system

Abstract

We provide the system. [Solution] A method for detecting fraudulent activity by analyzing acoustic information, A means for automatically generating a reaction based on the above analysis, A means of outputting the generated response as sound, A means of sending notifications to registered contacts if there are signs of fraud, A means of conducting natural conversations using AI agents, A system that includes this.

Description

Technical Field

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

Background Art

[0002] Patent Document 1 discloses a persona chatbot control method performed by at least one processor, including steps of receiving a user utterance, adding the user utterance to a prompt including an instruction sentence related to an explanation of a 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

Summary of the Invention

Problems to be Solved by the Invention

[0004] Recently, fraud using telephones has become sophisticated, and particularly, cases where elderly people become victims are occurring frequently. Such fraud is a social problem, and countermeasures are urgently needed. However, in the current telephone response, there is a lack of effective means for early detection and warning of fraud signs.

Means for Solving the Problems

[0005] The present invention provides a system that analyzes acoustic data to detect fraudulent activity, automatically generates a response based on the analysis, and outputs the generated response as audio. This system includes means for sending a notification to a pre-set contact when it determines that fraudulent activity is likely. This makes it possible to prevent fraudulent activity and ensure the safety of the elderly.

[0006] "Fraud" is a general term for any act in which a person deceives others and illegally obtains money or personal information.

[0007] "Audio data" refers to digital or analog information composed of voice or sound.

[0008] "Analysis" is the process of breaking down data and information, understanding its contents in detail, and detecting specific patterns and characteristics.

[0009] "Response generation" is the process of creating an appropriate response based on input information, and is particularly used in dialogue systems.

[0010] "Outputting as audio" is the process of physically reproducing text or digital information as sound and converting it into a form that can be heard by the human ear.

[0011] "Sending a notification" means informing relevant recipients of a specific event or situation electronically or by other means.

[0012] "Contact information" refers to information used when communicating with a specific individual or organization, such as a phone number or email address. [Brief explanation of the drawing]

[0013] [Figure 1] This is a conceptual diagram showing an example of the configuration of a data processing system according to the first embodiment. [Figure 2] This is a conceptual diagram showing an example of the essential functions of a data processing device and a smart device according to the first embodiment. [Figure 3] It is a conceptual diagram showing an example of the configuration of a data processing system according to the second embodiment. [Figure 4] It is a conceptual diagram showing an example of the main functions of a data processing device and smart glasses according to the second embodiment. [Figure 5] It is a conceptual diagram showing an example of the configuration of a data processing system according to the third embodiment. [Figure 6] It is a conceptual diagram showing an example of the main functions of a data processing device and a headset-type terminal according to the third embodiment. [Figure 7] It is a conceptual diagram showing an example of the configuration of a data processing system according to the fourth embodiment. [Figure 8] It is a conceptual diagram showing an example of the main functions of a data processing device and a robot according to the fourth embodiment. [Figure 9] It shows an emotion map to which multiple emotions are mapped. [Figure 10] It shows an emotion map to which multiple emotions are mapped. [Figure 11] It is a sequence diagram showing the processing flow of the data processing system in Example 1. [Figure 12] It is a sequence diagram showing the processing flow of the data processing system in Application Example 1. [Figure 13] It is a sequence diagram showing the processing flow of the data processing system in Example 2 when an emotion engine is combined. [Figure 14] It is a sequence diagram showing the processing flow of the data processing system in Application Example 2 when an emotion engine is combined.

Embodiments for Carrying Out the Invention

[0014] Hereinafter, an example of an embodiment of a system according to the technology of the present disclosure will be described with reference to the accompanying drawings.

[0015] First, the language used in the following description will be explained.

[0016] In the following embodiments, the numbered processor (hereinafter simply referred to as "processor") may be a single arithmetic unit or a combination of multiple arithmetic units. Also, the processor may be a single type of arithmetic unit or a combination of multiple types of arithmetic units. Examples of arithmetic units include a CPU (Central Processing Unit), a GPU (Graphics Processing Unit), a GPGPU (General-Purpose computing on Graphics Processing Units), an APU (Accelerated Processing Unit), and the like.

[0017] In the following embodiments, the numbered RAM (Random Access Memory) is a memory in which information is temporarily stored and is used as a work memory by the processor.

[0018] In the following embodiments, the numbered storage is one or more non-volatile storage devices that store various programs, various parameters, and the like. Examples of non-volatile storage devices include flash memory (SSD (Solid State Drive)), magnetic disks (e.g., hard disks), or magnetic tapes.

[0019] In the following embodiments, the numbered communication I / F (Interface) is an interface including a communication processor, an antenna, and the like. 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), or Bluetooth (registered trademark).

[0020] In the following embodiments, "A and / or B" is synonymous with "at least one of A and B." That is, "A and / or B" means that it may be A alone, or B alone, or a combination of A and B. Furthermore, in this specification, the same concept as "A and / or B" applies when expressing three or more things linked by "and / or."

[0021] [First Embodiment]

[0022] Figure 1 shows an example of the configuration of the data processing system 10 according to the first embodiment.

[0023] As shown in Figure 1, the 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.

[0024] The data processing device 12 comprises a computer 22, a database 24, and a communication interface 26. The computer 22 is an example of a "computer" related to the technology of this disclosure. The computer 22 comprises a processor 28, RAM 30, and storage 32. The processor 28, RAM 30, and storage 32 are connected to a bus 34. The database 24 and the communication interface 26 are also connected to the bus 34. The communication interface 26 is connected to a network 54. An example of the network 54 is a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0025] The smart device 14 comprises a computer 36, a reception device 38, an output device 40, a camera 42, and a communication interface 44. The computer 36 comprises a processor 46, RAM 48, and storage 50. The processor 46, RAM 48, and storage 50 are connected to a bus 52. The reception device 38, output device 40, and camera 42 are also connected to the bus 52.

[0026] The reception device 38 is equipped with a touch panel 38A and a microphone 38B, etc., and receives user input. The touch panel 38A receives user input by detecting contact with an object (e.g., a pen or finger). The microphone 38B receives user input by detecting the user's voice. The control unit 46A transmits data indicating the user input received by the touch panel 38A and 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.

[0027] 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 perceptible to the user 20 (e.g., audio and / or text). The display 40A displays visible information such as text and images according to instructions from the processor 46. The speaker 40B outputs audio according to instructions from the processor 46. The camera 42 is a small digital camera equipped with an optical system such as a lens, aperture, and shutter, and an image sensor such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor.

[0028] Communication interface 44 is connected to network 54. Communication interfaces 44 and 26 are responsible for the exchange of various types of information between processor 46 and processor 28 via network 54.

[0029] Figure 2 shows an example of the main functions of the data processing device 12 and the smart device 14.

[0030] As shown in Figure 2, in the data processing device 12, a specific processing 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" related to the technology of this 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 according to the specific processing program 56 executed on the RAM 30.

[0031] The storage 32 stores the data generation model 58 and the emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[0032] In the smart device 14, the processor 46 performs the reception output processing. The storage 50 stores the reception output program 60. The reception output program 60 is used in conjunction with a 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 processing is realized by the processor 46 operating as a control unit 46A according to the reception output program 60 executed on the RAM 48.

[0033] Next, the specific processing performed by the specific processing unit 290 of the data processing device 12 will be described. In the following description, the data processing device 12 will be referred to as the "server" and the smart device 14 as the "terminal".

[0034] This invention is an automated response system aimed at detecting and preventing fraudulent activity. Specifically, it is implemented by a system that analyzes acoustic data to detect signs of fraud and outputs an automatically generated response as audio.

[0035] The user's device detects incoming calls and sends audio data to a server for calls that may be fraudulent. The server receives this audio data and analyzes the speech using natural language processing techniques. During the analysis, machine learning models are used to detect specific keywords, phrases, and speaker tones associated with fraud.

[0036] If the server determines that fraudulent activity is highly likely, it activates an AI agent. This AI agent generates responses to continue the conversation on behalf of the user. The generated responses mimic the characteristics of an elderly person, incorporating natural pauses and a slow speaking style. This makes it possible to confuse the fraudster or prolong the conversation.

[0037] As soon as the AI ​​agent begins responding, the server sends a notification to pre-registered contacts. This notification includes a summary of the situation and real-time progress, making it easy for family members and trusted individuals to stay informed.

[0038] As a concrete example, let's consider a scenario where a scammer impersonates a bank employee to extract account information from a user. In this case, the AI ​​agent can divert the conversation by saying something like, "I haven't been able to go to the bank much since I joined the facility, that was a while ago..." and buy time. The ultimate goal is to thwart the scammer's attempt and ensure the user's safety.

[0039] The following describes the processing flow.

[0040] Step 1:

[0041] The terminal detects an incoming call. Once the call begins, it starts capturing audio data. This captured audio data is sent to the server in digital format.

[0042] Step 2:

[0043] The server receives audio data transmitted from the terminal and performs real-time analysis using natural language processing (NLP) technology. The goal of this analysis is to detect keywords and unnatural conversation patterns that suggest fraud.

[0044] Step 3:

[0045] Based on the analysis results, the server assesses the likelihood of fraudulent activity. If signs of fraud are detected, it triggers the activation of an AI agent.

[0046] Step 4:

[0047] The server activates an AI agent that generates natural-sounding responses mimicking those of an elderly person. These responses are designed to deflect or prolong conversations with scammers.

[0048] Step 5:

[0049] The server converts the generated response into voice data using speech synthesis technology and outputs it to the scammer via the terminal. This allows the AI ​​agent to continue the conversation on behalf of the user.

[0050] Step 6:

[0051] The server sends a notification to a pre-configured family member or trusted contact once the AI ​​agent begins responding. The notification includes the current situation and the progress of the conversation.

[0052] Step 7:

[0053] The server uses an AI agent to continue the conversation until the scammer ends the call or until instructed to do so by a family member. Once the call ends, the server reports the situation to the user.

[0054] (Example 1)

[0055] Next, we will describe Example 1. In the following description, the data processing device 12 will be referred to as the "server," and the smart device 14 will be referred to as the "terminal."

[0056] There is a need for technology that can prevent damage and losses caused by fraudulent activities and to respond quickly and effectively to potentially fraudulent calls.

[0057] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 1 is realized by the following means.

[0058] In this invention, the server includes means for analyzing voice information input from a communication device to detect fraudulent activity, means for automatically generating response information, and means for outputting the generated response information as sound. This makes it possible to monitor signs of fraudulent activity and take swift action when the likelihood is high.

[0059] A "communication device" refers to an electronic device that has the function of sending and receiving voice information from a user.

[0060] "Voice information" refers to information that represents the content of a phone call as digital data.

[0061] An "analysis tool" is a computer program that processes audio information, analyzes its content, and identifies specific patterns or phrases.

[0062] "Response information" refers to digital data for responses generated based on voice information.

[0063] A "notification" is an electronic message sent to a pre-designated recipient, containing status reports or warnings.

[0064] "Fraud" is an unethical act that attempts to obtain financial or physical benefits by deceiving others.

[0065] One "mode for carrying out the invention" of this invention is a system that automatically detects the possibility of a user making a voice call via a communication network being subjected to fraudulent activity by the other party to the call, and effectively responds to the situation.

[0066] The user's device detects incoming calls and collects voice information in real time once the call begins. This voice information is transmitted to the server via the device's communication module. Security measures, such as data encryption, are implemented during the transmission process.

[0067] The server receives audio information and analyzes it using natural language processing techniques and machine learning models. This analysis evaluates the likelihood of fraudulent activity based on specific keywords, phrases, and the speaker's tone. The software used includes machine learning libraries such as TENSORFLOW® and PyTorch.

[0068] If the analysis determines that there is a high probability of fraudulent activity, the server immediately activates an AI agent. This AI agent uses a generative AI model to generate natural-sounding responses and outputs them as voice. These responses mimic the characteristics of elderly people and have the effect of delaying the conversation with the fraudster.

[0069] Furthermore, when the server detects signs of fraud, it sends an electronic notification to pre-registered contacts. This notification includes information about the potential for fraud and the progress of the call, allowing recipients to immediately understand the situation.

[0070] A concrete example is when a scammer impersonates a bank employee to try and extract a user's personal information. In this case, the AI ​​agent can generate a response such as, "I've been hard of hearing lately, so I can't go to the bank very often..." to confuse the scammer and delay the fraudulent activity.

[0071] As an example of a prompt to the generating AI model, the text "Detect signs of fraud from the voice information during the call and initiate response action" can be used. This will cause the system to perform detection and response processing according to the program.

[0072] The flow of the specific processing in Example 1 will be explained using Figure 11.

[0073] Step 1:

[0074] The user's device detects an incoming call. The input is the incoming call signal, and the output is the trigger for initiating the call. Specifically, the device's communication module receives the incoming call, and once the call is accepted by the user, it prepares to proceed to the next step.

[0075] Step 2:

[0076] The user's device collects voice information in real time during a call and sends it to the server. The input is voice data, and the output is the transmission of voice data to the server. During this process, the device compresses and encrypts the data and transmits it securely over the communication network.

[0077] Step 3:

[0078] The server inputs the received audio data into a speech analysis module and analyzes it using natural language processing techniques. The input is encrypted audio data, and the output is the analysis result. Specifically, the server uses a machine learning model to extract keyword and phrase patterns, speaker tone, and assess the likelihood of fraud.

[0079] Step 4:

[0080] The server determines the likelihood of fraud based on the analysis results. The input is the analysis results, and the output is a fraud likelihood score. The server compares this score to a threshold, and if the likelihood of fraud is high, it proceeds to the next step.

[0081] Step 5:

[0082] If the server determines that there is a high probability of fraud, it activates an AI agent and generates a response using a generative AI model. The input is fraud detection information and the current conversation context, and the output is the generated response data. Specifically, the AI ​​agent generates natural-sounding responses that mimic those of an elderly person, delaying the conversation with the fraudster.

[0083] Step 6:

[0084] The server converts the responses generated by the AI ​​agent into audio signals and sends them to the scammer via the user's device. The input is the generated response data, and the output is the audio signal. This allows the AI ​​agent to continue the conversation.

[0085] Step 7:

[0086] When the server detects signs of fraudulent activity, it sends a notification to pre-registered contacts. Inputs include the potential for fraud and the progress of the call, while output is a notification message. Specifically, the server sends real-time progress information via email or SMS.

[0087] (Application Example 1)

[0088] Next, we will explain Application Example 1. In the following explanation, the data processing device 12 will be referred to as the "server," and the smart device 14 will be referred to as the "terminal."

[0089] In recent years, increasingly sophisticated fraudulent activities have become a social problem. Telephone scams targeting the elderly, in particular, are on the rise, and there is a need for quick and effective means to detect and prevent them. Traditional manual measures have their limitations, so more automated and effective methods are required.

[0090] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 1 is realized by the following means.

[0091] In this invention, the server includes means for analyzing acoustic information to detect fraudulent activity, means for automatically generating a response, and means for outputting the generated response as audio. This makes it possible to detect fraudulent activity in real time during a call and effectively prevent fraud by having an AI agent confuse the fraudster through natural dialogue.

[0092] "Acoustic information" refers to audio data from telephone calls and other voice communications, and analyzing this data forms the basis for detecting signs of fraudulent activity.

