in the second embodiment will be described with reference to FIG.

[1073] Step 1:

[1074] When a user makes or receives a call, the device automatically captures the call's audio data. Specifically, the device uses the microphone on the smartphone or dedicated device to collect audio data in real time and temporarily stores it in its internal memory. The input is the user's voice during the call, and the output is digital audio data.

[1075] Step 2:

[1076] The device converts the captured audio data into an optimal format (e.g., WAV, MP3) and compresses it. This is done using audio file format conversion software and a data compression algorithm (e.g., G.711, Opus). The input is the audio data obtained in step 1, and the output is the compressed audio data.

[1077] Step 3:

[1078] The device sends the compressed audio data to the server using a data transfer protocol over the Internet (e.g., HTTPS). The input is the compressed audio data, and the output is the audio data transferred to the server.

[1079] Step 4:

[1080] The server receives the voice data sent from the device and converts it into text data using speech recognition technologies such as Google Cloud Speech-to-Text and Amazon Transcribe. The input is the voice data transferred to the server, and the output is text data.

[1081] Step 5:

[1082] The server analyzes the converted text data using natural language processing tools (e.g., SpaCy, BERT). The analysis step extracts contextual information, entities (e.g., people's names, place names), sentiment, and intent. The input is text data, and the output is analyzed information (contextual information, entities, sentiment, and intent).

[1083] Step 6:

[1084] The server inputs the analyzed text data into an emotion recognition engine (e.g., IBM Watson Tone Analyzer) to analyze the user's emotional state. The emotion recognition engine extracts emotions such as joy, sadness, anger, fear, and anxiety from the text data. The input is the analyzed text data, and the output is emotional data.

[1085] Step 7:

[1086] The server assesses the fraud risk based on the analyzed text data and sentiment information. The fraud detection algorithm detects specific keywords and phrases (e.g., "emergency," "transfer," "urgent") and checks whether they match patterns in a fraud database. The input is text data and sentiment data, and the output is a fraud risk assessment result.

[1087] Step 8:

[1088] If the server evaluates the fraud risk as high, it immediately sends a warning signal to the terminal. The terminal receives this warning signal and displays a warning message on the screen for the user, and if necessary, also makes an audio notification. The input is the fraud risk evaluation result, and the output is a warning message for the user.

[1089] Step 9:

[1090] If the server determines that there is a high risk of fraud, it collects relevant data (e.g., conversation content, caller phone number, conversation date and time, etc.) and automatically forwards it to the police receiving system. This information is organized in an appropriate format (e.g., CSV file, JSON data). The input is the conversation content and related data, and the output is the data to be sent to the police.

[1091] Step 10:

[1092] If the risk of fraud is determined to be high, the server sends instructions to the device to activate an AI conversation agent. The device then receives this instruction, activates the AI ​​conversation agent, and generates a natural conversation on behalf of the user. The AI ​​conversation agent uses a generative AI model to continue the dialogue with the fraudster, buying time and assisting the police in their response. The input is instructions from the server, and the output is the dialogue with the fraudster.

[1093] (Application example 2)

[1094] Next, a description will be given of Application Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal."

[1095] In modern society, telephone fraud is on the rise, posing a serious problem, especially for elderly users and those in situations where they are prone to anxiety. Conventional fraud prevention measures lack real-time detection of telephone fraud and warning functions that take into account the user's emotional state, making it difficult to prevent damage before it occurs. It is necessary for fraud detection systems to take the user's emotional state into account to achieve more accurate fraud detection and prompt warnings.

[1096] The identification processing by the identification processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means. In this invention, the server includes means for acquiring a user's telephone conversation as voice data in real time and converting the acquired voice data into text data, means for analyzing the text data to determine the possibility of fraud, means for notifying the user when the possibility of fraud is determined, means for recognizing the user's emotional state using an emotion recognition engine, and means for additionally providing the recognized emotional state to the fraud detection algorithm. This enables real-time fraud detection and warnings and responses based on the user's emotional state.

[1097] "User telephone conversation" refers to a user's verbal communication conducted over the telephone.

[1098] "Real-time" refers to processing or reaction occurring immediately, without delay.

[1099] "Audio data" refers to information that represents audio in digital form.

[1100] "Text data" refers to information obtained by converting voice data into a character string format.

[1101] "Means for converting" refers to a method or device for converting data of a particular format into another format.

[1102] "Means for analyzing to determine likelihood of fraud" refers to a method or device that analyzes input data to assess the risk of fraud.

[1103] "Means for notifying" refers to a method or device for notifying a user of specific information.

[1104] "Means of providing information to law enforcement" refers to a method or device that transmits specific information in an appropriate format to law enforcement agencies.

[1105] "Artificial intelligence agent means" refers to a method or device that uses artificial intelligence to act or interact on behalf of a user.

[1106] An "emotion recognition engine" refers to software or algorithms used to analyze and identify a user's emotional state.

[1107] "Fraud detection algorithm" refers to a computational procedure or program used to assess the risk of fraud based on specific patterns or keywords.

[1108] The present invention provides a system for automatically detecting fraud risks by monitoring users' telephone conversations in real time. Furthermore, by combining an emotion engine that recognizes the user's emotional state, it provides more advanced fraud detection and user protection capabilities. An embodiment of this system is described in detail below.

[1109] Voice Input Processing

[1110] When a user makes or receives a call, the device automatically captures the call's audio data in real time. The hardware used is a smartphone or tablet, and the software is a voice recognition library (e.g., speech_recognition). The captured audio data is converted and compressed into an optimal format by the device, and then sent to the server.

[1111] Converting audio data to text data

[1112] The server receives the voice data sent from the device and converts it into text using speech recognition technology, using a highly accurate natural language processing (NLP) engine (e.g., Google's speech recognition API).

[1113] Conversation analysis

[1114] The server analyzes the text data to extract contextual information, entities (such as names of people and places), sentiment, and intent from the conversation. This analysis process uses natural language processing techniques (e.g., NLP engines). The analyzed information is then passed to a fraud detection algorithm.

[1115] Emotion recognition by emotion engine

[1116] The server uses an emotion engine to further analyze the text data. The emotion engine recognizes the user's emotional state (e.g., joy, sadness, anger, fear, anxiety, etc.) This data is provided as additional input to the fraud detection algorithm.

[1117] Fraud Detection Algorithms

[1118] The server evaluates the likelihood of fraud based on the analyzed text data and sentiment information. Fraud detection algorithms detect specific keywords and phrases (e.g., "emergency," "transfer," "urgent") and check whether they match a database of past frauds. If potential fraud is detected, it is flagged by the server.

[1119] User notification function

[1120] If the server determines that there is a high possibility of fraud, it immediately sends a warning signal to the device. The device then displays a warning message on the screen or makes an audio notification to the user. The strength and type of the warning are adjusted according to the emotional state detected by the emotion engine.

[1121] Police contact function

[1122] If fraud is suspected, the server will collect relevant data (e.g., the content of the conversation, the caller's phone number, the date and time of the conversation, etc.) and provide it to the police promptly. This information will be automatically forwarded in an appropriate format to the police's receiving system.

[1123] AI conversational agent

[1124] If it determines that there is a high possibility of fraud, the server sends an instruction to the device to activate an AI conversation agent. The device then activates the AI ​​conversation agent based on this instruction and continues the conversation on the user's behalf. The AI ​​conversation agent generates natural conversation and continues to converse with the fraudster, buying time and supporting the police response.

[1125] Specific examples

[1126] For example, if a user answers a phone call and the caller says, "I'm your family, please send me money now," the device immediately sends the voice data to the server. The server converts the voice data into text data and analyzes it using an NLP engine to detect keywords such as "family" and "please send me money." The emotion engine also detects the user's feelings of fear or anxiety. If it determines that there is a high possibility of fraud, it immediately issues a warning to the user and notifies the police. Furthermore, an AI conversation agent is activated to continue the conversation on the user's behalf, allowing time for the police to respond.

[1127] Prompt Sentence Examples

[1128] Please provide sample code and an explanation for building a system that, when a user receives a phone call and the caller says, "I'm your family member, please send me money now," converts the voice data into text data, analyzes it using natural language processing, and determines whether it is likely to be a fraud.

[1129] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[1130] (System program processing flow)

[1131] Step 1: Audio Input Processing

[1132] The device captures the user's telephone conversation as voice data in real time. Specifically, it uses the smartphone's microphone to record the conversation during the call and temporarily stores the data. The input is the voice data captured in real time, and the output is the stored voice data.

[1133] Step 2: Converting audio data to text data

[1134] The device sends the captured voice data to the server, which uses voice recognition technology to convert the voice data into text data. Here, Google's voice recognition API is used to perform highly accurate text conversion. The input is voice data, and the output is text data.

[1135] Step 3: Analyzing the conversation

[1136] The server analyzes the text data to extract the conversation content, context information, entities, sentiment, and intent. This analysis uses natural language processing technology. The input is the text data, and the output is the analyzed context information and entity information.