[0093] "Analysis" is the process of identifying the characteristics and patterns contained within acoustic information by analyzing it in detail, and then determining the possibility of fraudulent activity.

[0094] "Fraudulent activity" refers to acts of illegally obtaining information or swindling money through telephone calls, and is considered a legally and ethically problematic operation.

[0095] An "AI agent" is a program that uses artificial intelligence technology to engage in natural conversations like a human, and plays a crucial role in dealing with scammers.

[0096] "Natural language processing technology" is a field of computer science and artificial intelligence that involves techniques for computers to understand, interpret, and generate human language.

[0097] "Registered contacts" refers to contacts pre-configured in the system and are used to send notifications in emergencies.

[0098] "Response generation" is the process of creating a response that an AI agent generates when signs of fraudulent activity are detected.

[0099] "Natural dialogue" refers to smooth and easy-to-understand communication, similar to that between humans, and is crucial for preventing fraudsters from engaging in fraudulent activities.

[0100] This section describes the system used to implement this application. This system is designed to detect and prevent fraudulent activity in real time.

[0101] The server first collects acoustic information and converts it into text using Google Cloud Speech-to-Text API and other tools. This text data is then analyzed using natural language processing techniques to detect signs of fraudulent activity. The analysis utilizes machine learning libraries such as TensorFlow and PyTorch to identify keywords specific to fraud and the speaker's tone of voice.

[0102] If fraud is suspected, the server activates an AI agent and automatically generates a response using OpenAI's GPT model. This response mimics human characteristics, particularly the speech patterns of elderly people, to enable natural dialogue with the fraudster. The generated response is output as audio, intended to confuse the fraudster and prolong the conversation.

[0103] The user's device will be notified of the progress of this process, and the status will also be reported to registered contacts. This allows the user to make calls with peace of mind, and enables family members and related parties to understand the situation in real time.

[0104] As a concrete example, when a phone scam seems likely, this system generates a response such as, "I haven't been able to go to the bank much since I moved into the facility, but that was a long time ago..." to divert the scammer's conversation. In this application, the prompt message used as input to the generative AI model is, "Starting phone call analysis. Scam signs detected. Please generate the following elderly-sounding response..."

[0105] The flow of a specific process in Application Example 1 will be explained using Figure 12.

[0106] Step 1:

[0107] The terminal detects incoming calls. Once a call begins, it collects audio data in real time and sends it to the server. The input is audio data, and the output is the audio stream sent to the server.

[0108] Step 2:

[0109] The server converts the received audio data into text using the Google Cloud Speech-to-Text API. The input is an audio stream, and the output is text data extracted from the audio. This process converts the audio data into a parseable text format.

[0110] Step 3:

[0111] The server receives text data and performs analysis using natural language processing techniques. Machine learning models are used for the analysis to detect keywords and tones that indicate fraud. The input is the transformed text data, and the output is an assessment of the likelihood of fraud.

[0112] Step 4:

[0113] If the server determines there is a high probability of fraud, it activates an AI agent. This AI agent uses OpenAI's GPT model to generate natural, elderly-sounding responses. The input is the evaluation result regarding the likelihood of fraud, and the output is the generated dialogue text.

[0114] Step 5:

[0115] The generated dialogue text is converted into speech output using speech synthesis software and played back during the conversation with the scammer. The input is the generated dialogue text, and the output is speech as sound. This process can confuse the scammer and delay the conversation.

[0116] Step 6:

[0117] The server monitors the progress of the situation and sends notifications to registered communication destinations. The input is the fraud detection and AI agent response status, and the output is a notification message. This allows the user, their family, and stakeholders to understand the situation in real time.

[0118] Furthermore, an emotion engine that estimates the user's emotions may be incorporated. That is, the identification processing unit 290 may use the emotion identification model 59 to estimate the user's emotions and perform identification processing using the user's emotions.

[0119] This invention is a system aimed at the automatic detection and prevention of fraudulent activities, further incorporating an emotion engine that recognizes the user's emotions. The system protects the user from fraud by analyzing acoustic data to assess the likelihood of fraud, generating natural responses, and outputting them as voice. Furthermore, the emotion engine allows for real-time monitoring of the user's emotional state and dynamic adjustment of the response content.

[0120] The device detects an incoming call and sends audio data to the server. The server analyzes the received audio data and utilizes multiple natural language processing techniques to detect signs of fraud. The emotion engine identifies the user's emotions from the audio data, particularly monitoring for signs of stress and anxiety. This allows the server to understand the user's psychological state while interacting with the scammer.

[0121] If signs of fraud are detected, the server activates an AI agent, which uses generative AI technology to create responses that mimic those of an elderly person. At this time, an emotion engine considers the user's emotions; for example, if the user is nervous, it can generate a more reassuring response. This response is output using speech synthesis technology, continuing the conversation with the fraudster on the user's behalf.

[0122] Simultaneously, the server notifies pre-registered contacts of the potential fraud and the user's emotional state. The notification includes analysis results, the user's emotional state, and the status of ongoing responses, providing information for family members or trusted individuals to take appropriate action.

[0123] For example, if a scammer asks a user to disclose their asset information, the emotion engine can detect an increase in the user's stress level, and the server can generate a calm response such as, "I was asked something similar before, but I don't know the answer," thereby reducing the user's burden while confusing the scammer. In this way, it is possible to deter fraud and ensure the user's safety and peace of mind.

[0124] The following describes the processing flow.

[0125] Step 1:

[0126] The device detects an incoming call, and once the call begins, it starts capturing audio data in real time. This audio data is then ready to be sent to the server.

[0127] Step 2:

[0128] The server receives audio data transmitted from the terminal and immediately performs analysis using natural language processing technology. The goal is to detect keywords and speaker tones within the audio data that may indicate potential fraud.

[0129] Step 3:

[0130] Simultaneously, the server activates an emotion engine to analyze the user's emotions from the acoustic data. In particular, it objectively evaluates signs of stress and anxiety to understand the user's mental state.

[0131] Step 4:

[0132] If the server determines that a user is likely to be a scam, it activates an AI agent that generates a conversation to respond on behalf of the user using a generative AI. The generated responses mimic the voice of an elderly person and are tailored to reassure the user based on an evaluation by an emotion engine.

[0133] Step 5:

[0134] The server synthesizes the generated conversation as audio and outputs it to the scammer through the terminal. The aim is for the AI ​​agent to continue the conversation on behalf of the user, confusing and exhausting the scammer.

[0135] Step 6:

[0136] Simultaneously, the server sends a detailed notification to pre-registered contacts regarding the user's emotional state and the detection of signs of fraud. This allows family members and trusted individuals to understand the situation and take appropriate action.

[0137] Step 7:

[0138] The server will continue the conversation using an AI agent until the call ends or until there is a special intervention instruction from the family. After the call ends, the server will report a summary of the situation to the user and emergency contacts.

[0139] (Example 2)

[0140] Next, we will describe Example 2. In the following description, the data processing device 12 will be referred to as the "server" and the smart device 14 as the "terminal".

[0141] In modern society, telephone fraud is on the rise, with increasingly sophisticated tactics targeting the elderly. This significantly increases the likelihood of victims suffering economic and psychological losses. Furthermore, the emotional stress experienced by users in connection with fraud is a major concern. Therefore, in addition to fraud detection and prevention, systems that consider user emotions are needed.

[0142] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 2 is realized by the following means.

[0143] In this invention, the server includes means for analyzing acoustic information to evaluate fraudulent activity, means for automatically generating responses, and means for identifying the user's emotional state in real time. This enables early detection and prevention of fraudulent activity, as well as appropriate responses according to the user's emotional state.

[0144] "Acoustic information" refers to all data related to voice and sound, and is digital data obtained from voice communication such as telephone calls.

[0145] "Means of evaluating fraudulent activity" refers to processes and technologies for determining the possibility of fraud or other fraudulent activity by analyzing acoustic information.

[0146] "Means of automatically generating responses" refers to technologies in which a system generates an appropriate response or countermeasure based on analysis results without human intervention.

[0147] "Means of outputting as sound" refers to technologies for reproducing the generated response as actual sound through digital or synthesized speech.

[0148] "Means of sending notifications" refers to the technology or process for electronically transmitting important information to pre-registered contacts.

[0149] "Means for identifying a user's emotional state in real time" refers to technologies that use acoustic information and other data to instantly identify a user's current emotions.

[0150] "Means of dynamic adjustment" refers to technologies that automatically change functions and responses according to the user's state or the system's status.

[0151] "Multiple language processing techniques" refers to techniques that combine and use different natural language processing methods to understand and analyze language data with greater accuracy.

[0152] "Conversations that mimic the characteristics of a specific population group" refers to conversational styles generated by computers that mimic the speaking styles and vocabulary based on specific age groups or social backgrounds.

[0153] This invention is a system aimed at detecting and preventing fraudulent activity, integrating advanced analytical techniques that simultaneously monitor user emotions. The system is primarily based on the analysis of acoustic information and is realized through the cooperation of a server and terminals.

[0154] The device detects incoming calls, records the audio in real time, and generates acoustic information. This acoustic information is transmitted to a server in an encrypted format via the internet. Smartphones and landline telephone equipment are commonly used as the device.

[0155] The server analyzes the received acoustic information and converts the audio into text data using speech recognition technology. Then, natural language processing technology (e.g., NLTK or spaCy) is used to analyze the text data for signs of fraudulent activity. Through this analysis, the likelihood of fraudulent activity can be quickly assessed.

[0156] Furthermore, the server uses emotion analysis tools (e.g., IBM Watson® or Microsoft® Azure® Emotion API) to identify the user's emotions in real time from acoustic information. In particular, it pays attention to changes that indicate stress or anxiety, allowing it to understand the user's psychological state.

[0157] If signs of fraud are detected, the server uses a generative AI model to generate an appropriate response. An example of a prompt used in this process is: "Signs of fraud detected. The user is in a state of tension. Generate a calm response from an elderly person." Based on this prompt, the AI ​​model generates a response that promotes the user's sense of security.

[0158] The generated response is converted into speech using speech synthesis technology (e.g., Google Cloud Text-to-Speech) and output via the device as a response directed at the scammer. This allows the user to avoid direct interaction with the scammer.

[0159] In addition, the server notifies designated contacts of the analysis results and emotional state. This notification includes an assessment of the fraud risk, the user's emotional state, and details of the actions taken, providing information to pre-registered family and friends to understand the situation and take appropriate action. This ensures the user's safety on an ongoing basis.

[0160] The flow of the specific processing in Example 2 will be explained using Figure 13.

[0161] Step 1:

[0162] The device detects incoming calls and records audio in real time once the call begins. The input is the audio during the call, and the output is audio data. This data is transmitted encrypted to a server via the internet for subsequent analysis.

[0163] Step 2:

[0164] The server receives audio data transmitted from the terminal. The input is audio data from the terminal, and the server uses speech recognition technology to convert this audio data into text data. The output is text data, which is the audio converted into text. This process converts the audio information into a format that can be analyzed.

[0165] Step 3:

[0166] The server analyzes fraudulent activity based on the converted text data. The input is text data, and natural language processing techniques are used to identify signs of fraud. Here, specific keywords and contexts are detected, and the likelihood of fraud is assessed. The output is the result of the fraud risk assessment. This step accurately identifies the potential for fraudulent activity.

[0167] Step 4:

[0168] Simultaneously, the server analyzes the acoustic data using emotion analysis tools to identify the user's emotional state in real time. The input is acoustic data, and the output is an evaluation of the emotional state. Particular attention is paid to changes in stress and anxiety, and the user's psychological burden is measured.

[0169] Step 5:

[0170] The server generates a response using a generative AI model when signs of fraud are detected. The input consists of a prompt and data on fraud risk assessment and emotional state. Using the prompt "Signs of fraud detected. User is stressed. Generate a calm response for an elderly person," the server generates a response adjusted by the AI ​​model. The output is the reassuring, adjusted response.

[0171] Step 6:

[0172] The server uses speech synthesis technology to convert the generated response into speech and sends the generated audio to the terminal. The input is a text response from an AI model, and the output is a voice response to the scammer. The terminal plays this audio and continues the conversation with the scammer on behalf of the user.

[0173] Step 7:

[0174] The server notifies registered contacts of the analysis results and the user's emotional state. Inputs include fraud risk assessment, emotional state assessment, and generated responses. Output is a notification message, allowing the user's stakeholders to immediately understand the situation and take appropriate action.

[0175] (Application Example 2)

[0176] Next, we will explain application example 2. In the following explanation, the data processing device 12 will be referred to as a "server" and the smart device 14 as a "terminal".

[0177] In modern society, fraudulent phone calls are becoming more diverse and sophisticated, making it a major challenge to effectively protect potential victims. Furthermore, the psychological burden that fraudulent phone calls place on users due to stress and anxiety cannot be ignored, and there is a need to provide methods to alleviate this burden.

[0178] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means.

[0179] In this invention, the server includes a device for analyzing acoustic information to detect fraudulent activity, a device for analyzing the user's emotional state, and a device for dynamically adjusting the response based on the user's emotions. This makes it possible to respond quickly and accurately to fraudulent activity, and to protect the user from fraud while reducing their psychological burden.

[0180] "Fraudulent activity" refers to the act of deceiving people through illegal means to obtain money or important information.

[0181] "Acoustic information" refers to data used to record or transmit sound waves in digital or analog format.

[0182] An "analysis device" is a device that analyzes input data and extracts or determines specific information.

[0183] A "creation device" is a device that generates some kind of output based on an input.

[0184] "Notification" refers to a means of informing relevant parties of important matters.

[0185] "Emotional state" refers to an individual's emotional reactions and psychological health.

[0186] "Dynamic adjustment" means changing or adapting in real time in response to changes in circumstances or conditions.

[0187] The system for implementing this invention provides a function to detect fraudulent activity using acoustic information and protect users from deception. The server receives and analyzes the acoustic information to detect signs of fraudulent activity. Natural language processing techniques are used for the analysis, making it adaptable to various language environments. This allows for real-time evaluation of the likelihood of fraudulent activity.

[0188] The server further uses algorithms to identify emotions from acoustic information in order to understand the user's emotional state. In particular, it monitors signs of stress and anxiety to assess the user's condition. Based on this information, it generates an appropriate response when the user encounters fraudulent activity.

[0189] The generated responses are created using AI and dynamically adjusted according to the user's emotional state. For example, if the user is feeling anxious, a more reassuring response will be generated. This reduces the user's psychological burden while allowing them to continue the conversation with the scammer. These responses are output using speech synthesis technology.

[0190] Furthermore, if signs of fraudulent activity are detected, the server immediately sends a notification to registered contacts to ensure user safety. For example, if a fraudster requests a user to disclose financial information, the system will generate a response such as "We will verify your financial information later," causing confusion, while simultaneously notifying trusted individuals. In this way, the system provides users with a means to maintain a peaceful life.

[0191] An example of a prompt sentence to input into a generative AI model is, "When a scammer asks for bank account information, how can I safely end the conversation?"

[0192] The flow of a specific process in Application Example 2 will be explained using Figure 14.

[0193] Step 1:

[0194] The terminal acquires acoustic information during a call and sends it to the server. At this point, the input is raw audio data, which is converted to a digital format and transferred to the server. The output is the acoustic data sent to the server.

[0195] Step 2:

[0196] The server analyzes the received audio data and uses natural language processing techniques to detect signs of fraudulent activity. It uses the received audio data as input, analyzing each phrase and voice pattern to find indicators of fraudulent behavior. The output is an evaluation result indicating the likelihood of fraud. This evaluation is performed using text analysis with language models, among other methods.

[0197] Step 3:

[0198] The server simultaneously analyzes the user's emotional state from the acoustic data. The input is the acoustic data from step 1, which is used to identify the emotional state from the voice tone and manner of speaking. The output is the evaluation result of the user's emotional state (e.g., stress level, anxiety).

[0199] Step 4:

[0200] If signs of fraud are detected, the server uses a generative AI model to generate an appropriate response. The evaluation results obtained in steps 2 and 3 are used as input for generating this response. The output is a reconciled voice response that reassures the user and deceives the fraudster.

[0201] Step 5:

[0202] The server converts the generated voice response using speech synthesis technology and outputs it through the terminal. This ensures that appropriate responses are provided even without the user directly participating in the conversation. The input is the response data generated in step 4, and the output is heard from the terminal as synthesized speech.

[0203] Step 6:

[0204] Once the potential for fraud and the user's emotional state are assessed, the server immediately sends a notification to the relevant contacts. At this point, the input is the results of steps 2 and 3, and the output is a warning message based on the analysis results. This message is sent to trusted individuals so that the user can receive appropriate support.

[0205] 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 user input for the result of the specific processing. The control unit 46A transmits the audio data indicating 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.

[0206] Data generation model 58 is a type of so-called generative AI (Artificial Intelligence). An example of data generation model 58 is ChatGPT (registered trademark) (Internet search).<URL: https: / / openai.com / blog / chatgpt> ), Gemini (registered trademark) (Internet search) <url: https: gemini.google.com ?hl="ja">Examples of generative AI include the following. The data generation model 58 is obtained by performing deep learning on a neural network. The data generation model 58 is input with prompts containing instructions, and with inference data such as audio data representing speech, text data representing text, and image data representing images. The data generation model 58 infers from the input inference data according to the instructions indicated by the prompts, and outputs the inference results in data formats such as audio data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[0207] In the above embodiment, an example was given in which specific processing is performed by the data processing device 12, but the technology of this disclosure is not limited thereto, and the specific processing may also be performed by the smart device 14.

[0208] [Second Embodiment]

[0209] Figure 3 shows an example of the configuration of the data processing system 210 according to the second embodiment.

[0210] As shown in Figure 3, the data processing system 210 includes a data processing device 12 and smart glasses 214. An example of the data processing device 12 is a server.

[0211] The data processing device 12 comprises a computer 22, a database 24, and a communication interface 26. The computer 22 is an example of a "computer" related to the technology of this disclosure. The computer 22 comprises a processor 28, RAM 30, and storage 32. The processor 28, RAM 30, and storage 32 are connected to a bus 34. The database 24 and the communication interface 26 are also connected to the bus 34. The communication interface 26 is connected to a network 54. An example of the network 54 is a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0212] The smart glasses 214 include a computer 36, a microphone 238, a speaker 240, a camera 42, and a communication interface 44. The computer 36 includes a processor 46, RAM 48, and storage 50. The processor 46, RAM 48, and storage 50 are connected to a bus 52. The microphone 238, speaker 240, and camera 42 are also connected to the bus 52.

[0213] The microphone 238 receives voice signals from the user 20 and receives instructions from the user 20. The microphone 238 captures the voice signals from the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio according to the instructions from the processor 46.

[0214] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an image sensor such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the area around the user 20 (for example, an imaging range defined by a field of view equivalent to the width of a typical healthy person's field of vision).

[0215] Communication interface 44 is connected to network 54. Communication interfaces 44 and 26 are responsible for the exchange of various information between processor 46 and processor 28 via network 54. The exchange of various information between processor 46 and processor 28 using communication interfaces 44 and 26 is performed in a secure manner.

[0216] Figure 4 shows an example of the main functions of the data processing device 12 and the smart glasses 214. As shown in Figure 4, the data processing device 12 performs specific processing using the processor 28. The storage 32 stores the specific processing program 56.

[0217] The specific processing program 56 is an example of a "program" relating to the technology of this 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.

[0218] The storage 32 stores the data generation model 58 and the emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[0219] In the smart glasses 214, the processor 46 performs the reception output processing. The storage 50 stores the reception output program 60. 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 processing is realized by the processor 46 operating as a control unit 46A according to the reception output program 60 executed on the RAM 48.

[0220] Next, the identification processing performed by the identification processing unit 290 of the data processing device 12 will be described. 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".

[0221] This invention is an automated response system aimed at detecting and preventing fraudulent activity. Specifically, it is implemented by a system that analyzes acoustic data to detect signs of fraud and outputs an automatically generated response as audio.

[0222] The user's device detects incoming calls and sends audio data to a server for calls that may be fraudulent. The server receives this audio data and analyzes the speech using natural language processing techniques. During the analysis, machine learning models are used to detect specific keywords, phrases, and speaker tones associated with fraud.

[0223] If the server determines that fraudulent activity is highly likely, it activates an AI agent. This AI agent generates responses to continue the conversation on behalf of the user. The generated responses mimic the characteristics of an elderly person, incorporating natural pauses and a slow speaking style. This makes it possible to confuse the fraudster or prolong the conversation.

[0224] As soon as the AI ​​agent begins responding, the server sends a notification to pre-registered contacts. This notification includes a summary of the situation and real-time progress, making it easy for family members and trusted individuals to stay informed.

[0225] As a concrete example, let's consider a scenario where a scammer impersonates a bank employee to extract account information from a user. In this case, the AI ​​agent can divert the conversation by saying something like, "I haven't been able to go to the bank much since I joined the facility, that was a while ago..." and buy time. The ultimate goal is to thwart the scammer's attempt and ensure the user's safety.

[0226] The following describes the processing flow.

[0227] Step 1:

[0228] The terminal detects an incoming call. Once the call begins, it starts capturing audio data. This captured audio data is sent to the server in digital format.

[0229] Step 2:

[0230] The server receives audio data transmitted from the terminal and performs real-time analysis using natural language processing (NLP) technology. The goal of this analysis is to detect keywords and unnatural conversation patterns that suggest fraud.

[0231] Step 3:

[0232] Based on the analysis results, the server assesses the likelihood of fraudulent activity. If signs of fraud are detected, it triggers the activation of an AI agent.

[0233] Step 4:

[0234] The server activates an AI agent that generates natural-sounding responses mimicking those of an elderly person. These responses are designed to deflect or prolong conversations with scammers.

[0235] Step 5:

[0236] The server converts the generated response into voice data using speech synthesis technology and outputs it to the scammer via the terminal. This allows the AI ​​agent to continue the conversation on behalf of the user.

[0237] Step 6:

[0238] The server sends a notification to a pre-configured family member or trusted contact once the AI ​​agent begins responding. The notification includes the current situation and the progress of the conversation.

[0239] Step 7:

[0240] The server uses an AI agent to continue the conversation until the scammer ends the call or until instructed to do so by a family member. Once the call ends, the server reports the situation to the user.

[0241] (Example 1)

[0242] Next, we will describe Example 1. 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."

[0243] There is a need for technology that can prevent damage and losses caused by fraudulent activities and to respond quickly and effectively to potentially fraudulent calls.

[0244] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 1 is realized by the following means.

[0245] In this invention, the server includes means for analyzing voice information input from a communication device to detect fraudulent activity, means for automatically generating response information, and means for outputting the generated response information as sound. This makes it possible to monitor signs of fraudulent activity and take swift action when the likelihood is high.

[0246] A "communication device" refers to an electronic device that has the function of sending and receiving voice information from a user.

[0247] "Voice information" refers to information that represents the content of a phone call as digital data.

[0248] An "analysis tool" is a computer program that processes audio information, analyzes its content, and identifies specific patterns or phrases.

[0249] "Response information" refers to digital data for responses generated based on voice information.

[0250] A "notification" is an electronic message sent to a pre-designated recipient, containing status reports or warnings.

[0251] "Fraud" is an unethical act that attempts to obtain financial or physical benefits by deceiving others.

[0252] One "mode for carrying out the invention" of this invention is a system that automatically detects the possibility of a user making a voice call via a communication network being subjected to fraudulent activity by the other party to the call, and effectively responds to the situation.

[0253] The user's device detects incoming calls and collects voice information in real time once the call begins. This voice information is transmitted to the server via the device's communication module. Security measures, such as data encryption, are implemented during the transmission process.

[0254] The server receives audio information and analyzes it using natural language processing techniques and machine learning models. This analysis evaluates the likelihood of fraudulent activity based on specific keywords, phrases, and the speaker's tone. The software used includes machine learning libraries such as TensorFlow and PyTorch.

[0255] If the analysis determines that there is a high probability of fraudulent activity, the server immediately activates an AI agent. This AI agent uses a generative AI model to generate natural-sounding responses and outputs them as voice. These responses mimic the characteristics of elderly people and have the effect of delaying the conversation with the fraudster.

[0256] Furthermore, when the server detects signs of fraud, it sends an electronic notification to pre-registered contacts. This notification includes information about the potential for fraud and the progress of the call, allowing recipients to immediately understand the situation.

[0257] A concrete example is when a scammer impersonates a bank employee to try and extract a user's personal information. In this case, the AI ​​agent can generate a response such as, "I've been hard of hearing lately, so I can't go to the bank very often..." to confuse the scammer and delay the fraudulent activity.

[0258] As an example of a prompt to the generating AI model, the text "Detect signs of fraud from the voice information during the call and initiate response action" can be used. This will cause the system to perform detection and response processing according to the program.

[0259] The flow of the specific processing in Example 1 will be explained using Figure 11.

[0260] Step 1:

[0261] The user's device detects an incoming call. The input is the incoming call signal, and the output is the trigger for initiating the call. Specifically, the device's communication module receives the incoming call, and once the call is accepted by the user, it prepares to proceed to the next step.

[0262] Step 2:

[0263] The user's device collects voice information in real time during a call and sends it to the server. The input is voice data, and the output is the transmission of voice data to the server. During this process, the device compresses and encrypts the data and transmits it securely over the communication network.

[0264] Step 3:

[0265] The server inputs the received audio data into a speech analysis module and analyzes it using natural language processing techniques. The input is encrypted audio data, and the output is the analysis result. Specifically, the server uses a machine learning model to extract keyword and phrase patterns, speaker tone, and assess the likelihood of fraud.

[0266] Step 4:

[0267] The server determines the likelihood of fraud based on the analysis results. The input is the analysis results, and the output is a fraud likelihood score. The server compares this score to a threshold, and if the likelihood of fraud is high, it proceeds to the next step.

[0268] Step 5:

[0269] If the server determines that there is a high probability of fraud, it activates an AI agent and generates a response using a generative AI model. The input is fraud detection information and the current conversation context, and the output is the generated response data. Specifically, the AI ​​agent generates natural-sounding responses that mimic those of an elderly person, delaying the conversation with the fraudster.

[0270] Step 6:

[0271] The server converts the responses generated by the AI ​​agent into audio signals and sends them to the scammer via the user's device. The input is the generated response data, and the output is the audio signal. This allows the AI ​​agent to continue the conversation.

[0272] Step 7:

[0273] When the server detects signs of fraudulent activity, it sends a notification to pre-registered contacts. Inputs include the potential for fraud and the progress of the call, while output is a notification message. Specifically, the server sends real-time progress information via email or SMS.

[0274] (Application Example 1)

[0275] Next, we will explain Application Example 1. In the following explanation, the data processing device 12 will be referred to as the "server," and the smart glasses 214 will be referred to as the "terminal."

[0276] In recent years, increasingly sophisticated fraudulent activities have become a social problem. Telephone scams targeting the elderly, in particular, are on the rise, and there is a need for quick and effective means to detect and prevent them. Traditional manual measures have their limitations, so more automated and effective methods are required.

[0277] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 1 is realized by the following means.

[0278] In this invention, the server includes means for analyzing acoustic information to detect fraudulent activity, means for automatically generating a response, and means for outputting the generated response as audio. This makes it possible to detect fraudulent activity in real time during a call and effectively prevent fraud by having an AI agent confuse the fraudster through natural dialogue.

[0279] "Acoustic information" refers to voice data in telephone and other voice communications, and is the basis for detecting signs of fraud by analyzing this data.

[0280] "Analysis" is a process of specifically analyzing acoustic information to identify the features and patterns contained therein and determine the likelihood of fraud.

[0281] "Fraudulent activities" refer to acts of illegally obtaining information or defrauding money through telephone calls, which are legally and ethically problematic operations.

[0282] "AI agent" is a program that can conduct natural conversations like humans based on artificial intelligence technology and plays an important role in conversations with fraudsters.

[0283] "Natural language processing technology" is a field of computer science and artificial intelligence, and is the technology for enabling computers to understand, interpret, and generate human language.

[0284] "Registered contact" refers to a contact pre-set in the system and is used to send notifications in case of emergencies.

[0285] "Response generation" is a process of creating responses generated by an AI agent when signs of fraud are detected.

[0286] "Natural conversation" refers to smooth and easy-to-understand communication like that conducted by humans and is important for preventing the actions of fraudsters.

[0287] A system for realizing this application example will be described. This system is designed to detect and prevent fraudulent acts in real time.

[0288] The server first collects acoustic information and converts it into text using the Google Cloud Speech-to-Text API, among others. This text data is then analyzed using natural language processing techniques to detect signs of fraudulent activity. The analysis utilizes machine learning libraries such as TensorFlow and PyTorch to identify keywords specific to fraud and the speaker's tone of voice.

[0289] If fraud is suspected, the server activates an AI agent that automatically generates responses using OpenAI's GPT model. These responses mimic human characteristics, particularly the speech patterns of elderly people, to enable natural dialogue with the fraudster. The generated responses are output as audio, intended to confuse the fraudster and prolong the conversation.

[0290] The user's device will be notified of the progress of this process, and the status will also be reported to registered contacts. This allows the user to make calls with peace of mind, and enables family members and related parties to understand the situation in real time.

[0291] As a concrete example, when a phone scam seems likely, this system generates a response such as, "I haven't been able to go to the bank much since I moved into the facility, but that was a long time ago..." to divert the scammer's conversation. In this application, the prompt message used as input to the generative AI model is, "Starting phone call analysis. Scam signs detected. Please generate the following elderly-sounding response..."

[0292] The flow of a specific process in Application Example 1 will be explained using Figure 12.

[0293] Step 1:

[0294] The terminal detects incoming calls. Once a call begins, it collects audio data in real time and sends it to the server. The input is audio data, and the output is the audio stream sent to the server.

[0295] Step 2:

[0296] The server converts the received audio data into text using the Google Cloud Speech-to-Text API. The input is an audio stream, and the output is text data extracted from the audio. This process converts the audio data into a parseable text format.

[0297] Step 3:

[0298] The server receives text data and performs analysis using natural language processing techniques. Machine learning models are used for the analysis to detect keywords and tones that indicate fraud. The input is the transformed text data, and the output is an assessment of the likelihood of fraud.