[1137] Step 4: Emotion Recognition with the Emotion Engine

[1138] The server uses an emotion engine to recognize the user's emotional state from the text data. The emotion engine identifies emotions such as joy, sadness, anger, fear, and anxiety from the content and tone of the user's speech. The input is text data, and the output is emotional state data.

[1139] Step 5: Fraud Detection Algorithm

[1140] The server assesses the likelihood of fraud based on the analyzed text data and sentiment information. The fraud detection algorithm detects specific keywords and phrases and checks whether they match a historical fraud database. The input is contextual information, entity information, and sentiment data, and the output is a fraud risk assessment result.

[1141] Step 6: User Notification Function

[1142] If the server determines that there is a high possibility of fraud, it sends a warning signal to the terminal. The terminal receives this signal and displays a warning message on the screen or notifies the user by voice. The intensity and type of warning are adjusted according to the emotional state detected by the emotion engine. The input is the fraud evaluation result, and the output is a warning notification to the user.

[1143] Step 7: Contact the police

[1144] If the server determines that there is a possibility of fraud, it will collect relevant data (such as the content of the conversation, the caller's phone number, and the date and time of the conversation) and provide it to the police promptly. The information is automatically forwarded to the police's receiving system in an appropriate format. The input is the relevant data, and the output is a notification to the police.

[1145] Step 8: AI Conversational Agent

[1146] If the server determines that there is a high possibility of fraud, it sends an instruction to the terminal to activate an AI conversation agent. The terminal then activates the AI ​​conversation agent based on this instruction and continues the conversation on behalf of the user. The AI ​​conversation agent generates natural conversation and continues to converse with the fraudster, buying time and supporting police response. The input is the fraud assessment result, and the output is the activation of the AI ​​conversation agent.

[1147] The specific processing unit 290 transmits the result of the specific processing to the headset type terminal 314. In the headset type terminal 314, the control unit 46A causes the speaker 240 and the display 343 to output the result of the specific processing. The microphone 238 acquires audio indicating a user input regarding the result of the specific processing. The control unit 46A transmits audio data indicating the user input acquired by the microphone 238 to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the audio data.

[1148] The data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of the data generation model 58 is ChatGPT (Internet Search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search <url: https: gemini.google.com ?hl="ja">) and other generation AIs. The data generation model 58 is obtained by performing deep learning on a neural network. A prompt including an instruction is input to the data generation model 58, and inference data such as voice data indicating voice, text data indicating text, and image data indicating an image is also input. The data generation model 58 performs inference on the input inference data in accordance with the instruction indicated by the prompt, and outputs the inference result in a data format such as voice data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[1149] In the above embodiment, an example was given in which the specific processing is performed by the data processing device 12, but the technology of the present disclosure is not limited to this, and the specific processing may be performed by the headset type terminal 314.

[1150] [Fourth embodiment]

[1151] FIG. 7 shows an example of the configuration of a data processing system 410 according to the fourth embodiment.

[1152] 7, a data processing system 410 includes a data processing device 12 and a robot 414. An example of the data processing device 12 is a server.

[1153] The data processing device 12 includes a computer 22, a database 24, and a communication I / F 26. The computer 22 is an example of a "computer" according to the technology of the present disclosure. The computer 22 includes a processor 28, a RAM 30, and a storage 32. The processor 28, the RAM 30, and the storage 32 are connected to a bus 34. The database 24 and the communication I / F 26 are also connected to the bus 34. The communication I / F 26 is connected to a network 54. Examples of the network 54 include a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[1154] The robot 414 includes a computer 36, a microphone 238, a speaker 240, a camera 42, a communication I / F 44, and a control target 443. The computer 36 includes a processor 46, a RAM 48, and a storage 50. The processor 46, the RAM 48, and the storage 50 are connected to a bus 52. The microphone 238, the speaker 240, the camera 42, and the control target 443 are also connected to the bus 52.

[1155] The microphone 238 receives instructions and the like from the user 20 by receiving voice uttered by the user 20. The microphone 238 captures the voice uttered by the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio in accordance with instructions from the processor 46.

[1156] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an imaging element such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the surroundings of user 20 (for example, an imaging range defined by an angle of view equivalent to the field of vision of a typical healthy person).

[1157] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 are responsible for the exchange of various information between the processor 46 and the processor 28 via the network 54. The exchange of various information between the processor 46 and the processor 28 using the communication I / Fs 44 and 26 is carried out in a secure state.

[1158] The control object 443 includes a display device, LEDs in the eyes, and motors for driving the arms, hands, and feet. The posture and gestures of the robot 414 are controlled by controlling the motors of the arms, hands, and feet. Some of the emotions of the robot 414 can be expressed by controlling these motors. In addition, the facial expressions of the robot 414 can also be expressed by controlling the light emission state of the LEDs in the eyes of the robot 414.

[1159] Fig. 8 shows an example of the main functions of the data processing device 12 and the robot 414. As shown in Fig. 8, in the data processing device 12, a specific process is performed by the processor 28. A specific process program 56 is stored in the storage 32.

[1160] The specific processing program 56 is an example of a "program" according to the technology of the present disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific processing is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific processing program 56 executed on the RAM 30.

[1161] The storage 32 stores a data generation model 58 and an emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[1162] In the robot 414, the processor 46 performs the reception output process. A reception output program 60 is stored in the storage 50. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output process is realized by the processor 46 operating as the control unit 46A in accordance with the reception output program 60 executed on the RAM 48.

[1163] Next, a description will be given of the specific processing performed by the specific processing unit 290 of the data processing device 12. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal."

[1164] The present invention is a system for monitoring users' telephone conversations in real time and automatically detecting fraud risks. Detailed embodiments for specifically implementing the system are described below.

[1165] Voice Input Processing

[1166] When a user makes or receives a call, the terminal automatically captures the call's audio data in real time, which is then converted and compressed into an optimal format by the terminal and then sent to the server.

[1167] Converting audio data to text data

[1168] The server receives the voice data sent from the device and converts it into text data using speech recognition technology, which includes a highly accurate natural language processing (NLP) engine.

[1169] Conversation analysis

[1170] The server analyzes the text data and extracts contextual information, entities (such as names of people and places), sentiment, and intent from the conversation. This analysis process is performed using natural language processing techniques. The analyzed information is then passed to a fraud detection algorithm.

[1171] Fraud Detection Algorithms

[1172] The server evaluates the likelihood of fraud based on the analyzed text data. Fraud detection algorithms look for specific keywords and phrases (e.g., "urgent," "transfer," "urgent") and check whether they match previous fraud cases. If potential fraud is detected, it is flagged by the server.

[1173] User notification function

[1174] If the server determines that there is a high possibility of fraud, it immediately sends a warning signal to the device, which then displays a warning message on the screen or makes an audio notification based on the received warning signal.

[1175] Police contact function

[1176] If the server determines that there is a possibility of fraud, it will promptly provide the police with relevant data (e.g., the content of the conversation, the caller's phone number, the date and time of the conversation, etc.) This information will be automatically forwarded to the police's receiving system in an appropriate format.

[1177] AI conversational agent

[1178] If it determines that there is a high possibility of fraud, the server sends an instruction to the device to activate an AI conversation agent. The device then activates the AI ​​conversation agent based on the received instruction and continues the conversation on behalf of the user. The AI ​​conversation agent generates natural conversation and continues to converse with the fraudster, buying time and supporting police response.

[1179] Specific examples

[1180] For example, if a user answers a phone call and the caller says, "I'm your son, please send me money now," the device immediately sends the voice data to the server. The server converts the voice data into text data and analyzes it using an NLP engine. If keywords such as "son" and "please send me money" are detected, it is determined to be a fraudulent call. A warning is immediately issued to the user and the police are notified. Furthermore, an AI conversation agent is activated to continue the conversation on the user's behalf, allowing time for the police to respond.

[1181] Thus, the present invention provides a system that keeps users safe from telephone fraud and allows for fast and effective police response.

[1182] The processing flow will be explained below.

[1183] Step 1:

[1184] A user makes or receives a call. The terminal captures the voice data of this phone conversation in real time. The voice data is compressed within the terminal and prepared to be sent to the server for further processing.

[1185] Step 2:

[1186] The device transmits the captured audio data to the server, and this communication uses a secure protocol to ensure data confidentiality.

[1187] Step 3:

[1188] The server receives the received voice data and converts it into text using speech recognition technology. The speech recognition model utilizes a highly accurate natural language processing (NLP) engine.

[1189] Step 4:

[1190] The server analyzes the text data and uses an NLP engine to extract contextual information, entities (people's names, places, etc.), sentiment, and intent from the conversation. The results of this analysis are passed to a fraud detection algorithm.

[1191] Step 5:

[1192] The server runs fraud detection algorithms to evaluate potential fraud patterns. If certain keywords or phrases (e.g., "urgent," "transfer," "urgent") are detected, they are compared against a database of past frauds to determine the likelihood of a match.

[1193] Step 6:

[1194] If the server determines that there is a high possibility of fraud, it will raise a warning flag and record the possibility of fraud. This flag will be immediately notified to the user.