[0299] Step 4:

[0300] If the server determines there is a high probability of fraud, it activates an AI agent. This AI agent uses OpenAI's GPT model to generate natural, elderly-sounding responses. The input is the evaluation result regarding the likelihood of fraud, and the output is the generated dialogue text.

[0301] Step 5:

[0302] The generated dialogue text is converted into speech output using speech synthesis software and played back during the conversation with the scammer. The input is the generated dialogue text, and the output is speech as sound. This process can confuse the scammer and delay the conversation.

[0303] Step 6:

[0304] The server monitors the progress of the situation and sends notifications to registered communication destinations. The input is the fraud detection and AI agent response status, and the output is a notification message. This allows the user, their family, and stakeholders to understand the situation in real time.

[0305] Furthermore, an emotion engine for estimating the user's emotion may be combined. That is, the specific processing unit 290 may estimate the user's emotion using the emotion identification model 59 and perform specific processing using the user's emotion.

[0306] The present invention is a system for automatically detecting and preventing fraud, and further combines an emotion engine for recognizing the user's emotion. The system evaluates the possibility of fraud by analyzing acoustic data, generates a natural response, and outputs it as voice to protect the user from fraud. In addition, the emotion engine can monitor the user's emotional state in real time and dynamically adjust the content of the response.

[0307] The terminal detects an incoming call and transmits the acoustic data to the server. The server analyzes the received acoustic data and utilizes a plurality of natural language processing techniques to detect signs of fraud. The emotion engine identifies the user's emotion from the acoustic data and monitors signs of stress and anxiety in particular. Thereby, the server can grasp what kind of mental state the user is in during the conversation with the fraudster.

[0308] When signs of fraud are detected, the server activates an AI agent and creates a response imitating an elderly person using generative AI technology. At this time, the emotion engine takes into account the user's emotion, and for example, when the user is nervous, a response that gives a greater sense of security can be generated. This response is output by voice synthesis technology and continues the conversation with the fraudster on behalf of the user.

[0309] At the same time, the server notifies the pre-registered designated contact of the possibility of fraud and the user's emotional state. The notification includes the analysis result, the user's emotional state, and the ongoing response situation, and provides information for family members and trusted persons to take appropriate actions.

[0310] For example, if a scammer asks a user to disclose their asset information, the emotion engine can detect an increase in the user's stress level, and the server can generate a calm response such as, "I was asked something similar before, but I don't know the answer," thereby reducing the user's burden while confusing the scammer. In this way, it is possible to deter fraud and ensure the user's safety and peace of mind.

[0311] The following describes the processing flow.

[0312] Step 1:

[0313] The device detects an incoming call, and once the call begins, it starts capturing audio data in real time. This audio data is then ready to be sent to the server.

[0314] Step 2:

[0315] The server receives audio data transmitted from the terminal and immediately performs analysis using natural language processing technology. The goal is to detect keywords and speaker tones within the audio data that may indicate potential fraud.

[0316] Step 3:

[0317] Simultaneously, the server activates an emotion engine to analyze the user's emotions from the acoustic data. In particular, it objectively evaluates signs of stress and anxiety to understand the user's mental state.

[0318] Step 4:

[0319] If the server determines that a user is likely to be a scam, it activates an AI agent that generates a conversation to respond on behalf of the user using a generative AI. The generated responses mimic the voice of an elderly person and are tailored to reassure the user based on an evaluation by an emotion engine.

[0320] Step 5:

[0321] The server synthesizes the generated conversation as audio and outputs it to the scammer through the terminal. The aim is for the AI ​​agent to continue the conversation on behalf of the user, confusing and exhausting the scammer.

[0322] Step 6:

[0323] Simultaneously, the server sends a detailed notification to pre-registered contacts regarding the user's emotional state and the detection of signs of fraud. This allows family members and trusted individuals to understand the situation and take appropriate action.

[0324] Step 7:

[0325] The server will continue the conversation using an AI agent until the call ends or until there is a special intervention instruction from the family. After the call ends, the server will report a summary of the situation to the user and emergency contacts.

[0326] (Example 2)

[0327] Next, we will describe Example 2. 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".

[0328] In modern society, telephone fraud is on the rise, with increasingly sophisticated tactics targeting the elderly. This significantly increases the likelihood of victims suffering economic and psychological losses. Furthermore, the emotional stress experienced by users in connection with fraud is a major concern. Therefore, in addition to fraud detection and prevention, systems that consider user emotions are needed.

[0329] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 2 is realized by the following means.

[0330] In this invention, the server includes means for analyzing acoustic information to evaluate fraudulent activity, means for automatically generating responses, and means for identifying the user's emotional state in real time. This enables early detection and prevention of fraudulent activity, as well as appropriate responses according to the user's emotional state.

[0331] "Acoustic information" refers to all data related to voice and sound, and is digital data obtained from voice communication such as telephone calls.

[0332] "Means of evaluating fraudulent activity" refers to processes and technologies for determining the possibility of fraud or other fraudulent activity by analyzing acoustic information.

[0333] "Means of automatically generating responses" refers to technologies in which a system generates an appropriate response or countermeasure based on analysis results without human intervention.

[0334] "Means of outputting as sound" refers to technologies for reproducing the generated response as actual sound through digital or synthesized speech.

[0335] "Means of sending notifications" refers to the technology or process for electronically transmitting important information to pre-registered contacts.

[0336] "Means for identifying a user's emotional state in real time" refers to technologies that use acoustic information and other data to instantly identify a user's current emotions.

[0337] "Means of dynamic adjustment" refers to technologies that automatically change functions and responses according to the user's state or the system's status.

[0338] "Multiple language processing techniques" refers to techniques that combine and use different natural language processing methods to understand and analyze language data with greater accuracy.

[0339] "Conversations that mimic the characteristics of a specific population group" refers to conversational styles generated by computers that mimic the speaking styles and vocabulary based on specific age groups or social backgrounds.

[0340] This invention is a system aimed at detecting and preventing fraudulent activity, integrating advanced analytical techniques that simultaneously monitor user emotions. The system is primarily based on the analysis of acoustic information and is realized through the cooperation of a server and terminals.

[0341] The device detects incoming calls, records the audio in real time, and generates acoustic information. This acoustic information is transmitted to a server in an encrypted format via the internet. Smartphones and landline telephone equipment are commonly used as the device.

[0342] The server analyzes the received acoustic information and converts the audio into text data using speech recognition technology. Then, natural language processing technology (e.g., NLTK or spaCy) is used to analyze the text data for signs of fraudulent activity. Through this analysis, the likelihood of fraudulent activity can be quickly assessed.

[0343] Furthermore, the server uses emotion analysis tools (e.g., IBM Watson or Microsoft Azure Emotion API) to identify the user's emotions in real time from the acoustic information. In particular, it pays attention to changes that indicate stress or anxiety, allowing it to understand the user's psychological state.

[0344] If signs of fraud are detected, the server uses a generative AI model to generate an appropriate response. An example of a prompt used in this process is: "Signs of fraud detected. The user is in a state of tension. Generate a calm response from an elderly person." Based on this prompt, the AI ​​model generates a response that promotes the user's sense of security.

[0345] The generated response is converted into speech using speech synthesis technology (e.g., Google Cloud Text-to-Speech) and output via the device as a response directed at the scammer. This allows the user to avoid direct interaction with the scammer.

[0346] In addition, the server notifies designated contacts of the analysis results and emotional state. This notification includes an assessment of the fraud risk, the user's emotional state, and details of the actions taken, providing information to pre-registered family and friends to understand the situation and take appropriate action. This ensures the user's safety on an ongoing basis.

[0347] The flow of the specific processing in Example 2 will be explained using Figure 13.

[0348] Step 1:

[0349] The device detects incoming calls and records audio in real time once the call begins. The input is the audio during the call, and the output is audio data. This data is transmitted encrypted to a server via the internet for subsequent analysis.

[0350] Step 2:

[0351] The server receives audio data transmitted from the terminal. The input is audio data from the terminal, and the server uses speech recognition technology to convert this audio data into text data. The output is text data, which is the audio converted into text. This process converts the audio information into a format that can be analyzed.

[0352] Step 3:

[0353] The server analyzes fraudulent activity based on the converted text data. The input is text data, and natural language processing techniques are used to identify signs of fraud. Here, specific keywords and contexts are detected, and the likelihood of fraud is assessed. The output is the result of the fraud risk assessment. This step accurately identifies the potential for fraudulent activity.

[0354] Step 4:

[0355] Simultaneously, the server analyzes the acoustic data using emotion analysis tools to identify the user's emotional state in real time. The input is acoustic data, and the output is an evaluation of the emotional state. Particular attention is paid to changes in stress and anxiety, and the user's psychological burden is measured.

[0356] Step 5:

[0357] The server generates a response using a generative AI model when signs of fraud are detected. The input consists of a prompt and data on fraud risk assessment and emotional state. Using the prompt "Signs of fraud detected. User is stressed. Generate a calm response for an elderly person," the server generates a response adjusted by the AI ​​model. The output is the reassuring, adjusted response.

[0358] Step 6:

[0359] The server uses speech synthesis technology to convert the generated response into speech and sends the generated audio to the terminal. The input is a text response from an AI model, and the output is a voice response to the scammer. The terminal plays this audio and continues the conversation with the scammer on behalf of the user.

[0360] Step 7:

[0361] The server notifies registered contacts of the analysis results and the user's emotional state. Inputs include fraud risk assessment, emotional state assessment, and generated responses. Output is a notification message, allowing the user's stakeholders to immediately understand the situation and take appropriate action.

[0362] (Application Example 2)

[0363] Next, we will explain application example 2. In the following explanation, the data processing device 12 will be referred to as the "server," and the smart glasses 214 will be referred to as the "terminal."

[0364] In modern society, fraudulent phone calls are becoming more diverse and sophisticated, making it a major challenge to effectively protect potential victims. Furthermore, the psychological burden that fraudulent phone calls place on users due to stress and anxiety cannot be ignored, and there is a need to provide methods to alleviate this burden.

[0365] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means.

[0366] In this invention, the server includes a device for analyzing acoustic information to detect fraudulent activity, a device for analyzing the user's emotional state, and a device for dynamically adjusting the response based on the user's emotions. This makes it possible to respond quickly and accurately to fraudulent activity, and to protect the user from fraud while reducing their psychological burden.

[0367] "Fraudulent activity" refers to the act of deceiving people through illegal means to obtain money or important information.

[0368] "Acoustic information" refers to data used to record or transmit sound waves in digital or analog format.

[0369] An "analysis device" is a device that analyzes input data and extracts or determines specific information.

[0370] A "creation device" is a device that generates some kind of output based on an input.

[0371] "Notification" refers to a means of informing relevant parties of important matters.

[0372] "Emotional state" refers to an individual's emotional reactions and psychological health.

[0373] "Dynamic adjustment" means changing or adapting in real time in response to changes in circumstances or conditions.

[0374] The system for implementing this invention provides a function to detect fraudulent activity using acoustic information and protect users from deception. The server receives and analyzes the acoustic information to detect signs of fraudulent activity. Natural language processing techniques are used for the analysis, making it adaptable to various language environments. This allows for real-time evaluation of the likelihood of fraudulent activity.

[0375] The server further uses algorithms to identify emotions from acoustic information in order to understand the user's emotional state. In particular, it monitors signs of stress and anxiety to assess the user's condition. Based on this information, it generates an appropriate response when the user encounters fraudulent activity.

[0376] The generated responses are created using AI and dynamically adjusted according to the user's emotional state. For example, if the user is feeling anxious, a more reassuring response will be generated. This reduces the user's psychological burden while allowing them to continue the conversation with the scammer. These responses are output using speech synthesis technology.

[0377] Furthermore, if signs of fraudulent activity are detected, the server immediately sends a notification to registered contacts to ensure user safety. For example, if a fraudster requests a user to disclose financial information, the system will generate a response such as "We will verify your financial information later," causing confusion, while simultaneously notifying trusted individuals. In this way, the system provides users with a means to maintain a peaceful life.

[0378] An example of a prompt sentence to input into a generative AI model is, "When a scammer asks for bank account information, how can I safely end the conversation?"

[0379] The flow of a specific process in Application Example 2 will be explained using Figure 14.

[0380] Step 1:

[0381] The terminal acquires acoustic information during a call and sends it to the server. At this point, the input is raw audio data, which is converted to a digital format and transferred to the server. The output is the acoustic data sent to the server.

[0382] Step 2:

[0383] The server analyzes the received audio data and uses natural language processing techniques to detect signs of fraudulent activity. It uses the received audio data as input, analyzing each phrase and voice pattern to find indicators of fraudulent behavior. The output is an evaluation result indicating the likelihood of fraud. This evaluation is performed using text analysis with language models, among other methods.

[0384] Step 3:

[0385] The server simultaneously analyzes the user's emotional state from the acoustic data. The input is the acoustic data from step 1, which is used to identify the emotional state from the voice tone and manner of speaking. The output is the evaluation result of the user's emotional state (e.g., stress level, anxiety).

[0386] Step 4:

[0387] If signs of fraud are detected, the server uses a generative AI model to generate an appropriate response. The evaluation results obtained in steps 2 and 3 are used as input for generating this response. The output is a reconciled voice response that reassures the user and deceives the fraudster.

[0388] Step 5:

[0389] The server converts the generated voice response using speech synthesis technology and outputs it through the terminal. This ensures that appropriate responses are provided even without the user directly participating in the conversation. The input is the response data generated in step 4, and the output is heard from the terminal as synthesized speech.

[0390] Step 6:

[0391] Once the potential for fraud and the user's emotional state are assessed, the server immediately sends a notification to the relevant contacts. At this point, the input is the results of steps 2 and 3, and the output is a warning message based on the analysis results. This message is sent to trusted individuals so that the user can receive appropriate support.

[0392] 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 user input for the result of the specific processing. The control unit 46A transmits the audio data indicating user input acquired by the microphone 238 to the data processing unit 12. In the data processing unit 12, the specific processing unit 290 acquires the audio data.

[0393] Data generation model 58 is a type of so-called generative AI (Artificial Intelligence). One example of data generation model 58 is ChatGPT (Internet search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search) <url: https: gemini.google.com ?hl="ja">Examples of generative AI include the following. The data generation model 58 is obtained by performing deep learning on a neural network. The data generation model 58 is input with prompts containing instructions, and with inference data such as audio data representing speech, text data representing text, and image data representing images. The data generation model 58 infers from the input inference data according to the instructions indicated by the prompts, and outputs the inference results in data formats such as audio data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[0394] In the above embodiment, an example was given in which specific processing is performed by the data processing device 12, but the technology of this disclosure is not limited thereto, and the specific processing may also be performed by the smart glasses 214.

[0395] [Third Embodiment]

[0396] Figure 5 shows an example of the configuration of the data processing system 310 according to the third embodiment.

[0397] As shown in Figure 5, the data processing system 310 includes a data processing device 12 and a headset terminal 314. An example of the data processing device 12 is a server.