[1195] Step 7:

[1196] When a warning flag is raised, the server immediately sends a warning signal to the terminal, which then displays a warning message on the screen or notifies the user by voice.

[1197] Step 8:

[1198] The server collects relevant information about the flagged call (e.g., the caller's phone number, the content of the conversation, the date and time, etc.) and prepares it for automatic provision to the police. This information is then forwarded in an appropriate format to the police's receiving system.

[1199] Step 9:

[1200] If the server determines that there is a high possibility of fraud, it sends an instruction to the device to activate an AI conversation agent. Based on this instruction, the device activates the AI ​​conversation agent and continues the conversation on behalf of the user.

[1201] Step 10:

[1202] The AI ​​conversational agent generates natural conversations on behalf of the user and continues the dialogue with the fraudster, buying time and assisting law enforcement. For example, the AI ​​agent can ask, "Okay, where should I send the money?" to elicit additional information from the fraudster.

[1203] Step 11:

[1204] All conversation data is recorded by the server and stored in a format that can be analyzed later. Newly detected fraud patterns and techniques are added to the database and used as training data to improve future detection accuracy.

[1205] Through these steps, the system can protect users from telephone fraud in real time and assist police in responding quickly.

[1206] Example 1

[1207] Next, a description will be given of Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal."

[1208] Telephone scams are becoming more sophisticated every year, making vulnerable individuals such as the elderly particularly vulnerable to fraud. Current manual countermeasures often delay detection and response to scams, often failing to detect them until actual harm has occurred. Furthermore, continuing to talk to scammers is mentally stressful for users, so a fast and effective response method is needed.

[1209] The specific processing by the specific processing unit 290 of the data processing device 12 in the first embodiment is realized by the following means.

[1210] In this invention, the server includes: means for capturing a user's telephone conversation as voice data in real time; means for converting and compressing the captured voice data into an optimal format and transmitting it to the server; means for receiving the voice data and converting it into text data using voice recognition technology; means for analyzing the text data and extracting contextual information, entities, emotions, and intent; means for assessing the possibility of fraud based on the text data; means for notifying the user if a possibility of fraud is determined; means for providing information to the police if a possibility of fraud is determined; and an artificial intelligence agent for continuing the conversation on behalf of the user. This makes it possible to detect telephone fraud in real time, quickly warn the user, and contact the police if necessary, thereby preventing damage before it occurs. Furthermore, by having the artificial intelligence agent continue the conversation with the fraudster, it is possible to reduce the user's mental burden and ensure time for the police to respond.

[1211] "User" refers to a person who uses the system.

[1212] A "telephone conversation" refers to a voice communication that a user has over the telephone.

[1213] "Voice data" refers to the digital recording of a user's telephone conversation.

[1214] "Format" refers to the data structure and file format used to save audio data.

[1215] "Compression" refers to a process for reducing the data size of audio data.

[1216] "Server" refers to a computer system for processing and analyzing audio data.

[1217] "Speech recognition technology" refers to technology for analyzing voice data and converting its contents into text data.

[1218] "Text data" refers to character string data converted from audio data.

[1219] "Contextual information" refers to information about the background and situation of the conversation content.

[1220] "Entity" refers to proper nouns or specific information (e.g., people's names, place names, organization names, etc.) that appear in the conversation content.

[1221] "Emotion" refers to the speaker's emotional state (e.g., joy, anger, sadness, etc.) extracted from the conversation content.

[1222] "Intention" refers to the purpose or aim that a speaker is trying to convey through a conversation.

[1223] "Fraud Potential" refers to the result of assessing whether the content of a conversation poses a risk of fraudulent activity.

[1224] "Notification" refers to a message or signal that displays a warning or information to a user.

[1225] "Police" refers to a public security agency or its affiliates.

[1226] "Means of providing information" refers to the methods and functions for transmitting relevant information from the server to the police.

[1227] An "artificial intelligence agent" refers to an intelligent program that automatically converses on behalf of a user.

[1228] The present invention is a real-time monitoring system for protecting users from telephone fraud. This system monitors users' telephone conversations in real time and automatically detects and responds to fraud risks. Detailed embodiments for specifically implementing the present invention are described below.

[1229] Voice Input Processing

[1230] When a user makes or receives a call, the device automatically captures the call's audio data in real time, converts the captured audio data into an optimal format (e.g., from WAV to MP3), compresses it, and then sends it to the server.

[1231] Converting audio data to text data

[1232] The server receives the voice data sent from the device and converts it into text data using voice recognition technology (e.g., a high-precision voice recognition engine). Software used here includes the Google Cloud Speech-to-Text API.

[1233] Conversation analysis

[1234] The server analyzes the text data. Natural language processing techniques (e.g., SpaCy) are used to extract contextual information, entities (people's names, places, etc.), sentiment, and intent. The analyzed information is then passed to a fraud detection algorithm for evaluation.

[1235] Fraud Detection Algorithms

[1236] The server evaluates the likelihood of fraud based on the analyzed text data. Fraud detection algorithms look for specific keywords and phrases (e.g., "urgent," "transfer," "urgent") and check whether they match previous fraud cases. Potential fraud is detected and flagged.

[1237] User notification function

[1238] If the server determines that there is a high possibility of fraud, it immediately sends a warning signal to the device, which then displays a warning message on the screen or makes an audio notification based on the received warning signal.

[1239] Police contact function

[1240] If a possible fraud is identified, the server will promptly provide the police with relevant data (e.g., the content of the conversation, the caller's phone number, the date and time of the conversation, etc.) This information will be automatically forwarded to the police's receiving system in an appropriate format.

[1241] AI conversational agent

[1242] If it determines that there is a high possibility of fraud, the server sends an instruction to the device to launch an AI conversation agent. The device then launches the AI ​​conversation agent based on this instruction and continues the conversation on the user's behalf. The AI ​​conversation agent generates natural conversation and continues to converse with the fraudster, buying time and supporting police response.

[1243] Specific examples

[1244] For example, if a user answers a phone call and the caller says, "I'm your son, please send me money now," the device immediately sends the voice data to the server. The server converts the voice data into text data and analyzes it using an NLP engine. If keywords such as "son" and "please send me money" are detected, it is determined to be a fraudulent call. A warning is immediately issued to the user and the police are notified. Furthermore, an AI conversation agent is activated to continue the conversation on the user's behalf, allowing time for the police to respond.

[1245] Prompt Sentence Examples

[1246] By inputting the prompt "Please explain in detail how the system will respond if a user encounters a phone scam," the generative AI model will explain the detailed operation of each processing step. Furthermore, by using the prompt "Please tell me how to convert voice data into text and assess the risk of fraud," a detailed explanation of the speech recognition and fraud detection algorithms will be obtained.

[1247] The present invention provides a system that keeps users safe from telephone fraud and enables fast and effective police response.

[1248] The flow of the identification process in the first embodiment will be described with reference to FIG.

[1249] Program processing flow

[1250] Step 1: Getting voice input

[1251] When the device detects a user's phone call, it captures audio data in real time through the microphone. The input is the user's voice, and the output is digital audio data. The device converts this captured audio data into an optimal format (e.g., WAV format). After conversion, the data is compressed (e.g., converted to MP3 format).

[1252] Step 2: Sending audio data to the server

[1253] The device sends the compressed audio data to the server. The input is the compressed audio data obtained in step 1, and the output is the data to be sent to the server. This data is sent using a secure protocol (e.g., HTTPS).

[1254] Step 3: Converting audio data to text data

[1255] The server receives the voice data sent from the device and converts it into text data using a speech recognition engine (e.g., Google Cloud Speech-to-Text API). The input is compressed voice data, and the output is text data. Specifically, the server analyzes the voice file and outputs the spoken content as a string of characters.

[1256] Step 4: Analyzing the conversation

[1257] The server analyzes the text data. It uses natural language processing techniques (e.g., SpaCy) to extract contextual information, entities, sentiment, and intent. The input is text data, and the output is analyzed data. Specifically, the server tokenizes the text data and analyzes the meaning of each token.

[1258] Step 5: Fraud Potential Assessment

[1259] The server evaluates the likelihood of fraud based on the analysis results. It checks whether certain keywords or phrases (e.g., "urgent," "transfer," "urgent") match a database of past frauds. The input is the analyzed data, and the output is an assessment of the likelihood of fraud. Based on the assessment, if there is a high likelihood of fraud, it is flagged.

[1260] Step 6: User Notification

[1261] If it is determined that there is a high possibility of fraud, the server immediately sends a warning signal to the terminal. The input is the fraud evaluation result, and the output is the warning signal to the terminal. Based on the received warning signal, the terminal displays a warning message on the screen or issues an audio notification.

[1262] Step 7: Contact the police

[1263] If it is determined that there is a high possibility of fraud, the server will provide the police with relevant data (e.g., the content of the conversation, the caller's phone number, the date and time of the conversation, etc.). The input is the fraud assessment result and relevant data, and the output is the data to be sent to the police. The server formats the data in an appropriate format and forwards it to the police's receiving system.

[1264] Step 8: Launching the AI ​​Conversation Agent

[1265] The server sends an instruction to launch the AI ​​conversation agent to the device. The input is the fraud evaluation result, and the output is the startup instruction to the device. The device then launches the AI ​​conversation agent based on the instruction and continues the conversation on behalf of the user. Specifically, a natural conversation is conducted based on the prompt sentences generated by the AI ​​conversation agent.

[1266] Through the above processing steps, the system of the present invention monitors telephone fraud in real time and responds quickly and efficiently.

[1267] (Application example 1)

[1268] Next, a description will be given of Application Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal."

[1269] In recent years, telephone fraud has been on the rise, and countermeasures are urgently needed. However, with current systems, it takes time for users to realize they have been scammed, which often results in greater damage. Furthermore, delays in contacting authorities such as the police make it difficult to take prompt action. Furthermore, the pressure of continuing a conversation with the fraudster places a heavy psychological burden on users. Therefore, there is a need for a system that can detect telephone fraud in real time, immediately warn users, and, if necessary, contact authorities such as the police and take action on their behalf.

[1270] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 1 is realized by the following means.

[1271] In this invention, the server includes: means for acquiring a user's telephone conversation as voice data in real time; means for converting the acquired voice data into text data; means for analyzing the text data to determine the possibility of fraud; means for monitoring the voice data in real time and assessing the risk of fraud; means for determining the risk of fraud based on the detection of specific keywords and contextual information and displaying a warning; means for notifying the user when a possibility of fraud is determined; means for providing information to authorities such as the police when a possibility of fraud is determined; and artificial intelligence agent means for continuing the conversation on behalf of the user. This allows the user to know the risk of telephone fraud in real time and respond quickly. Furthermore, since authorities such as the police can be contacted quickly, a quick response can be expected. Furthermore, by having the artificial intelligence agent continue the conversation on behalf of the user, the user's mental burden can be reduced.

[1272] A "user" is someone who uses the system.

[1273] A "telephone conversation" is a dialogue conducted through voice communication.

[1274] "Voice data" refers to information that has been digitized and recorded from a user's telephone conversation.

[1275] "Text data" is voice data converted into character information.

[1276] "Analyzing text data" means using natural language processing technology to understand the content of the text data and extract specific meanings and information.

[1277] "Determining the likelihood of fraud" means that the system evaluates whether there is a risk of fraud based on the analyzed text data.

[1278] "Notifying the user" means that if the system determines that there is a possibility of fraud, it will provide the user with information to warn or caution them.

[1279] "Providing information to police or other authorities" means transmitting the content of the conversation and related information to police or other law enforcement authorities if it is determined that there is a possibility of fraud.

[1280] An "artificial intelligence agent" is a program that carries on a telephone conversation on behalf of a user and mimics natural dialogue.

[1281] "Monitoring voice data in real time" means instantly monitoring a user's telephone conversation and analyzing the voice data as needed.

[1282] "Assessing the risk of fraud" means determining the possibility of fraud based on the content of the text data.

[1283] "Detecting specific keywords and contextual information" means finding important words and context related to fraudulent activity from the analyzed text data.

[1284] "Display a warning" means that the system will display a warning message on the user's terminal if a fraud risk is detected.

[1285] The present invention relates to a system for monitoring a user's telephone conversations in real time and assessing the risk of fraud. Detailed embodiments for specifically implementing this system are described below.

[1286] System Configuration

[1287] The system mainly consists of the following components:

[1288] 1. User's device

[1289] 2. Server

[1290] 3. Artificial Intelligence Agents

[1291] User's device

[1292] The user's device is a communication device such as a smartphone, and captures the user's telephone conversation as voice data in real time. The device converts the captured voice data into an optimal format, compresses it, and sends it to the server. As a specific example, the device uses the "Google Cloud Speech-to-Text API" as voice recognition technology to convert the voice data into text data.

[1293] server

[1294] The server receives the voice data sent from the device and analyzes it in real time. The voice data is converted into text data using speech recognition technology, and then analyzed using a natural language processing (NLP) engine. The server performs the following processes based on the analyzed text data:

[1295] Detect specific keywords and contextual information to assess fraud risk.

[1296] If the risk is deemed high, a warning will be sent to the user.

[1297] Providing information to police and other authorities as necessary.

[1298] The server detects specific keywords (e.g., "urgent," "transfer," "urgent") and contextual information based on an algorithm to assess the risk of fraud. This determines whether there is a risk of fraud. Natural language processing technologies used in this process include Google Cloud Natural Language API.

[1299] Artificial Intelligence Agent

[1300] An AI agent is activated when a high risk of fraud is detected and continues the conversation on behalf of the user. The AI ​​agent mimics natural dialogue and maintains a conversation with the fraudster, freeing up police response time.

[1301] Specific examples

[1302] For example, if a user answers a phone call and the caller says, "I'm your son, please send me money now," the device immediately sends the voice data to the server. The server converts the voice data into text data and analyzes it using an NLP engine. If keywords such as "son" and "please send me money" are detected, it is determined to be a fraudulent call. A warning is immediately issued to the user, and the police are notified. An artificial intelligence agent is then activated to continue the conversation on the user's behalf.

[1303] Prompt Sentence Examples

[1304] Here is an example prompt for creating an AI model:

[1305] "Create an NLP model that converts phone voice data into text in real time and detects fraud risk."

[1306] "Convert the audio data into text data using the Google Cloud Speech-to-Text API."

[1307] As can be seen, the present invention provides a system that keeps users safe from telephone fraud and allows for a fast and effective police response.

[1308] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[1309] Step 1:

[1310] The device receives the user's phone conversation in real time, capturing voice data through a microphone and converting it into a digital format.

[1311] Input: User's phone conversation (voice data)

[1312] Data processing: capturing audio data and converting it to a digital format

[1313] Output: Digital audio data

[1314] Step 2:

[1315] The device converts the acquired audio data into the optimal format and compresses it.

[1316] Input: Digital audio data

[1317] Data processing: format conversion and compression

[1318] Output: Compressed audio data

[1319] Step 3:

[1320] The terminal transmits the compressed audio data to the server.

[1321] Input: Compressed audio data

[1322] Data calculation: Data transmission

[1323] Output: Data transfer to the server

[1324] Step 4:

[1325] The server receives the voice data sent from the device and converts it into text data using voice recognition technology, using the Google Cloud Speech-to-Text API.

[1326] Input: Compressed audio data

[1327] Data processing: speech recognition and text conversion

[1328] Output: Text data

[1329] Step 5:

[1330] The server analyzes the text data using a natural language processing (NLP) engine, using the Google Cloud Natural Language API to extract specific keywords and contextual information from the text data.

[1331] Input: Text data

[1332] Data processing: Information extraction through natural language analysis

[1333] Output: Analysis results (specific keywords and contextual information)

[1334] Step 6:

[1335] The server evaluates the fraud risk based on the analysis results, checking whether specific keywords (e.g., "emergency," "transfer," "urgent") are included to determine the fraud risk.

[1336] Input: Analysis results

[1337] Data Computing: Applying Fraud Risk Assessment Algorithms

[1338] Output: Fraud risk rating (high, medium, low)

[1339] Step 7:

[1340] If the server determines that the risk of fraud is high, it will send a warning to the user's device.

[1341] Input: Fraud Risk Rating (High)

[1342] Data Calculation: Generating and Sending Warning Messages

[1343] Output: Warning notification to the user terminal

[1344] Step 8:

[1345] The user's device will display the received warning on the screen and also alert them with audio notifications and vibrations.

[1346] Input: Warning notification from the server

[1347] Data processing: Display warning message

[1348] Output: User attention (screen display and audio notification)

[1349] Step 9:

[1350] If the server determines there is a high risk of fraud, it will provide relevant information to authorities such as the police, sending data including the content of the conversation and the caller's number.

[1351] Input: Fraud Risk Rating (High)

[1352] Data Computing: Generating and transmitting relevant information

[1353] Output: Providing information to authorities such as police

[1354] Step 10:

[1355] If the server determines that the fraud risk is high, it sends an instruction to the terminal to activate an AI agent, which continues the conversation on behalf of the user.

[1356] Input: Fraud Risk Rating (High)

[1357] Data Calculation: Instructions for launching artificial intelligence agents

[1358] Output: Launch of artificial intelligence agent

[1359] Step 11:

[1360] The artificial intelligence agent mimics natural dialogue and continues the conversation on behalf of the user, maintaining a conversation with the fraudster to free up police response time.

[1361] Input: Conversational content and user instructions

[1362] Data processing: Natural conversation generation

[1363] Output: Ongoing dialogue with the fraudster

[1364] Furthermore, an emotion engine that estimates the user's emotion may be further combined. That is, the identification processing unit 290 may estimate the user's emotion using the emotion identification model 59, and perform identification processing using the user's emotion.