[0398] The data processing device 12 comprises a computer 22, a database 24, and a communication interface 26. The computer 22 is an example of a "computer" related to the technology of this disclosure. The computer 22 comprises a processor 28, RAM 30, and storage 32. The processor 28, RAM 30, and storage 32 are connected to a bus 34. The database 24 and the communication interface 26 are also connected to the bus 34. The communication interface 26 is connected to a network 54. An example of the network 54 is a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0399] The headset terminal 314 includes a computer 36, a microphone 238, a speaker 240, a camera 42, a communication interface 44, and a display 343. The computer 36 includes a processor 46, RAM 48, and storage 50. The processor 46, RAM 48, and storage 50 are connected to a bus 52. The microphone 238, speaker 240, camera 42, and display 343 are also connected to the bus 52.

[0400] The microphone 238 receives voice signals from the user 20 and receives instructions from the user 20. The microphone 238 captures the voice signals from the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio according to the instructions from the processor 46.

[0401] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an image sensor such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the area around the user 20 (for example, an imaging range defined by a field of view equivalent to the width of a typical healthy person's field of vision).

[0402] Communication interface 44 is connected to network 54. Communication interfaces 44 and 26 are responsible for the exchange of various information between processor 46 and processor 28 via network 54. The exchange of various information between processor 46 and processor 28 using communication interfaces 44 and 26 is performed in a secure manner.

[0403] Figure 6 shows an example of the main functions of the data processing device 12 and the headset terminal 314. As shown in Figure 6, the data processing device 12 performs specific processing using the processor 28. The storage 32 stores the specific processing program 56.

[0404] The specific processing program 56 is an example of a "program" relating to the technology of this 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.

[0405] The storage 32 stores the data generation model 58 and the emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[0406] In the headset terminal 314, the processor 46 performs the reception output processing. The storage 50 stores the reception output program 60. 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 processing is realized by the processor 46 operating as a control unit 46A according to the reception output program 60 executed on the RAM 48.

[0407] Next, the specific processing performed by the specific processing unit 290 of the data processing device 12 will be described. In the following description, the data processing device 12 will be referred to as the "server" and the headset terminal 314 will be referred to as the "terminal".

[0408] This invention is an automated response system aimed at detecting and preventing fraudulent activity. Specifically, it is implemented by a system that analyzes acoustic data to detect signs of fraud and outputs an automatically generated response as audio.

[0409] The user's device detects incoming calls and sends audio data to a server for calls that may be fraudulent. The server receives this audio data and analyzes the speech using natural language processing techniques. During the analysis, machine learning models are used to detect specific keywords, phrases, and speaker tones associated with fraud.

[0410] If the server determines that fraudulent activity is highly likely, it activates an AI agent. This AI agent generates responses to continue the conversation on behalf of the user. The generated responses mimic the characteristics of an elderly person, incorporating natural pauses and a slow speaking style. This makes it possible to confuse the fraudster or prolong the conversation.

[0411] As soon as the AI ​​agent begins responding, the server sends a notification to pre-registered contacts. This notification includes a summary of the situation and real-time progress, making it easy for family members and trusted individuals to stay informed.

[0412] As a concrete example, let's consider a scenario where a scammer impersonates a bank employee to extract account information from a user. In this case, the AI ​​agent can divert the conversation by saying something like, "I haven't been able to go to the bank much since I joined the facility, that was a while ago..." and buy time. The ultimate goal is to thwart the scammer's attempt and ensure the user's safety.

[0413] The following describes the processing flow.

[0414] Step 1:

[0415] The terminal detects an incoming call. Once the call begins, it starts capturing audio data. This captured audio data is sent to the server in digital format.

[0416] Step 2:

[0417] The server receives audio data transmitted from the terminal and performs real-time analysis using natural language processing (NLP) technology. The goal of this analysis is to detect keywords and unnatural conversation patterns that suggest fraud.

[0418] Step 3:

[0419] Based on the analysis results, the server assesses the likelihood of fraudulent activity. If signs of fraud are detected, it triggers the activation of an AI agent.

[0420] Step 4:

[0421] The server activates an AI agent that generates natural-sounding responses mimicking those of an elderly person. These responses are designed to deflect or prolong conversations with scammers.

[0422] Step 5:

[0423] The server converts the generated response into voice data using speech synthesis technology and outputs it to the scammer via the terminal. This allows the AI ​​agent to continue the conversation on behalf of the user.

[0424] Step 6:

[0425] The server sends a notification to a pre-configured family member or trusted contact once the AI ​​agent begins responding. The notification includes the current situation and the progress of the conversation.

[0426] Step 7:

[0427] The server uses an AI agent to continue the conversation until the scammer ends the call or until instructed to do so by a family member. Once the call ends, the server reports the situation to the user.

[0428] (Example 1)

[0429] Next, we will describe Example 1. 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."

[0430] There is a need for technology that can prevent damage and losses caused by fraudulent activities and to respond quickly and effectively to potentially fraudulent calls.

[0431] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 1 is realized by the following means.

[0432] In this invention, the server includes means for analyzing voice information input from a communication device to detect fraudulent activity, means for automatically generating response information, and means for outputting the generated response information as sound. This makes it possible to monitor signs of fraudulent activity and take swift action when the likelihood is high.

[0433] A "communication device" refers to an electronic device that has the function of sending and receiving voice information from a user.

[0434] "Voice information" refers to information that represents the content of a phone call as digital data.

[0435] An "analysis tool" is a computer program that processes audio information, analyzes its content, and identifies specific patterns or phrases.

[0436] "Response information" refers to digital data for responses generated based on voice information.

[0437] A "notification" is an electronic message sent to a pre-designated recipient, containing status reports or warnings.

[0438] "Fraud" is an unethical act that attempts to obtain financial or physical benefits by deceiving others.

[0439] One "mode for carrying out the invention" of this invention is a system that automatically detects the possibility of a user making a voice call via a communication network being subjected to fraudulent activity by the other party to the call, and effectively responds to the situation.

[0440] The user's device detects incoming calls and collects voice information in real time once the call begins. This voice information is transmitted to the server via the device's communication module. Security measures, such as data encryption, are implemented during the transmission process.

[0441] The server receives audio information and analyzes it using natural language processing techniques and machine learning models. This analysis evaluates the likelihood of fraudulent activity based on specific keywords, phrases, and the speaker's tone. The software used includes machine learning libraries such as TensorFlow and PyTorch.

[0442] If the analysis determines that there is a high probability of fraudulent activity, the server immediately activates an AI agent. This AI agent uses a generative AI model to generate natural-sounding responses and outputs them as voice. These responses mimic the characteristics of elderly people and have the effect of delaying the conversation with the fraudster.

[0443] Furthermore, when the server detects signs of fraud, it sends an electronic notification to pre-registered contacts. This notification includes information about the potential for fraud and the progress of the call, allowing recipients to immediately understand the situation.

[0444] A concrete example is when a scammer impersonates a bank employee to try and extract a user's personal information. In this case, the AI ​​agent can generate a response such as, "I've been hard of hearing lately, so I can't go to the bank very often..." to confuse the scammer and delay the fraudulent activity.

[0445] As an example of a prompt to the generating AI model, the text "Detect signs of fraud from the voice information during the call and initiate response action" can be used. This will cause the system to perform detection and response processing according to the program.

[0446] The flow of the specific processing in Example 1 will be explained using Figure 11.

[0447] Step 1:

[0448] The user's device detects an incoming call. The input is the incoming call signal, and the output is the trigger for initiating the call. Specifically, the device's communication module receives the incoming call, and once the call is accepted by the user, it prepares to proceed to the next step.

[0449] Step 2:

[0450] The user's device collects voice information in real time during a call and sends it to the server. The input is voice data, and the output is the transmission of voice data to the server. During this process, the device compresses and encrypts the data and transmits it securely over the communication network.

[0451] Step 3:

[0452] The server inputs the received audio data into a speech analysis module and analyzes it using natural language processing techniques. The input is encrypted audio data, and the output is the analysis result. Specifically, the server uses a machine learning model to extract keyword and phrase patterns, speaker tone, and assess the likelihood of fraud.

[0453] Step 4:

[0454] The server determines the likelihood of fraud based on the analysis results. The input is the analysis results, and the output is a fraud likelihood score. The server compares this score to a threshold, and if the likelihood of fraud is high, it proceeds to the next step.

[0455] Step 5:

[0456] If the server determines that there is a high probability of fraud, it activates an AI agent and generates a response using a generative AI model. The input is fraud detection information and the current conversation context, and the output is the generated response data. Specifically, the AI ​​agent generates natural-sounding responses that mimic those of an elderly person, delaying the conversation with the fraudster.

[0457] Step 6:

[0458] The server converts the responses generated by the AI ​​agent into audio signals and sends them to the scammer via the user's device. The input is the generated response data, and the output is the audio signal. This allows the AI ​​agent to continue the conversation.

[0459] Step 7:

[0460] When the server detects signs of fraudulent activity, it sends a notification to pre-registered contacts. Inputs include the potential for fraud and the progress of the call, while output is a notification message. Specifically, the server sends real-time progress information via email or SMS.

[0461] (Application Example 1)

[0462] Next, we will explain Application Example 1. In the following explanation, 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."

[0463] In recent years, increasingly sophisticated fraudulent activities have become a social problem. Telephone scams targeting the elderly, in particular, are on the rise, and there is a need for quick and effective means to detect and prevent them. Traditional manual measures have their limitations, so more automated and effective methods are required.

[0464] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 1 is realized by the following means.

[0465] In this invention, the server includes means for analyzing acoustic information to detect fraudulent activity, means for automatically generating a response, and means for outputting the generated response as audio. This makes it possible to detect fraudulent activity in real time during a call and effectively prevent fraud by having an AI agent confuse the fraudster through natural dialogue.

[0466] "Acoustic information" refers to audio data from telephone calls and other voice communications, and analyzing this data forms the basis for detecting signs of fraudulent activity.

[0467] "Analysis" is the process of identifying the characteristics and patterns contained within acoustic information by analyzing it in detail, and then determining the possibility of fraudulent activity.

[0468] "Fraudulent activity" refers to acts of illegally obtaining information or swindling money through telephone calls, and is considered a legally and ethically problematic operation.

[0469] An "AI agent" is a program that uses artificial intelligence technology to engage in natural conversations like a human, and plays a crucial role in dealing with scammers.

[0470] "Natural language processing technology" is a field of computer science and artificial intelligence that involves techniques for computers to understand, interpret, and generate human language.

[0471] "Registered contacts" refers to contacts pre-configured in the system and are used to send notifications in emergencies.

[0472] "Response generation" is the process of creating a response that an AI agent generates when signs of fraudulent activity are detected.

[0473] "Natural dialogue" refers to smooth and easy-to-understand communication, similar to that between humans, and is crucial for preventing fraudsters from engaging in fraudulent activities.

[0474] This section describes the system used to implement this application. This system is designed to detect and prevent fraudulent activity in real time.

[0475] The server first collects acoustic information and converts it into text using the Google Cloud Speech-to-Text API, among others. This text data is then analyzed using natural language processing techniques to detect signs of fraudulent activity. The analysis utilizes machine learning libraries such as TensorFlow and PyTorch to identify keywords specific to fraud and the speaker's tone of voice.

[0476] If fraud is suspected, the server activates an AI agent that automatically generates responses using OpenAI's GPT model. These responses mimic human characteristics, particularly the speech patterns of elderly people, to enable natural dialogue with the fraudster. The generated responses are output as audio, intended to confuse the fraudster and prolong the conversation.

[0477] The user's device will be notified of the progress of this process, and the status will also be reported to registered contacts. This allows the user to make calls with peace of mind, and enables family members and related parties to understand the situation in real time.

[0478] As a concrete example, when a phone scam seems likely, this system generates a response such as, "I haven't been able to go to the bank much since I moved into the facility, but that was a long time ago..." to divert the scammer's conversation. In this application, the prompt message used as input to the generative AI model is, "Starting phone call analysis. Scam signs detected. Please generate the following elderly-sounding response..."

[0479] The flow of a specific process in Application Example 1 will be explained using Figure 12.

[0480] Step 1:

[0481] The terminal detects incoming calls. Once a call begins, it collects audio data in real time and sends it to the server. The input is audio data, and the output is the audio stream sent to the server.

[0482] Step 2:

[0483] The server converts the received audio data into text using the Google Cloud Speech-to-Text API. The input is an audio stream, and the output is text data extracted from the audio. This process converts the audio data into a parseable text format.

[0484] Step 3:

[0485] The server receives text data and performs analysis using natural language processing techniques. Machine learning models are used for the analysis to detect keywords and tones that indicate fraud. The input is the transformed text data, and the output is an assessment of the likelihood of fraud.

[0486] Step 4:

[0487] If the server determines there is a high probability of fraud, it activates an AI agent. This AI agent uses OpenAI's GPT model to generate natural, elderly-sounding responses. The input is the evaluation result regarding the likelihood of fraud, and the output is the generated dialogue text.

[0488] Step 5:

[0489] The generated dialogue text is converted into speech output using speech synthesis software and played back during the conversation with the scammer. The input is the generated dialogue text, and the output is speech as sound. This process can confuse the scammer and delay the conversation.

[0490] Step 6:

[0491] The server monitors the progress of the situation and sends notifications to registered communication destinations. The input is the fraud detection and AI agent response status, and the output is a notification message. This allows the user, their family, and stakeholders to understand the situation in real time.

[0492] Furthermore, an emotion engine that estimates the user's emotions may be incorporated. That is, the identification processing unit 290 may use the emotion identification model 59 to estimate the user's emotions and perform identification processing using the user's emotions.

[0493] This invention is a system aimed at the automatic detection and prevention of fraudulent activities, further incorporating an emotion engine that recognizes the user's emotions. The system protects the user from fraud by analyzing acoustic data to assess the likelihood of fraud, generating natural responses, and outputting them as voice. Furthermore, the emotion engine allows for real-time monitoring of the user's emotional state and dynamic adjustment of the response content.

[0494] The device detects an incoming call and sends audio data to the server. The server analyzes the received audio data and utilizes multiple natural language processing techniques to detect signs of fraud. The emotion engine identifies the user's emotions from the audio data, particularly monitoring for signs of stress and anxiety. This allows the server to understand the user's psychological state while interacting with the scammer.

[0495] If signs of fraud are detected, the server activates an AI agent, which uses generative AI technology to create responses that mimic those of an elderly person. At this time, an emotion engine considers the user's emotions; for example, if the user is nervous, it can generate a more reassuring response. This response is output using speech synthesis technology, continuing the conversation with the fraudster on the user's behalf.

[0496] Simultaneously, the server notifies pre-registered contacts of the potential fraud and the user's emotional state. The notification includes analysis results, the user's emotional state, and the status of ongoing responses, providing information for family members or trusted individuals to take appropriate action.

[0497] For example, if a scammer asks a user to disclose their asset information, the emotion engine can detect an increase in the user's stress level, and the server can generate a calm response such as, "I was asked something similar before, but I don't know the answer," thereby reducing the user's burden while confusing the scammer. In this way, it is possible to deter fraud and ensure the user's safety and peace of mind.

[0498] The following describes the processing flow.

[0499] Step 1:

[0500] The device detects an incoming call, and once the call begins, it starts capturing audio data in real time. This audio data is then ready to be sent to the server.

[0501] Step 2:

[0502] The server receives audio data transmitted from the terminal and immediately performs analysis using natural language processing technology. The goal is to detect keywords and speaker tones within the audio data that may indicate potential fraud.