[1365] The present invention is a system that monitors users' telephone conversations in real time and automatically detects fraud risks. Furthermore, by combining it with an emotion engine that recognizes the user's emotional state, it provides more advanced fraud detection and user protection functions. Detailed embodiments for specifically implementing this system are described below.

[1366] Voice Input Processing

[1367] When a user makes or receives a call, the terminal automatically captures the call's audio data in real time, which is then converted and compressed into an optimal format by the terminal and then sent to the server.

[1368] Converting audio data to text data

[1369] The server receives the voice data sent from the device and converts it into text data using voice recognition technology, which includes a highly accurate natural language processing (NLP) engine.

[1370] Conversation analysis

[1371] The server analyzes the text data and extracts contextual information, entities (such as names of people and places), sentiment, and intent from the conversation. This analysis process is performed using natural language processing techniques. The analyzed information is then passed to a fraud detection algorithm.

[1372] Emotion recognition by emotion engine

[1373] The server uses an emotion engine to further analyze the text data, which recognizes the user's emotional state (e.g., joy, sadness, anger, fear, anxiety, etc.) This data is provided as additional input to the fraud detection algorithm.

[1374] Fraud Detection Algorithms

[1375] The server evaluates the likelihood of fraud based on the analyzed text data and sentiment information. Fraud detection algorithms detect specific keywords and phrases (e.g., "emergency," "transfer," "urgent") and check whether they match a database of past frauds. If potential fraud is detected, it is flagged by the server.

[1376] User notification function

[1377] If the server determines that there is a high possibility of fraud, it immediately sends a warning signal to the device. The device then displays a warning message on the screen or makes an audio notification to the user. The strength and type of the warning are adjusted according to the emotional state detected by the emotion engine.

[1378] Police contact function

[1379] If the server determines that there is a possibility of fraud, it will collect relevant data (e.g., the content of the conversation, the caller's phone number, the date and time of the conversation, etc.) and provide it to the police promptly. This information will be automatically forwarded in an appropriate format to the police's receiving system.

[1380] AI conversational agent

[1381] If it determines that there is a high possibility of fraud, the server sends an instruction to the device to activate an AI conversation agent. The device then activates the AI ​​conversation agent based on this instruction and continues the conversation on the user's behalf. The AI ​​conversation agent generates natural conversation and continues to converse with the fraudster, buying time and supporting the police response.

[1382] Specific examples

[1383] For example, if a user answers a phone call and the caller says, "I'm your son, please send me money now," the device immediately sends the voice data to the server. The server converts the voice data into text data and analyzes it using an NLP engine to detect keywords such as "son" and "please send me money." The emotion engine also detects the user's feelings of fear or anxiety. If it determines that there is a high possibility of fraud, it immediately issues a warning to the user and notifies the police. Furthermore, an AI conversation agent is activated to continue the conversation on the user's behalf, allowing time for the police to respond.

[1384] As described above, the present invention is a system that protects users from telephone fraud in real time and further achieves more accurate fraud detection and warning by taking into account the user's emotional state.

[1385] The processing flow will be explained below.

[1386] Step 1:

[1387] A user makes or receives a call. The terminal captures the voice data of this call in real time. The captured voice data is compressed within the terminal and prepared for transmission to the server.

[1388] Step 2:

[1389] The device transmits the captured audio data to the server, and this communication uses a secure protocol to ensure data confidentiality.

[1390] Step 3:

[1391] The server processes the received voice data and converts it into text using speech recognition technology, which includes a highly accurate natural language processing (NLP) engine.

[1392] Step 4:

[1393] The server analyzes the text data, using an NLP engine to extract contextual information, entities (such as names of people or places), sentiment, and intent from the conversation, and passes the results of this analysis to a fraud detection algorithm.

[1394] Step 5:

[1395] The server then passes the text data to an emotion engine to evaluate the user's emotional state, which reads emotions from the user's tone of voice and word choice and adds them to the analysis results.

[1396] Step 6:

[1397] The server runs a fraud detection algorithm to assess the likelihood of fraud based on the analyzed text data and sentiment information. If certain keywords or phrases (e.g., "urgent," "transfer," "urgent") are detected, they are compared with a database of past frauds to determine the degree of match.

[1398] Step 7:

[1399] If the server determines that there is a high possibility of fraud, it will raise a warning flag and record the possibility of fraud. This flag will be immediately notified to the user.

[1400] Step 8:

[1401] When a warning flag is raised, the server immediately sends a warning signal to the device. The device receives this signal and displays a warning message on the screen or issues an audio notification to the user. The strength and type of the warning are adjusted according to the emotional state detected by the emotion engine.

[1402] Step 9:

[1403] If the server determines that there is a high possibility of fraud, it will collect detailed relevant information (e.g., the content of the conversation, the caller's phone number, the date and time of the conversation, etc.) and provide it to the police promptly. This information will be automatically forwarded in an appropriate format to the police's receiving system.

[1404] Step 10:

[1405] If the server determines that there is a high possibility of fraud, it sends an instruction to the device to activate an AI conversation agent. Based on this instruction, the device activates the AI ​​conversation agent and continues the conversation on behalf of the user.

[1406] Step 11:

[1407] The AI ​​conversational agent generates natural conversations on behalf of the user and continues the dialogue with the fraudster. For example, the AI ​​agent can ask, "Okay, where should I send the money?" to elicit additional information from the fraudster, buying time and assisting police in their response.

[1408] Step 12:

[1409] The server records all conversation data and stores it in a format that can be analyzed later. Newly detected fraud patterns and techniques are added to the database and used as training data to improve future detection accuracy.

[1410] Through these steps, the system can protect users from phone fraud in real time and assist police in responding quickly. Furthermore, by taking into account the user's emotional state, the accuracy of fraud detection and warning can be further improved.

[1411] Example 2

[1412] Next, a description will be given of Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal."

[1413] In recent years, telephone fraud has been on the rise, with many people falling victim to it. Furthermore, fraudulent activities are becoming more complex and sophisticated, making it difficult to detect and respond to them in real time using conventional methods. Furthermore, the victim's emotional state can be an important factor in conversations with fraudsters, but conventional systems do not take this into account. This creates a need for rapid and accurate fraud detection and response.

[1414] The identification process by the identification processing unit 290 of the data processing device 12 in Example 2 is realized by the following means. In this invention, the server includes means for acquiring a user's telephone conversation as voice data in real time, means for converting the acquired voice data into text data, means for analyzing the text data to determine the possibility of fraud, means for notifying the user when the possibility of fraud is determined, means for providing information to relevant authorities when the possibility of fraud is determined, an artificial intelligence agent means for continuing the conversation on behalf of the user, means for analyzing the user's emotional state, and means for reevaluating the possibility of fraud based on the emotional state. This makes it possible to quickly and accurately detect fraudulent acts and take appropriate measures taking into account the user's emotional state.

[1415] "User" means any person or entity that uses the System to conduct telephone conversations.

[1416] A "telephone conversation" refers to a voice-based communication that a user has with another person via telephone.

[1417] "Voice data" refers to voice information captured in real time as a digital representation of a user's telephone conversation.

[1418] "Text data" refers to text information converted from voice data using voice recognition technology.

[1419] "Analysis" refers to the process of extracting the content and meaning of text or audio data and converting it into an understandable format.

[1420] "Possibility of fraud" refers to the result of assessing the degree of likelihood of fraud based on the acquired conversation content.

[1421] "Notification" refers to the act of the system displaying or audibly conveying warnings or information to the user.

[1422] "Relevant authorities" refers to organisations and bodies that are provided with information to deal with fraudulent activity, including the police.

[1423] An "artificial intelligence agent" refers to a sophisticated automated response program that can continue a dialogue on behalf of a user.

[1424] "Emotional state" refers to a user's mental or emotional response or state (e.g., joy, sadness, anger, fear, anxiety, etc.).

[1425] "Reappraisal" refers to the process of reviewing and updating the initial appraisal results based on the acquired emotional state.

[1426] The present invention is a system that monitors users' telephone conversations in real time and automatically detects fraud risks. Furthermore, by combining it with an emotion engine that recognizes the user's emotional state, it provides more advanced fraud detection and user protection functions.

[1427] Voice Input Processing

[1428] The device automatically captures the call audio data when the user makes or receives a call. The audio data is collected in real time and temporarily stored in the internal memory. This process can be performed using a smartphone or a dedicated device.

[1429] Audio data transmission and format conversion

[1430] The device converts the captured audio data into the optimal format (e.g., WAV, MP3) and compresses it using commonly used data compression algorithms (e.g., G.711, Opus). After conversion, the audio data is sent from the device to the server.

[1431] Converting audio data to text

[1432] The server receives the voice data sent from the device and converts it into text data using speech recognition technologies such as Google Cloud Speech-to-Text and Amazon Transcribe. This process utilizes a highly accurate natural language processing (NLP) engine.

[1433] Text data analysis

[1434] The server then analyzes the converted text data using natural language processing tools such as SpaCy and BERT. During the analysis step, contextual information, entities (e.g., people's names, place names), sentiment, and intent are extracted.

[1435] Performing emotion recognition

[1436] The server uses an emotion recognition engine such as IBM Watson Tone Analyzer to analyze the user's emotional state (happiness, sadness, anger, fear, anxiety, etc.) from the text data. This emotional information is used as additional input for subsequent fraud risk assessment.

[1437] Fraud risk assessment

[1438] The server assesses the fraud risk based on the analyzed text data and sentiment information. The fraud detection algorithm looks for specific keywords and phrases (e.g., "emergency," "transfer," "urgent") and matches them with patterns in a fraud database. If a high fraud risk is detected, the server flags the system.

[1439] User warning notification

[1440] If the server determines that there is a high risk of fraud, it immediately sends a warning signal to the terminal. The terminal then receives this warning signal and displays a warning message on the screen and also makes an audio notification to the user. For example, a message such as "There is a possibility of fraud, please be careful" may be displayed.

[1441] Data sharing with the police

[1442] If the server determines that there is a possibility of fraud, it collects relevant data such as the content of the conversation, the originating phone number, the date and time of the conversation, etc. This data is organized in an appropriate format (e.g., CSV file, JSON data) and automatically forwarded to the receiving system of the relevant agency, such as the police.

[1443] Launching an AI conversation agent

[1444] If the risk of fraud is deemed high, the server sends instructions to the device to activate an AI conversation agent. The device then receives this instruction, activates the AI ​​conversation agent, and generates natural conversation on behalf of the user. The AI ​​conversation agent uses the generative AI model to continue dialogue with the fraudster, buying time and assisting the police in their response.

[1445] Specific examples

[1446] For example, if a user answers a phone call and says, "I'm your son, please send me money now," the device immediately sends the voice data to the server. The server then converts the voice data into text using Google Cloud Speech-to-Text and uses SpaCy to extract keywords such as "son" and "please send me money." The emotion engine then analyzes the user's fear and anxiety to detect a high risk of fraud. The server then immediately sends a warning signal to the device, displays a warning message to the user, and notifies the police. Furthermore, an AI conversation agent continues the conversation on behalf of the user, allowing time for the police to respond.

[1447] Prompt Sentence Examples

[1448] "If a user is having a potentially fraudulent conversation on the phone, capture the audio data in real time and assess the likelihood of it being fraudulent. If it is, create a system that alerts the user and activates an AI conversation agent to continue the conversation on the user's behalf."

[1449] The flow of the identification process in the second embodiment will be described with reference to FIG.

[1450] Step 1:

[1451] When a user makes or receives a call, the device automatically captures the call's audio data. Specifically, the device uses the microphone on the smartphone or dedicated device to collect audio data in real time and temporarily stores it in its internal memory. The input is the user's voice during the call, and the output is digital audio data.

[1452] Step 2:

[1453] The device converts the captured audio data into an optimal format (e.g., WAV, MP3) and compresses it. This is done using audio file format conversion software and a data compression algorithm (e.g., G.711, Opus). The input is the audio data obtained in step 1, and the output is the compressed audio data.

[1454] Step 3:

[1455] The device sends the compressed audio data to the server using a data transfer protocol over the Internet (e.g., HTTPS). The input is the compressed audio data, and the output is the audio data transferred to the server.

[1456] Step 4:

[1457] The server receives the voice data sent from the device and converts it into text data using speech recognition technologies such as Google Cloud Speech-to-Text and Amazon Transcribe. The input is the voice data transferred to the server, and the output is text data.

[1458] Step 5:

[1459] The server analyzes the converted text data using natural language processing tools (e.g., SpaCy, BERT). The analysis step extracts contextual information, entities (e.g., people's names, place names), sentiment, and intent. The input is text data, and the output is analyzed information (contextual information, entities, sentiment, and intent).

[1460] Step 6:

[1461] The server inputs the analyzed text data into an emotion recognition engine (e.g., IBM Watson Tone Analyzer) to analyze the user's emotional state. The emotion recognition engine extracts emotions such as joy, sadness, anger, fear, and anxiety from the text data. The input is the analyzed text data, and the output is emotional data.

[1462] Step 7:

[1463] The server assesses the fraud risk based on the analyzed text data and sentiment information. The fraud detection algorithm detects specific keywords and phrases (e.g., "emergency," "transfer," "urgent") and checks whether they match patterns in a fraud database. The input is text data and sentiment data, and the output is a fraud risk assessment result.

[1464] Step 8:

[1465] If the server evaluates the fraud risk as high, it immediately sends a warning signal to the terminal. The terminal receives this warning signal and displays a warning message on the screen for the user, and if necessary, also makes an audio notification. The input is the fraud risk evaluation result, and the output is a warning message for the user.

[1466] Step 9:

[1467] If the server determines that there is a high risk of fraud, it collects relevant data (e.g., conversation content, caller phone number, conversation date and time, etc.) and automatically forwards it to the police receiving system. This information is organized in an appropriate format (e.g., CSV file, JSON data). The input is the conversation content and related data, and the output is the data to be sent to the police.

[1468] Step 10:

[1469] If the risk of fraud is determined to be high, the server sends instructions to the device to activate an AI conversation agent. The device then receives this instruction, activates the AI ​​conversation agent, and generates a natural conversation on behalf of the user. The AI ​​conversation agent uses a generative AI model to continue the dialogue with the fraudster, buying time and assisting the police in their response. The input is instructions from the server, and the output is the dialogue with the fraudster.

[1470] (Application example 2)

[1471] Next, a description will be given of Application Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal."

[1472] In modern society, telephone fraud is on the rise, posing a serious problem, especially for elderly users and those in situations where they are prone to anxiety. Conventional fraud prevention measures lack real-time detection of telephone fraud and warning functions that take into account the user's emotional state, making it difficult to prevent damage before it occurs. It is necessary for fraud detection systems to take the user's emotional state into account to achieve more accurate fraud detection and prompt warnings.

[1473] The identification processing by the identification processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means. In this invention, the server includes means for acquiring a user's telephone conversation as voice data in real time and converting the acquired voice data into text data, means for analyzing the text data to determine the possibility of fraud, means for notifying the user when the possibility of fraud is determined, means for recognizing the user's emotional state using an emotion recognition engine, and means for additionally providing the recognized emotional state to the fraud detection algorithm. This enables real-time fraud detection and warnings and responses based on the user's emotional state.

[1474] "User telephone conversation" refers to a user's verbal communication conducted over the telephone.

[1475] "Real-time" refers to processing or reaction occurring immediately, without delay.

[1476] "Audio data" refers to information that represents audio in digital form.

[1477] "Text data" refers to information obtained by converting voice data into a character string format.

[1478] "Means for converting" refers to a method or device for converting data of a particular format into another format.

[1479] "Means for analyzing to determine likelihood of fraud" refers to a method or device that analyzes input data to assess the risk of fraud.

[1480] "Means for notifying" refers to a method or device for notifying a user of specific information.

[1481] "Means of providing information to law enforcement" refers to a method or device that transmits specific information in an appropriate format to law enforcement agencies.

[1482] "Artificial intelligence agent means" refers to a method or device that uses artificial intelligence to act or interact on behalf of a user.

[1483] An "emotion recognition engine" refers to software or algorithms used to analyze and identify a user's emotional state.

[1484] "Fraud detection algorithm" refers to a computational procedure or program used to assess the risk of fraud based on specific patterns or keywords.

[1485] The present invention provides a system for automatically detecting fraud risks by monitoring users' telephone conversations in real time. Furthermore, by combining an emotion engine that recognizes the user's emotional state, it provides more advanced fraud detection and user protection capabilities. An embodiment of this system is described in detail below.

[1486] Voice Input Processing

[1487] When a user makes or receives a call, the device automatically captures the call's audio data in real time. The hardware used is a smartphone or tablet, and the software is a voice recognition library (e.g., speech_recognition). The captured audio data is converted and compressed into an optimal format by the device, and then sent to the server.

[1488] Converting audio data to text data

[1489] The server receives the voice data sent from the device and converts it into text using speech recognition technology, using a highly accurate natural language processing (NLP) engine (e.g., Google's speech recognition API).

[1490] Conversation analysis

[1491] The server analyzes the text data to extract contextual information, entities (such as names of people and places), sentiment, and intent from the conversation. This analysis process uses natural language processing techniques (e.g., NLP engines). The analyzed information is then passed to a fraud detection algorithm.

[1492] Emotion recognition by emotion engine

[1493] The server uses an emotion engine to further analyze the text data. The emotion engine recognizes the user's emotional state (e.g., joy, sadness, anger, fear, anxiety, etc.) This data is provided as additional input to the fraud detection algorithm.

[1494] Fraud Detection Algorithms

[1495] The server evaluates the likelihood of fraud based on the analyzed text data and sentiment information. Fraud detection algorithms detect specific keywords and phrases (e.g., "emergency," "transfer," "urgent") and check whether they match a database of past frauds. If potential fraud is detected, it is flagged by the server.