[0503] Step 3:

[0504] Simultaneously, the server activates an emotion engine to analyze the user's emotions from the acoustic data. In particular, it objectively evaluates signs of stress and anxiety to understand the user's mental state.

[0505] Step 4:

[0506] If the server determines that a user is likely to be a scam, it activates an AI agent that generates a conversation to respond on behalf of the user using a generative AI. The generated responses mimic the voice of an elderly person and are tailored to reassure the user based on an evaluation by an emotion engine.

[0507] Step 5:

[0508] The server synthesizes the generated conversation as audio and outputs it to the scammer through the terminal. The aim is for the AI ​​agent to continue the conversation on behalf of the user, confusing and exhausting the scammer.

[0509] Step 6:

[0510] Simultaneously, the server sends a detailed notification to pre-registered contacts regarding the user's emotional state and the detection of signs of fraud. This allows family members and trusted individuals to understand the situation and take appropriate action.

[0511] Step 7:

[0512] The server will continue the conversation using an AI agent until the call ends or until there is a special intervention instruction from the family. After the call ends, the server will report a summary of the situation to the user and emergency contacts.

[0513] (Example 2)

[0514] Next, we will describe Example 2. 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."

[0515] In modern society, telephone fraud is on the rise, with increasingly sophisticated tactics targeting the elderly. This significantly increases the likelihood of victims suffering economic and psychological losses. Furthermore, the emotional stress experienced by users in connection with fraud is a major concern. Therefore, in addition to fraud detection and prevention, systems that consider user emotions are needed.

[0516] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 2 is realized by the following means.

[0517] In this invention, the server includes means for analyzing acoustic information to evaluate fraudulent activity, means for automatically generating responses, and means for identifying the user's emotional state in real time. This enables early detection and prevention of fraudulent activity, as well as appropriate responses according to the user's emotional state.

[0518] "Acoustic information" refers to all data related to voice and sound, and is digital data obtained from voice communication such as telephone calls.

[0519] "Means of evaluating fraudulent activity" refers to processes and technologies for determining the possibility of fraud or other fraudulent activity by analyzing acoustic information.

[0520] "Means of automatically generating responses" refers to technologies in which a system generates an appropriate response or countermeasure based on analysis results without human intervention.

[0521] "Means of outputting as sound" refers to technologies for reproducing the generated response as actual sound through digital or synthesized speech.

[0522] "Means of sending notifications" refers to the technology or process for electronically transmitting important information to pre-registered contacts.

[0523] "Means for identifying a user's emotional state in real time" refers to technologies that use acoustic information and other data to instantly identify a user's current emotions.

[0524] "Means of dynamic adjustment" refers to technologies that automatically change functions and responses according to the user's state or the system's status.

[0525] "Multiple language processing techniques" refers to techniques that combine and use different natural language processing methods to understand and analyze language data with greater accuracy.

[0526] "Conversations that mimic the characteristics of a specific population group" refers to conversational styles generated by computers that mimic the speaking styles and vocabulary based on specific age groups or social backgrounds.

[0527] This invention is a system aimed at detecting and preventing fraudulent activity, integrating advanced analytical techniques that simultaneously monitor user emotions. The system is primarily based on the analysis of acoustic information and is realized through the cooperation of a server and terminals.

[0528] The device detects incoming calls, records the audio in real time, and generates acoustic information. This acoustic information is transmitted to a server in an encrypted format via the internet. Smartphones and landline telephone equipment are commonly used as the device.

[0529] The server analyzes the received acoustic information and converts the audio into text data using speech recognition technology. Then, natural language processing technology (e.g., NLTK or spaCy) is used to analyze the text data for signs of fraudulent activity. Through this analysis, the likelihood of fraudulent activity can be quickly assessed.

[0530] Furthermore, the server uses emotion analysis tools (e.g., IBM Watson or Microsoft Azure Emotion API) to identify the user's emotions in real time from the acoustic information. In particular, it pays attention to changes that indicate stress or anxiety, allowing it to understand the user's psychological state.

[0531] If signs of fraud are detected, the server uses a generative AI model to generate an appropriate response. An example of a prompt used in this process is: "Signs of fraud detected. The user is in a state of tension. Generate a calm response from an elderly person." Based on this prompt, the AI ​​model generates a response that promotes the user's sense of security.

[0532] The generated response is converted into speech using speech synthesis technology (e.g., Google Cloud Text-to-Speech) and output via the device as a response directed at the scammer. This allows the user to avoid direct interaction with the scammer.

[0533] In addition, the server notifies designated contacts of the analysis results and emotional state. This notification includes an assessment of the fraud risk, the user's emotional state, and details of the actions taken, providing information to pre-registered family and friends to understand the situation and take appropriate action. This ensures the user's safety on an ongoing basis.

[0534] The flow of the specific processing in Example 2 will be explained using Figure 13.

[0535] Step 1:

[0536] The device detects incoming calls and records audio in real time once the call begins. The input is the audio during the call, and the output is audio data. This data is transmitted encrypted to a server via the internet for subsequent analysis.

[0537] Step 2:

[0538] The server receives audio data transmitted from the terminal. The input is audio data from the terminal, and the server uses speech recognition technology to convert this audio data into text data. The output is text data, which is the audio converted into text. This process converts the audio information into a format that can be analyzed.

[0539] Step 3:

[0540] The server analyzes fraudulent activity based on the converted text data. The input is text data, and natural language processing techniques are used to identify signs of fraud. Here, specific keywords and contexts are detected, and the likelihood of fraud is assessed. The output is the result of the fraud risk assessment. This step accurately identifies the potential for fraudulent activity.

[0541] Step 4:

[0542] Simultaneously, the server analyzes the acoustic data using emotion analysis tools to identify the user's emotional state in real time. The input is acoustic data, and the output is an evaluation of the emotional state. Particular attention is paid to changes in stress and anxiety, and the user's psychological burden is measured.

[0543] Step 5:

[0544] The server generates a response using a generative AI model when signs of fraud are detected. The input consists of a prompt and data on fraud risk assessment and emotional state. Using the prompt "Signs of fraud detected. User is stressed. Generate a calm response for an elderly person," the server generates a response adjusted by the AI ​​model. The output is the reassuring, adjusted response.

[0545] Step 6:

[0546] The server uses speech synthesis technology to convert the generated response into speech and sends the generated audio to the terminal. The input is a text response from an AI model, and the output is a voice response to the scammer. The terminal plays this audio and continues the conversation with the scammer on behalf of the user.

[0547] Step 7:

[0548] The server notifies registered contacts of the analysis results and the user's emotional state. Inputs include fraud risk assessment, emotional state assessment, and generated responses. Output is a notification message, allowing the user's stakeholders to immediately understand the situation and take appropriate action.

[0549] (Application Example 2)

[0550] Next, we will explain application example 2. In the following explanation, 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."

[0551] In modern society, fraudulent phone calls are becoming more diverse and sophisticated, making it a major challenge to effectively protect potential victims. Furthermore, the psychological burden that fraudulent phone calls place on users due to stress and anxiety cannot be ignored, and there is a need to provide methods to alleviate this burden.

[0552] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means.

[0553] In this invention, the server includes a device for analyzing acoustic information to detect fraudulent activity, a device for analyzing the user's emotional state, and a device for dynamically adjusting the response based on the user's emotions. This makes it possible to respond quickly and accurately to fraudulent activity, and to protect the user from fraud while reducing their psychological burden.

[0554] "Fraudulent activity" refers to the act of deceiving people through illegal means to obtain money or important information.

[0555] "Acoustic information" refers to data used to record or transmit sound waves in digital or analog format.

[0556] An "analysis device" is a device that analyzes input data and extracts or determines specific information.

[0557] A "creation device" is a device that generates some kind of output based on an input.

[0558] "Notification" refers to a means of informing relevant parties of important matters.

[0559] "Emotional state" refers to an individual's emotional reactions and psychological health.

[0560] "Dynamic adjustment" means changing or adapting in real time in response to changes in circumstances or conditions.

[0561] The system for implementing this invention provides a function to detect fraudulent activity using acoustic information and protect users from deception. The server receives and analyzes the acoustic information to detect signs of fraudulent activity. Natural language processing techniques are used for the analysis, making it adaptable to various language environments. This allows for real-time evaluation of the likelihood of fraudulent activity.

[0562] The server further uses algorithms to identify emotions from acoustic information in order to understand the user's emotional state. In particular, it monitors signs of stress and anxiety to assess the user's condition. Based on this information, it generates an appropriate response when the user encounters fraudulent activity.

[0563] The generated responses are created using AI and dynamically adjusted according to the user's emotional state. For example, if the user is feeling anxious, a more reassuring response will be generated. This reduces the user's psychological burden while allowing them to continue the conversation with the scammer. These responses are output using speech synthesis technology.

[0564] Furthermore, if signs of fraudulent activity are detected, the server immediately sends a notification to registered contacts to ensure user safety. For example, if a fraudster requests a user to disclose financial information, the system will generate a response such as "We will verify your financial information later," causing confusion, while simultaneously notifying trusted individuals. In this way, the system provides users with a means to maintain a peaceful life.

[0565] An example of a prompt sentence to input into a generative AI model is, "When a scammer asks for bank account information, how can I safely end the conversation?"

[0566] The flow of a specific process in Application Example 2 will be explained using Figure 14.

[0567] Step 1:

[0568] The terminal acquires acoustic information during a call and sends it to the server. At this point, the input is raw audio data, which is converted to a digital format and transferred to the server. The output is the acoustic data sent to the server.

[0569] Step 2:

[0570] The server analyzes the received audio data and uses natural language processing techniques to detect signs of fraudulent activity. It uses the received audio data as input, analyzing each phrase and voice pattern to find indicators of fraudulent behavior. The output is an evaluation result indicating the likelihood of fraud. This evaluation is performed using text analysis with language models, among other methods.

[0571] Step 3:

[0572] The server simultaneously analyzes the user's emotional state from the acoustic data. The input is the acoustic data from step 1, which is used to identify the emotional state from the voice tone and manner of speaking. The output is the evaluation result of the user's emotional state (e.g., stress level, anxiety).

[0573] Step 4:

[0574] If signs of fraud are detected, the server uses a generative AI model to generate an appropriate response. The evaluation results obtained in steps 2 and 3 are used as input for generating this response. The output is a reconciled voice response that reassures the user and deceives the fraudster.

[0575] Step 5:

[0576] The server converts the generated voice response using speech synthesis technology and outputs it through the terminal. This ensures that appropriate responses are provided even without the user directly participating in the conversation. The input is the response data generated in step 4, and the output is heard from the terminal as synthesized speech.

[0577] Step 6:

[0578] Once the potential for fraud and the user's emotional state are assessed, the server immediately sends a notification to the relevant contacts. At this point, the input is the results of steps 2 and 3, and the output is a warning message based on the analysis results. This message is sent to trusted individuals so that the user can receive appropriate support.

[0579] The specific processing unit 290 transmits the result of the specific processing to the headset terminal 314. In the headset terminal 314, the control unit 46A causes the speaker 240 and display 343 to output the result of the specific processing. The microphone 238 acquires audio indicating user input for the result of the specific processing. The control unit 46A transmits the audio data indicating user input acquired by the microphone 238 to the data processing unit 12. In the data processing unit 12, the specific processing unit 290 acquires the audio data.

[0580] Data generation model 58 is a type of so-called generative AI (Artificial Intelligence). One example of data generation model 58 is ChatGPT (Internet search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search) <url: https: gemini.google.com ?hl="ja">Examples of generative AI include the following. The data generation model 58 is obtained by performing deep learning on a neural network. The data generation model 58 is input with prompts containing instructions, and with inference data such as audio data representing speech, text data representing text, and image data representing images. The data generation model 58 infers from the input inference data according to the instructions indicated by the prompts, and outputs the inference results in data formats such as audio data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[0581] In the above embodiment, an example was given in which specific processing is performed by the data processing device 12, but the technology of this disclosure is not limited thereto, and specific processing may also be performed by the headset terminal 314.

[0582] [Fourth Embodiment]

[0583] Figure 7 shows an example of the configuration of the data processing system 410 according to the fourth embodiment.

[0584] As shown in Figure 7, the 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.

[0585] The data processing device 12 comprises a computer 22, a database 24, and a communication interface 26. The computer 22 is an example of a "computer" related to the technology of this disclosure. The computer 22 comprises a processor 28, RAM 30, and storage 32. The processor 28, RAM 30, and storage 32 are connected to a bus 34. The database 24 and the communication interface 26 are also connected to the bus 34. The communication interface 26 is connected to a network 54. An example of the network 54 is a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0586] The robot 414 includes a computer 36, a microphone 238, a speaker 240, a camera 42, a communication interface 44, and a controlled object 443. The computer 36 includes a processor 46, RAM 48, and storage 50. The processor 46, RAM 48, and storage 50 are connected to a bus 52. The microphone 238, speaker 240, camera 42, and controlled object 443 are also connected to the bus 52.

[0587] The microphone 238 receives voice signals from the user 20 and receives instructions from the user 20. The microphone 238 captures the voice signals from the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio according to the instructions from the processor 46.

[0588] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an image sensor such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the area around the user 20 (for example, an imaging range defined by a field of view equivalent to the width of a typical healthy person's field of vision).

[0589] Communication interface 44 is connected to network 54. Communication interfaces 44 and 26 are responsible for the exchange of various information between processor 46 and processor 28 via network 54. The exchange of various information between processor 46 and processor 28 using communication interfaces 44 and 26 is performed in a secure manner.

[0590] The controlled object 443 includes a display device, LEDs in the eyes, and motors that drive 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 robot 414's emotions can be expressed by controlling these motors. Furthermore, the robot 414's facial expressions can also be expressed by controlling the illumination state of the LEDs in its eyes.

[0591] Figure 8 shows an example of the main functions of the data processing device 12 and the robot 414. As shown in Figure 8, the data processing device 12 performs specific processing using the processor 28. The storage 32 stores the specific processing program 56.

[0592] The specific processing program 56 is an example of a "program" relating to the technology of this 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.

[0593] The storage 32 stores the data generation model 58 and the emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[0594] In robot 414, the processor 46 performs the reception output processing. The storage 50 stores the reception output program 60. 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 processing is realized by the processor 46 operating as a control unit 46A according to the reception output program 60 executed on the RAM 48.

[0595] Next, the specific processing performed by the specific processing unit 290 of the data processing device 12 will be described. In the following description, the data processing device 12 will be referred to as the "server" and the robot 414 as the "terminal".

[0596] This invention is an automated response system aimed at detecting and preventing fraudulent activity. Specifically, it is implemented by a system that analyzes acoustic data to detect signs of fraud and outputs an automatically generated response as audio.

[0597] The user's device detects incoming calls and sends audio data to a server for calls that may be fraudulent. The server receives this audio data and analyzes the speech using natural language processing techniques. During the analysis, machine learning models are used to detect specific keywords, phrases, and speaker tones associated with fraud.

[0598] If the server determines that fraudulent activity is highly likely, it activates an AI agent. This AI agent generates responses to continue the conversation on behalf of the user. The generated responses mimic the characteristics of an elderly person, incorporating natural pauses and a slow speaking style. This makes it possible to confuse the fraudster or prolong the conversation.