[1496] User notification function

[1497] If the server determines that there is a high possibility of fraud, it immediately sends a warning signal to the device. The device then displays a warning message on the screen or makes an audio notification to the user. The strength and type of the warning are adjusted according to the emotional state detected by the emotion engine.

[1498] Police contact function

[1499] If fraud is suspected, the server will collect relevant data (e.g., the content of the conversation, the caller's phone number, the date and time of the conversation, etc.) and provide it to the police promptly. This information will be automatically forwarded in an appropriate format to the police's receiving system.

[1500] AI conversational agent

[1501] If it determines that there is a high possibility of fraud, the server sends an instruction to the device to activate an AI conversation agent. The device then activates the AI ​​conversation agent based on this instruction and continues the conversation on the user's behalf. The AI ​​conversation agent generates natural conversation and continues to converse with the fraudster, buying time and supporting the police response.

[1502] Specific examples

[1503] For example, if a user answers a phone call and the caller says, "I'm your family, please send me money now," the device immediately sends the voice data to the server. The server converts the voice data into text data and analyzes it using an NLP engine to detect keywords such as "family" and "please send me money." The emotion engine also detects the user's feelings of fear or anxiety. If it determines that there is a high possibility of fraud, it immediately issues a warning to the user and notifies the police. Furthermore, an AI conversation agent is activated to continue the conversation on the user's behalf, allowing time for the police to respond.

[1504] Prompt Sentence Examples

[1505] Please provide sample code and an explanation for building a system that, when a user receives a phone call and the caller says, "I'm your family member, please send me money now," converts the voice data into text data, analyzes it using natural language processing, and determines whether it is likely to be a fraud.

[1506] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[1507] (System program processing flow)

[1508] Step 1: Audio Input Processing

[1509] The device captures the user's telephone conversation as voice data in real time. Specifically, it uses the smartphone's microphone to record the conversation during the call and temporarily stores the data. The input is the voice data captured in real time, and the output is the stored voice data.

[1510] Step 2: Converting audio data to text data

[1511] The device sends the captured voice data to the server, which uses voice recognition technology to convert the voice data into text data. Here, Google's voice recognition API is used to perform highly accurate text conversion. The input is voice data, and the output is text data.

[1512] Step 3: Analyzing the conversation

[1513] The server analyzes the text data to extract the conversation content, context information, entities, sentiment, and intent. This analysis uses natural language processing technology. The input is the text data, and the output is the analyzed context information and entity information.

[1514] Step 4: Emotion Recognition with the Emotion Engine

[1515] The server uses an emotion engine to recognize the user's emotional state from the text data. The emotion engine identifies emotions such as joy, sadness, anger, fear, and anxiety from the content and tone of the user's speech. The input is text data, and the output is emotional state data.

[1516] Step 5: Fraud Detection Algorithm

[1517] The server assesses the likelihood of fraud based on the analyzed text data and sentiment information. The fraud detection algorithm detects specific keywords and phrases and checks whether they match a historical fraud database. The input is contextual information, entity information, and sentiment data, and the output is a fraud risk assessment result.

[1518] Step 6: User Notification Function

[1519] If the server determines that there is a high possibility of fraud, it sends a warning signal to the terminal. The terminal receives this signal and displays a warning message on the screen or notifies the user by voice. The intensity and type of warning are adjusted according to the emotional state detected by the emotion engine. The input is the fraud evaluation result, and the output is a warning notification to the user.

[1520] Step 7: Contact the police

[1521] If the server determines that there is a possibility of fraud, it will collect relevant data (such as the content of the conversation, the caller's phone number, and the date and time of the conversation) and provide it to the police promptly. The information is automatically forwarded to the police's receiving system in an appropriate format. The input is the relevant data, and the output is a notification to the police.

[1522] Step 8: AI Conversational Agent

[1523] If the server determines that there is a high possibility of fraud, it sends an instruction to the terminal to activate an AI conversation agent. The terminal then activates the AI ​​conversation agent based on this instruction and continues the conversation on behalf of the user. The AI ​​conversation agent generates natural conversation and continues to converse with the fraudster, buying time and supporting police response. The input is the fraud assessment result, and the output is the activation of the AI ​​conversation agent.

[1524] The specific processing unit 290 transmits the result of the specific processing to the robot 414. In the robot 414, the control unit 46A causes the speaker 240 and the control target 443 to output the result of the specific processing. The microphone 238 acquires voice indicating a user input regarding the result of the specific processing. The control unit 46A transmits voice data indicating the user input acquired by the microphone 238 to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the voice data.

[1525] The data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of the data generation model 58 is ChatGPT (Internet Search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search <url: https: gemini.google.com ?hl="ja">) and other generation AIs. The data generation model 58 is obtained by performing deep learning on a neural network. A prompt including an instruction is input to the data generation model 58, and inference data such as voice data indicating voice, text data indicating text, and image data indicating an image is also input. The data generation model 58 performs inference on the input inference data in accordance with the instruction indicated by the prompt, and outputs the inference result in a data format such as voice data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[1526] In the above embodiment, an example was given in which the specific processing is performed by the data processing device 12, but the technology of the present disclosure is not limited to this, and the specific processing may be performed by the robot 414.

[1527] The emotion identification model 59 as an emotion engine may determine the user's emotion according to a specific mapping. Specifically, the emotion identification model 59 may determine the user's emotion according to an emotion map (see FIG. 9), which is a specific mapping. Similarly, the emotion identification model 59 may determine the robot's emotion, and the identification processing unit 290 may perform identification processing using the robot's emotion.

[1528] FIG. 9 is a diagram illustrating an emotion map 400 on which multiple emotions are mapped. In the emotion map 400, emotions are arranged in concentric circles radiating from the center. Emotions closer to the center of the concentric circles are more primitive. Emotions representing states and actions arising from a state of mind are arranged on the outer edges of the concentric circles. The concept of emotion includes both affect and mental states. Emotions generally generated from reactions occurring in the brain are arranged on the left side of the concentric circles. Emotions generally induced by situational judgment are arranged on the right side of the concentric circles. Emotions generally generated from reactions occurring in the brain and induced by situational judgment are arranged on the upper and lower sides of the concentric circles. Furthermore, the emotion of "pleasure" is arranged on the upper side of the concentric circles, and the emotion of "discomfort" is arranged on the lower side. In this way, in the emotion map 400, multiple emotions are mapped based on the structure by which emotions are generated, and emotions that tend to occur simultaneously are mapped close to each other.

[1529] These emotions are distributed in the 3 o'clock direction on emotion map 400, and typically fluctuate between relief and anxiety. In the right half of emotion map 400, situational awareness dominates over internal sensations, resulting in a sense of calm.

[1530] The inside of emotion map 400 represents what is going on in the mind, and the outside of emotion map 400 represents behavior, so the further you go outside emotion map 400, the more visible the emotions become (the more they are expressed in behavior).

[1531] Human emotions are based on various balances, such as posture and blood sugar levels. When these balances deviate from the ideal, a state of discomfort is indicated, and when they approach the ideal, a state of pleasure is indicated. Emotions can also be created for robots, automobiles, and motorcycles, based on various balances, such as posture and remaining battery life. When these balances deviate from the ideal, a state of discomfort is indicated, and when they approach the ideal, a state of pleasure is indicated. An emotion map can be generated, for example, based on Dr. Mitsuyoshi's emotion map (Research on Voice Emotion Recognition and Emotional Brain Physiological Signal Analysis Systems, Tokushima University, Doctoral Dissertation: https: / / ci.nii.ac.jp / naid / 500000375379). The left half of the emotion map lists emotions belonging to the "reaction" domain, where sensation is dominant. The right half of the emotion map lists emotions belonging to the "situation" domain, where situational awareness is dominant.

[1532] The emotion map defines two emotions that promote learning. One is a negative emotion on the situation side, around the middle of "repentance" or "reflection." In other words, this occurs when the robot experiences negative emotions such as "I never want to feel this way again" or "I don't want to be scolded again." The other is a positive emotion on the response side, around "desire." In other words, this occurs when the robot experiences positive feelings such as "I want more" or "I want to know more."

[1533] The emotion identification model 59 inputs user input into a pre-trained neural network, obtains emotion values ​​indicating each emotion shown in the emotion map 400, and determines the user's emotion. This neural network is pre-trained based on multiple pieces of training data that are combinations of user input and emotion values ​​indicating each emotion shown in the emotion map 400. Furthermore, this neural network is trained so that emotions that are located close to each other have similar values, as in the emotion map 900 shown in FIG. 10. FIG. 10 shows an example in which multiple emotions, "relieved," "calm," and "reassuring," have similar emotion values.

[1534] The system according to the present disclosure has been described above mainly with respect to the functions of the data processing device 12, but the system according to the present disclosure is not necessarily implemented on a server. The system according to the present disclosure may be implemented as a general information processing system. The present disclosure may be implemented, for example, as a software program running on a personal computer or an application running on a smartphone, etc. The method according to the present disclosure may be provided to users in the form of SaaS (Software as a Service).