[0599] As soon as the AI ​​agent begins responding, the server sends a notification to pre-registered contacts. This notification includes a summary of the situation and real-time progress, making it easy for family members and trusted individuals to stay informed.

[0600] As a concrete example, let's consider a scenario where a scammer impersonates a bank employee to extract account information from a user. In this case, the AI ​​agent can divert the conversation by saying something like, "I haven't been able to go to the bank much since I joined the facility, that was a while ago..." and buy time. The ultimate goal is to thwart the scammer's attempt and ensure the user's safety.

[0601] The following describes the processing flow.

[0602] Step 1:

[0603] The terminal detects an incoming call. Once the call begins, it starts capturing audio data. This captured audio data is sent to the server in digital format.

[0604] Step 2:

[0605] The server receives audio data transmitted from the terminal and performs real-time analysis using natural language processing (NLP) technology. The goal of this analysis is to detect keywords and unnatural conversation patterns that suggest fraud.

[0606] Step 3:

[0607] Based on the analysis results, the server assesses the likelihood of fraudulent activity. If signs of fraud are detected, it triggers the activation of an AI agent.

[0608] Step 4:

[0609] The server activates an AI agent that generates natural-sounding responses mimicking those of an elderly person. These responses are designed to deflect or prolong conversations with scammers.

[0610] Step 5:

[0611] The server converts the generated response into voice data using speech synthesis technology and outputs it to the scammer via the terminal. This allows the AI ​​agent to continue the conversation on behalf of the user.

[0612] Step 6:

[0613] The server sends a notification to a pre-configured family member or trusted contact once the AI ​​agent begins responding. The notification includes the current situation and the progress of the conversation.

[0614] Step 7:

[0615] The server uses an AI agent to continue the conversation until the scammer ends the call or until instructed to do so by a family member. Once the call ends, the server reports the situation to the user.

[0616] (Example 1)

[0617] Next, we will describe Example 1. In the following description, the data processing device 12 will be referred to as the "server" and the robot 414 as the "terminal".

[0618] There is a need for technology that can prevent damage and losses caused by fraudulent activities and to respond quickly and effectively to potentially fraudulent calls.

[0619] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 1 is realized by the following means.

[0620] In this invention, the server includes means for analyzing voice information input from a communication device to detect fraudulent activity, means for automatically generating response information, and means for outputting the generated response information as sound. This makes it possible to monitor signs of fraudulent activity and take swift action when the likelihood is high.

[0621] A "communication device" refers to an electronic device that has the function of sending and receiving voice information from a user.

[0622] "Voice information" refers to information that represents the content of a phone call as digital data.

[0623] An "analysis tool" is a computer program that processes audio information, analyzes its content, and identifies specific patterns or phrases.

[0624] "Response information" refers to digital data for responses generated based on voice information.

[0625] A "notification" is an electronic message sent to a pre-designated recipient, containing status reports or warnings.

[0626] "Fraud" is an unethical act that attempts to obtain financial or physical benefits by deceiving others.

[0627] One "mode for carrying out the invention" of this invention is a system that automatically detects the possibility of a user making a voice call via a communication network being subjected to fraudulent activity by the other party to the call, and effectively responds to the situation.

[0628] The user's device detects incoming calls and collects voice information in real time once the call begins. This voice information is transmitted to the server via the device's communication module. Security measures, such as data encryption, are implemented during the transmission process.

[0629] The server receives audio information and analyzes it using natural language processing techniques and machine learning models. This analysis evaluates the likelihood of fraudulent activity based on specific keywords, phrases, and the speaker's tone. The software used includes machine learning libraries such as TensorFlow and PyTorch.

[0630] If the analysis determines that there is a high probability of fraudulent activity, the server immediately activates an AI agent. This AI agent uses a generative AI model to generate natural-sounding responses and outputs them as voice. These responses mimic the characteristics of elderly people and have the effect of delaying the conversation with the fraudster.

[0631] Furthermore, when the server detects signs of fraud, it sends an electronic notification to pre-registered contacts. This notification includes information about the potential for fraud and the progress of the call, allowing recipients to immediately understand the situation.

[0632] A concrete example is when a scammer impersonates a bank employee to try and extract a user's personal information. In this case, the AI ​​agent can generate a response such as, "I've been hard of hearing lately, so I can't go to the bank very often..." to confuse the scammer and delay the fraudulent activity.

[0633] As an example of a prompt to the generating AI model, the text "Detect signs of fraud from the voice information during the call and initiate response action" can be used. This will cause the system to perform detection and response processing according to the program.

[0634] The flow of the specific processing in Example 1 will be explained using Figure 11.

[0635] Step 1:

[0636] The user's device detects an incoming call. The input is the incoming call signal, and the output is the trigger for initiating the call. Specifically, the device's communication module receives the incoming call, and once the call is accepted by the user, it prepares to proceed to the next step.

[0637] Step 2:

[0638] The user's device collects voice information in real time during a call and sends it to the server. The input is voice data, and the output is the transmission of voice data to the server. During this process, the device compresses and encrypts the data and transmits it securely over the communication network.

[0639] Step 3:

[0640] The server inputs the received audio data into a speech analysis module and analyzes it using natural language processing techniques. The input is encrypted audio data, and the output is the analysis result. Specifically, the server uses a machine learning model to extract keyword and phrase patterns, speaker tone, and assess the likelihood of fraud.

[0641] Step 4:

[0642] The server determines the likelihood of fraud based on the analysis results. The input is the analysis results, and the output is a fraud likelihood score. The server compares this score to a threshold, and if the likelihood of fraud is high, it proceeds to the next step.

[0643] Step 5:

[0644] If the server determines that there is a high probability of fraud, it activates an AI agent and generates a response using a generative AI model. The input is fraud detection information and the current conversation context, and the output is the generated response data. Specifically, the AI ​​agent generates natural-sounding responses that mimic those of an elderly person, delaying the conversation with the fraudster.

[0645] Step 6:

[0646] The server converts the responses generated by the AI ​​agent into audio signals and sends them to the scammer via the user's device. The input is the generated response data, and the output is the audio signal. This allows the AI ​​agent to continue the conversation.

[0647] Step 7:

[0648] When the server detects signs of fraudulent activity, it sends a notification to pre-registered contacts. Inputs include the potential for fraud and the progress of the call, while output is a notification message. Specifically, the server sends real-time progress information via email or SMS.

[0649] (Application Example 1)

[0650] Next, we will explain Application Example 1. In the following explanation, the data processing device 12 will be referred to as the "server" and the robot 414 as the "terminal".

[0651] In recent years, increasingly sophisticated fraudulent activities have become a social problem. Telephone scams targeting the elderly, in particular, are on the rise, and there is a need for quick and effective means to detect and prevent them. Traditional manual measures have their limitations, so more automated and effective methods are required.

[0652] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 1 is realized by the following means.

[0653] In this invention, the server includes means for analyzing acoustic information to detect fraudulent activity, means for automatically generating a response, and means for outputting the generated response as audio. This makes it possible to detect fraudulent activity in real time during a call and effectively prevent fraud by having an AI agent confuse the fraudster through natural dialogue.

[0654] "Acoustic information" refers to audio data from telephone calls and other voice communications, and analyzing this data forms the basis for detecting signs of fraudulent activity.

[0655] "Analysis" is the process of identifying the characteristics and patterns contained within acoustic information by analyzing it in detail, and then determining the possibility of fraudulent activity.

[0656] "Fraudulent activity" refers to acts of illegally obtaining information or swindling money through telephone calls, and is considered a legally and ethically problematic operation.

[0657] An "AI agent" is a program that uses artificial intelligence technology to engage in natural conversations like a human, and plays a crucial role in dealing with scammers.

[0658] "Natural language processing technology" is a field of computer science and artificial intelligence that involves techniques for computers to understand, interpret, and generate human language.

[0659] "Registered contacts" refers to contacts pre-configured in the system and are used to send notifications in emergencies.

[0660] "Response generation" is the process of creating a response that an AI agent generates when signs of fraudulent activity are detected.

[0661] "Natural dialogue" refers to smooth and easy-to-understand communication, similar to that between humans, and is crucial for preventing fraudsters from engaging in fraudulent activities.

[0662] This section describes the system used to implement this application. This system is designed to detect and prevent fraudulent activity in real time.

[0663] The server first collects acoustic information and converts it into text using the Google Cloud Speech-to-Text API, among others. This text data is then analyzed using natural language processing techniques to detect signs of fraudulent activity. The analysis utilizes machine learning libraries such as TensorFlow and PyTorch to identify keywords specific to fraud and the speaker's tone of voice.

[0664] If fraud is suspected, the server activates an AI agent that automatically generates responses using OpenAI's GPT model. These responses mimic human characteristics, particularly the speech patterns of elderly people, to enable natural dialogue with the fraudster. The generated responses are output as audio, intended to confuse the fraudster and prolong the conversation.

[0665] The user's device will be notified of the progress of this process, and the status will also be reported to registered contacts. This allows the user to make calls with peace of mind, and enables family members and related parties to understand the situation in real time.

[0666] As a concrete example, when a phone scam seems likely, this system generates a response such as, "I haven't been able to go to the bank much since I moved into the facility, but that was a long time ago..." to divert the scammer's conversation. In this application, the prompt message used as input to the generative AI model is, "Starting phone call analysis. Scam signs detected. Please generate the following elderly-sounding response..."

[0667] The flow of a specific process in Application Example 1 will be explained using Figure 12.

[0668] Step 1:

[0669] The terminal detects incoming calls. Once a call begins, it collects audio data in real time and sends it to the server. The input is audio data, and the output is the audio stream sent to the server.

[0670] Step 2:

[0671] The server converts the received audio data into text using the Google Cloud Speech-to-Text API. The input is an audio stream, and the output is text data extracted from the audio. This process converts the audio data into a parseable text format.

[0672] Step 3:

[0673] The server receives text data and performs analysis using natural language processing techniques. Machine learning models are used for the analysis to detect keywords and tones that indicate fraud. The input is the transformed text data, and the output is an assessment of the likelihood of fraud.

[0674] Step 4:

[0675] If the server determines there is a high probability of fraud, it activates an AI agent. This AI agent uses OpenAI's GPT model to generate natural, elderly-sounding responses. The input is the evaluation result regarding the likelihood of fraud, and the output is the generated dialogue text.

[0676] Step 5:

[0677] The generated dialogue text is converted into speech output using speech synthesis software and played back during the conversation with the scammer. The input is the generated dialogue text, and the output is speech as sound. This process can confuse the scammer and delay the conversation.

[0678] Step 6:

[0679] The server monitors the progress of the situation and sends notifications to registered communication destinations. The input is the fraud detection and AI agent response status, and the output is a notification message. This allows the user, their family, and stakeholders to understand the situation in real time.

[0680] Furthermore, an emotion engine that estimates the user's emotions may be incorporated. That is, the identification processing unit 290 may use the emotion identification model 59 to estimate the user's emotions and perform identification processing using the user's emotions.

[0681] This invention is a system aimed at the automatic detection and prevention of fraudulent activities, further incorporating an emotion engine that recognizes the user's emotions. The system protects the user from fraud by analyzing acoustic data to assess the likelihood of fraud, generating natural responses, and outputting them as voice. Furthermore, the emotion engine allows for real-time monitoring of the user's emotional state and dynamic adjustment of the response content.

[0682] The device detects an incoming call and sends audio data to the server. The server analyzes the received audio data and utilizes multiple natural language processing techniques to detect signs of fraud. The emotion engine identifies the user's emotions from the audio data, particularly monitoring for signs of stress and anxiety. This allows the server to understand the user's psychological state while interacting with the scammer.

[0683] If signs of fraud are detected, the server activates an AI agent, which uses generative AI technology to create responses that mimic those of an elderly person. At this time, an emotion engine considers the user's emotions; for example, if the user is nervous, it can generate a more reassuring response. This response is output using speech synthesis technology, continuing the conversation with the fraudster on the user's behalf.

[0684] Simultaneously, the server notifies pre-registered contacts of the potential fraud and the user's emotional state. The notification includes analysis results, the user's emotional state, and the status of ongoing responses, providing information for family members or trusted individuals to take appropriate action.

[0685] For example, if a scammer asks a user to disclose their asset information, the emotion engine can detect an increase in the user's stress level, and the server can generate a calm response such as, "I was asked something similar before, but I don't know the answer," thereby reducing the user's burden while confusing the scammer. In this way, it is possible to deter fraud and ensure the user's safety and peace of mind.

[0686] The following describes the processing flow.

[0687] Step 1:

[0688] The device detects an incoming call, and once the call begins, it starts capturing audio data in real time. This audio data is then ready to be sent to the server.

[0689] Step 2:

[0690] The server receives audio data transmitted from the terminal and immediately performs analysis using natural language processing technology. The goal is to detect keywords and speaker tones within the audio data that may indicate potential fraud.

[0691] Step 3:

[0692] Simultaneously, the server activates an emotion engine to analyze the user's emotions from the acoustic data. In particular, it objectively evaluates signs of stress and anxiety to understand the user's mental state.

[0693] Step 4:

[0694] If the server determines that a user is likely to be a scam, it activates an AI agent that generates a conversation to respond on behalf of the user using a generative AI. The generated responses mimic the voice of an elderly person and are tailored to reassure the user based on an evaluation by an emotion engine.

[0695] Step 5:

[0696] The server synthesizes the generated conversation as audio and outputs it to the scammer through the terminal. The aim is for the AI ​​agent to continue the conversation on behalf of the user, confusing and exhausting the scammer.

[0697] Step 6:

[0698] Simultaneously, the server sends a detailed notification to pre-registered contacts regarding the user's emotional state and the detection of signs of fraud. This allows family members and trusted individuals to understand the situation and take appropriate action.

[0699] Step 7:

[0700] The server will continue the conversation using an AI agent until the call ends or until there is a special intervention instruction from the family. After the call ends, the server will report a summary of the situation to the user and emergency contacts.

[0701] (Example 2)

[0702] Next, we will describe Example 2. In the following description, the data processing device 12 will be referred to as the "server" and the robot 414 as the "terminal".

[0703] In modern society, telephone fraud is on the rise, with increasingly sophisticated tactics targeting the elderly. This significantly increases the likelihood of victims suffering economic and psychological losses. Furthermore, the emotional stress experienced by users in connection with fraud is a major concern. Therefore, in addition to fraud detection and prevention, systems that consider user emotions are needed.

[0704] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 2 is realized by the following means.

[0705] In this invention, the server includes means for analyzing acoustic information to evaluate fraudulent activity, means for automatically generating responses, and means for identifying the user's emotional state in real time. This enables early detection and prevention of fraudulent activity, as well as appropriate responses according to the user's emotional state.

[0706] "Acoustic information" refers to all data related to voice and sound, and is digital data obtained from voice communication such as telephone calls.

[0707] "Means of evaluating fraudulent activity" refers to processes and technologies for determining the possibility of fraud or other fraudulent activity by analyzing acoustic information.

[0708] "Means of automatically generating responses" refers to technologies in which a system generates an appropriate response or countermeasure based on analysis results without human intervention.

[0709] "Means of outputting as sound" refers to technologies for reproducing the generated response as actual sound through digital or synthesized speech.

[0710] "Means of sending notifications" refers to the technology or process for electronically transmitting important information to pre-registered contacts.