[1535] In the above embodiment, an example was given in which the specific processing is performed by one computer 22, but the technology of the present disclosure is not limited to this, and the specific processing may be distributed and performed by a plurality of computers including the computer 22. For example, the data generation model 58 may be provided in an external device of the data processing device 12, and data may be generated in the external device in accordance with input data.

[1536] In the above embodiment, an example in which the specific processing program 56 is stored in the storage 32 has been described, but the technology of the present disclosure is not limited to this. For example, the specific processing program 56 may be stored in a portable, computer-readable, non-transitory storage medium such as a USB (Universal Serial Bus) memory. The specific processing program 56 stored in the non-transitory storage medium is installed in the computer 22 of the data processing device 12. The processor 28 executes the specific processing in accordance with the specific processing program 56.

[1537] Alternatively, the specific processing program 56 may be stored in a storage device such as a server connected to the data processing device 12 via the network 54, and the specific processing program 56 may be downloaded and installed on the computer 22 in response to a request from the data processing device 12.

[1538] It is not necessary to store all of the specific processing program 56 in a storage device such as a server connected to the data processing device 12 via the network 54, or to store all of the specific processing program 56 in the storage 32; only a portion of the specific processing program 56 may be stored.

[1539] The hardware resource for executing a specific process can be any of the following processors: An example of a processor is a CPU, which is a general-purpose processor that functions as a hardware resource for executing a specific process by executing software, i.e., a program. Another example of a processor is a dedicated electrical circuit, such as an FPGA (Field-Programmable Gate Array), a PLD (Programmable Logic Device), or an ASIC (Application Specific Integrated Circuit), which is a processor with a circuit configuration designed specifically for executing a specific process. Each processor has built-in or connected memory, and each processor uses the memory to execute the specific process.

[1540] The hardware resource that executes the specific processing may be configured with one of these various processors, or may be configured with a combination of two or more processors of the same or different types (for example, a combination of multiple FPGAs, or a combination of a CPU and an FPGA). Also, the hardware resource that executes the specific processing may be a single processor.

[1541] As an example of a system configured with a single processor, first, one processor is configured by combining one or more CPUs and software, and this processor functions as a hardware resource that executes a specific process. Second, there is a system that uses a processor that realizes the functions of an entire system including multiple hardware resources that execute a specific process on a single IC chip, as typified by SoC (System-on-a-chip). In this way, a specific process is realized using one or more of the above-mentioned various processors as hardware resources.

[1542] Furthermore, the hardware structure of these various processors can be, more specifically, an electric circuit that combines circuit elements such as semiconductor devices. The specific processing described above is merely an example. Therefore, it goes without saying that unnecessary steps may be deleted, new steps may be added, or the processing order may be rearranged, without departing from the spirit of the invention.

[1543] The above-described description and illustrations are a detailed explanation of the parts related to the technology of the present disclosure and are merely an example of the technology of the present disclosure. For example, the above description of the configuration, functions, actions, and effects is an explanation of an example of the configuration, functions, actions, and effects of the parts related to the technology of the present disclosure. Therefore, it goes without saying that unnecessary parts may be deleted, new elements may be added, or replacements may be made to the above-described description and illustrations within the scope of the gist of the technology of the present disclosure. Furthermore, to avoid confusion and facilitate understanding of the parts related to the technology of the present disclosure, the above-described description and illustrations omit explanations of common technical knowledge that do not require particular explanation to enable the implementation of the technology of the present disclosure.

[1544] All publications, patent applications, and technical standards mentioned in this specification are herein incorporated by reference to the same extent as if each individual publication, patent application, or technical standard was specifically and individually indicated to be incorporated by reference.

[1545] The following is further disclosed regarding the above embodiment.

[1546] (Claim 1)

[1547] Capture users' telephone conversations as voice data in real time,

[1548] A means for converting the acquired voice data into text data;

[1549] A means of analyzing text data to determine the possibility of fraud;

[1550] a means for notifying the user if a potential fraud is determined;

[1551] a means of providing information to the police if fraud is determined to be a possibility;

[1552] an artificial intelligence agent means for continuing a conversation on behalf of the user;

[1553] A system including:

[1554] (Claim 2)

[1555] 10. The system of claim 1, comprising means for using natural language processing techniques in determining likelihood of fraud.

[1556] (Claim 3)

[1557] The system of claim 1, further comprising means for detecting specific keywords and contextual information based on the results of the analysis of the text data to determine the likelihood of fraud.

[1558] "Example 1"

[1559] (Claim 1)

[1560] means for acquiring a user's telephone conversation as voice data in real time;

[1561] means for converting and compressing the acquired audio data into an optimal format and transmitting the data to a server;

[1562] means for receiving voice data and converting it into text data using voice recognition technology;

[1563] a means for analyzing the text data and extracting contextual information, entities, sentiment, and intent;

[1564] a means for assessing likelihood of fraud based on the text data;

[1565] a means for notifying the user if a potential fraud is determined;

[1566] a means of providing information to the police if fraud is determined to be a possibility;

[1567] an artificial intelligence agent means for continuing a conversation on behalf of the user;

[1568] A system including:

[1569] (Claim 2)

[1570] 10. The system of claim 1, comprising means for using natural language processing techniques in determining likelihood of fraud.

[1571] (Claim 3)

[1572] The system of claim 1, further comprising means for detecting specific keywords and contextual information based on the results of the analysis of the text data to determine the likelihood of fraud.

[1573] "Application Example 1"

[1574] (Claim 1)

[1575] Capture users' telephone conversations as voice data in real time,

[1576] A means for converting the acquired voice data into text data;

[1577] A means of analyzing text data to determine the possibility of fraud;

[1578] a means for notifying the user if a potential fraud is determined;

[1579] A means of providing information to police and other authorities if fraud is determined to be a possibility; and

[1580] an artificial intelligence agent means for continuing a conversation on behalf of the user;

[1581] A means to monitor voice data in real time and assess fraud risk; and

[1582] A means of determining fraud risk and displaying a warning based on the detection of specific keywords and contextual information;

[1583] A system including:

[1584] (Claim 2)

[1585] 10. The system of claim 1, further comprising means for analyzing the text data using natural language processing techniques to determine likelihood of fraud.

[1586] (Claim 3)

[1587] 2. The system of claim 1, further comprising means for detecting specific keywords and contextual information based on the results of the analysis of the text data to assess fraud risk and display a warning to the user in real time.

[1588] "Example 2: Combining Emotion Engines"

[1589] (Claim 1)

[1590] means for acquiring a user's telephone conversation as voice data in real time;

[1591] A means for converting the acquired voice data into text data;

[1592] A means of analyzing text data to determine the possibility of fraud;

[1593] a means for notifying the user if a potential fraud is determined;

[1594] a means of providing information to relevant authorities when potential fraud is identified; and

[1595] an artificial intelligence agent means for continuing a conversation on behalf of the user;

[1596] means for analyzing the emotional state of a user;

[1597] a means of reassessing the likelihood of fraud based on emotional state;

[1598] A system including:

[1599] (Claim 2)

[1600] 10. The system of claim 1, including means for using natural language processing techniques in determining likelihood of fraud.

[1601] (Claim 3)

[1602] The system of claim 1, further comprising means for detecting specific keywords and contextual information based on the results of the analysis of the text data to determine the possibility of fraud.

[1603] "Application example 2 when combining emotion engines"

[1604] (Claim 1)

[1605] Capture users' telephone conversations as voice data in real time,

[1606] A means for converting the acquired voice data into text data;

[1607] A means of analyzing text data to determine the possibility of fraud;

[1608] a means for notifying the user if a potential fraud is determined;

[1609] a means of providing information to the police if fraud is determined to be a possibility;

[1610] an artificial intelligence agent means for continuing a conversation on behalf of the user;

[1611] means for recognizing an emotional state of a user using an emotion recognition engine;

[1612] means for additionally providing the recognized emotional state to the fraud detection algorithm;

[1613] A system including:

[1614] (Claim 2)

[1615] 10. The system of claim 1, comprising means for using natural language processing techniques in determining likelihood of fraud.

[1616] (Claim 3)

[1617] 10. The system of claim 1, further comprising means for detecting specific keywords, contextual information, and emotional states based on the analysis of the text data to determine likelihood of fraud. [Explanation of symbols]

[1618] 10, 210, 310, 410 Data Processing Systems 12 Data Processing Device 14 Smart Devices 214 Smart Glasses 314 Headset-type terminal 414 Robot< / url:> < / url:> < / url:> < / url:>

Claims

1. Capture users' telephone conversations as voice data in real time, A means for converting the acquired voice data into text data; A means of analyzing text data to determine the possibility of fraud; a means for notifying the user if a potential fraud is determined; a means of providing information to the police if fraud is determined to be a possibility; an artificial intelligence agent means for continuing a conversation on behalf of the user; A system including:

2. 10. The system of claim 1, further comprising means for using natural language processing techniques in determining likelihood of fraud.

3. The system according to claim 1, further comprising means for detecting specific keywords and contextual information based on the results of the analysis of the text data to determine the possibility of fraud.

Images