[0711] "Means for identifying a user's emotional state in real time" refers to technologies that use acoustic information and other data to instantly identify a user's current emotions.

[0712] "Means of dynamic adjustment" refers to technologies that automatically change functions and responses according to the user's state or the system's status.

[0713] "Multiple language processing techniques" refers to techniques that combine and use different natural language processing methods to understand and analyze language data with greater accuracy.

[0714] "Conversations that mimic the characteristics of a specific population group" refers to conversational styles generated by computers that mimic the speaking styles and vocabulary based on specific age groups or social backgrounds.

[0715] This invention is a system aimed at detecting and preventing fraudulent activity, integrating advanced analytical techniques that simultaneously monitor user emotions. The system is primarily based on the analysis of acoustic information and is realized through the cooperation of a server and terminals.

[0716] The device detects incoming calls, records the audio in real time, and generates acoustic information. This acoustic information is transmitted to a server in an encrypted format via the internet. Smartphones and landline telephone equipment are commonly used as the device.

[0717] The server analyzes the received acoustic information and converts the audio into text data using speech recognition technology. Then, natural language processing technology (e.g., NLTK or spaCy) is used to analyze the text data for signs of fraudulent activity. Through this analysis, the likelihood of fraudulent activity can be quickly assessed.

[0718] Furthermore, the server uses emotion analysis tools (e.g., IBM Watson or Microsoft Azure Emotion API) to identify the user's emotions in real time from the acoustic information. In particular, it pays attention to changes that indicate stress or anxiety, allowing it to understand the user's psychological state.

[0719] If signs of fraud are detected, the server uses a generative AI model to generate an appropriate response. An example of a prompt used in this process is: "Signs of fraud detected. The user is in a state of tension. Generate a calm response from an elderly person." Based on this prompt, the AI ​​model generates a response that promotes the user's sense of security.

[0720] The generated response is converted into speech using speech synthesis technology (e.g., Google Cloud Text-to-Speech) and output via the device as a response directed at the scammer. This allows the user to avoid direct interaction with the scammer.

[0721] In addition, the server notifies designated contacts of the analysis results and emotional state. This notification includes an assessment of the fraud risk, the user's emotional state, and details of the actions taken, providing information to pre-registered family and friends to understand the situation and take appropriate action. This ensures the user's safety on an ongoing basis.

[0722] The flow of the specific processing in Example 2 will be explained using Figure 13.

[0723] Step 1:

[0724] The device detects incoming calls and records audio in real time once the call begins. The input is the audio during the call, and the output is audio data. This data is transmitted encrypted to a server via the internet for subsequent analysis.

[0725] Step 2:

[0726] The server receives audio data transmitted from the terminal. The input is audio data from the terminal, and the server uses speech recognition technology to convert this audio data into text data. The output is text data, which is the audio converted into text. This process converts the audio information into a format that can be analyzed.

[0727] Step 3:

[0728] The server analyzes fraudulent activity based on the converted text data. The input is text data, and natural language processing techniques are used to identify signs of fraud. Here, specific keywords and contexts are detected, and the likelihood of fraud is assessed. The output is the result of the fraud risk assessment. This step accurately identifies the potential for fraudulent activity.

[0729] Step 4:

[0730] Simultaneously, the server analyzes the acoustic data using emotion analysis tools to identify the user's emotional state in real time. The input is acoustic data, and the output is an evaluation of the emotional state. Particular attention is paid to changes in stress and anxiety, and the user's psychological burden is measured.

[0731] Step 5:

[0732] The server generates a response using a generative AI model when signs of fraud are detected. The input consists of a prompt and data on fraud risk assessment and emotional state. Using the prompt "Signs of fraud detected. User is stressed. Generate a calm response for an elderly person," the server generates a response adjusted by the AI ​​model. The output is the reassuring, adjusted response.

[0733] Step 6:

[0734] The server uses speech synthesis technology to convert the generated response into speech and sends the generated audio to the terminal. The input is a text response from an AI model, and the output is a voice response to the scammer. The terminal plays this audio and continues the conversation with the scammer on behalf of the user.

[0735] Step 7:

[0736] The server notifies registered contacts of the analysis results and the user's emotional state. Inputs include fraud risk assessment, emotional state assessment, and generated responses. Output is a notification message, allowing the user's stakeholders to immediately understand the situation and take appropriate action.

[0737] (Application Example 2)

[0738] Next, we will explain application example 2. In the following explanation, the data processing device 12 will be referred to as the "server" and the robot 414 as the "terminal".

[0739] In modern society, fraudulent phone calls are becoming more diverse and sophisticated, making it a major challenge to effectively protect potential victims. Furthermore, the psychological burden that fraudulent phone calls place on users due to stress and anxiety cannot be ignored, and there is a need to provide methods to alleviate this burden.

[0740] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means.

[0741] In this invention, the server includes a device for analyzing acoustic information to detect fraudulent activity, a device for analyzing the user's emotional state, and a device for dynamically adjusting the response based on the user's emotions. This makes it possible to respond quickly and accurately to fraudulent activity, and to protect the user from fraud while reducing their psychological burden.

[0742] "Fraudulent activity" refers to the act of deceiving people through illegal means to obtain money or important information.

[0743] "Acoustic information" refers to data used to record or transmit sound waves in digital or analog format.

[0744] An "analysis device" is a device that analyzes input data and extracts or determines specific information.

[0745] A "creation device" is a device that generates some kind of output based on an input.

[0746] "Notification" refers to a means of informing relevant parties of important matters.

[0747] "Emotional state" refers to an individual's emotional reactions and psychological health.

[0748] "Dynamic adjustment" means changing or adapting in real time in response to changes in circumstances or conditions.

[0749] The system for implementing this invention provides a function to detect fraudulent activity using acoustic information and protect users from deception. The server receives and analyzes the acoustic information to detect signs of fraudulent activity. Natural language processing techniques are used for the analysis, making it adaptable to various language environments. This allows for real-time evaluation of the likelihood of fraudulent activity.

[0750] The server further uses algorithms to identify emotions from acoustic information in order to understand the user's emotional state. In particular, it monitors signs of stress and anxiety to assess the user's condition. Based on this information, it generates an appropriate response when the user encounters fraudulent activity.

[0751] The generated responses are created using AI and dynamically adjusted according to the user's emotional state. For example, if the user is feeling anxious, a more reassuring response will be generated. This reduces the user's psychological burden while allowing them to continue the conversation with the scammer. These responses are output using speech synthesis technology.

[0752] Furthermore, if signs of fraudulent activity are detected, the server immediately sends a notification to registered contacts to ensure user safety. For example, if a fraudster requests a user to disclose financial information, the system will generate a response such as "We will verify your financial information later," causing confusion, while simultaneously notifying trusted individuals. In this way, the system provides users with a means to maintain a peaceful life.

[0753] An example of a prompt sentence to input into a generative AI model is, "When a scammer asks for bank account information, how can I safely end the conversation?"

[0754] The flow of a specific process in Application Example 2 will be explained using Figure 14.

[0755] Step 1:

[0756] The terminal acquires acoustic information during a call and sends it to the server. At this point, the input is raw audio data, which is converted to a digital format and transferred to the server. The output is the acoustic data sent to the server.

[0757] Step 2:

[0758] The server analyzes the received audio data and uses natural language processing techniques to detect signs of fraudulent activity. It uses the received audio data as input, analyzing each phrase and voice pattern to find indicators of fraudulent behavior. The output is an evaluation result indicating the likelihood of fraud. This evaluation is performed using text analysis with language models, among other methods.

[0759] Step 3:

[0760] The server simultaneously analyzes the user's emotional state from the acoustic data. The input is the acoustic data from step 1, which is used to identify the emotional state from the voice tone and manner of speaking. The output is the evaluation result of the user's emotional state (e.g., stress level, anxiety).

[0761] Step 4:

[0762] If signs of fraud are detected, the server uses a generative AI model to generate an appropriate response. The evaluation results obtained in steps 2 and 3 are used as input for generating this response. The output is a reconciled voice response that reassures the user and deceives the fraudster.

[0763] Step 5:

[0764] The server converts the generated voice response using speech synthesis technology and outputs it through the terminal. This ensures that appropriate responses are provided even without the user directly participating in the conversation. The input is the response data generated in step 4, and the output is heard from the terminal as synthesized speech.

[0765] Step 6:

[0766] Once the potential for fraud and the user's emotional state are assessed, the server immediately sends a notification to the relevant contacts. At this point, the input is the results of steps 2 and 3, and the output is a warning message based on the analysis results. This message is sent to trusted individuals so that the user can receive appropriate support.

[0767] 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 controlled object 443 to output the result of the specific processing. The microphone 238 acquires audio indicating user input for the result of the specific processing. The control unit 46A transmits the audio data indicating user input acquired by the microphone 238 to the data processing unit 12. In the data processing unit 12, the specific processing unit 290 acquires the audio data.

[0768] Data generation model 58 is a type of so-called generative AI (Artificial Intelligence). One example of data generation model 58 is ChatGPT (Internet search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search) <url: https: gemini.google.com ?hl="ja">Examples of generative AI include the following. The data generation model 58 is obtained by performing deep learning on a neural network. The data generation model 58 is input with prompts containing instructions, and with inference data such as audio data representing speech, text data representing text, and image data representing images. The data generation model 58 infers from the input inference data according to the instructions indicated by the prompts, and outputs the inference results in data formats such as audio data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[0769] In the above embodiment, an example was given in which specific processing is performed by the data processing device 12, but the technology of this disclosure is not limited thereto, and the specific processing may also be performed by the robot 414.

[0770] Furthermore, the emotion identification model 59, acting 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 a specific mapping, which is an emotion map (see Figure 9). Similarly, the emotion identification model 59 may also determine the robot's emotion, and the identification processing unit 290 may perform identification processing using the robot's emotion.

[0771] Figure 9 shows an emotion map 400 in which multiple emotions are mapped. In the emotion map 400, emotions are arranged in concentric circles radiating from the center. The closer to the center of the concentric circles, the more primitive the emotions are located. Further out of the concentric circles, emotions representing states and actions arising from mental states are located. Emotion is a concept that includes feelings and mental states. On the left side of the concentric circles, emotions that are generally generated from reactions occurring in the brain are located. On the right side of the concentric circles, emotions that are generally induced by situational judgment are located. Above and below the concentric circles, emotions that are generally generated from reactions occurring in the brain and induced by situational judgment are located. In addition, the emotion of "pleasure" is located on the upper side of the concentric circles, and the emotion of "displeasure" is located on the lower side. Thus, in the emotion map 400, multiple emotions are mapped based on the structure in which emotions arise, and emotions that are likely to occur simultaneously are mapped close together.

[0772] These emotions are distributed at the 3 o'clock position on the Emotion Map 400, and usually fluctuate between feelings of security and anxiety. In the right half of the Emotion Map 400, situational awareness takes precedence over internal feelings, resulting in a calm impression.

[0773] The inside of the Emotion Map 400 represents inner thoughts, while the outside represents actions. Therefore, the further you go from the outside of the Emotion Map 400, the more visible (expressed in actions) your emotions become.

[0774] Here, human emotions are based on various balances, such as posture and blood sugar levels. When these balances deviate from the ideal, it results in discomfort, and when they approach the ideal, it results in pleasure. Similarly, in robots, cars, motorcycles, etc., emotions can be created based on various balances, such as posture and battery level. When these balances deviate from the ideal, it results in discomfort, and when they approach the ideal, it results in pleasure. The emotion map can be generated, for example, based on Dr. Mitsuyoshi's emotion map (Research on a system for analyzing brain physiological signals of speech emotion recognition and emotion, Tokushima University, doctoral dissertation: https: / / ci.nii.ac.jp / naid / 500000375379). The left half of the emotion map contains emotions belonging to a region called "response," where sensation is dominant. The right half of the emotion map contains emotions belonging to a region called "situation," where situational awareness is dominant.

[0775] The emotion map defines two emotions that promote learning. One is the emotion around the middle of the negative "repentance" and "reflection" on the situation side. In other words, it is 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 the emotion around the positive "desire" on the reaction side. In other words, it is when the robot has positive feelings such as "I want more" or "I want to know more."

[0776] The emotion identification model 59 inputs user input into a pre-trained neural network, obtains emotion values ​​representing each emotion shown in the emotion map 400, and determines the user's emotion. This neural network is pre-trained based on multiple training data sets, which are combinations of user input and emotion values ​​representing each emotion shown in the emotion map 400. Furthermore, this neural network is trained so that emotions located close together have similar values, as shown in the emotion map 900 in Figure 10. Figure 10 shows an example where multiple emotions such as "reassured," "calm," and "confident" have similar emotion values.

[0777] The above description primarily focuses on the functions of the data processing device 12 in relation to this disclosure. However, the system related to this disclosure is not necessarily implemented on a server. The system related to this disclosure may be implemented as a general information processing system. This disclosure may be implemented, for example, as a software program that runs on a personal computer or as an application that runs on a smartphone. The method related to this disclosure may be provided to users in SaaS (Software as a Service) format.

[0778] In the above embodiment, an example was given in which a specific process is performed by a single computer 22. However, the technology of this disclosure is not limited thereto, and a distributed processing of the specific process may be performed by multiple computers, including computer 22. For example, a data generation model 58 may be provided in an external device of the data processing device 12, and the external device may generate data according to the input data.

[0779] In the above embodiment, an example was given in which the specific processing program 56 is stored in the storage 32, but the technology of this disclosure is not limited thereto. For example, the specific processing program 56 may be stored in a portable, computer-readable, non-temporary storage medium such as a USB (Universal Serial Bus) memory. The specific processing program 56 stored in the non-temporary storage medium is installed in the computer 22 of the data processing device 12. The processor 28 executes specific processing according to the specific processing program 56.

[0780] 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.

[0781] Furthermore, it is not necessary to store the entirety 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 the entirety of the specific processing program 56 in the storage 32; it is acceptable to store only a portion of the specific processing program 56.

[0782] The following types of processors can be used as hardware resources to perform specific processing. Examples of processors include a CPU, a general-purpose processor that functions as a hardware resource to perform specific processing by executing software, i.e., a program. Other examples of processors include dedicated electrical circuits, such as FPGAs (Field-Programmable Gate Arrays), PLDs (Programmable Logic Devices), or ASICs (Application Specific Integrated Circuits), which have circuit configurations specifically designed to perform specific processing. All of these processors have built-in or connected memory, and all of them perform specific processing by using memory.

[0783] The hardware resource that performs a specific process may consist of one of these various processors, or it may consist of 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). Alternatively, the hardware resource that performs a specific process may consist of a single processor.

[0784] Examples of configurations using a single processor include, firstly, a configuration in which one or more CPUs and software are combined to form a single processor, and this processor functions as a hardware resource that performs a specific process. Secondly, there is a configuration using a processor that realizes the functions of the entire system, including multiple hardware resources that perform a specific process, on a single IC chip, as exemplified by SoCs (System-on-a-chip). In this way, a specific process is realized using one or more of the above types of processors as hardware resources.

[0785] Furthermore, the hardware structure of these various processors can more specifically utilize electrical circuits that combine circuit elements such as semiconductor devices. Also, the specific processing described above is merely an example. Therefore, it goes without saying that unnecessary steps can be deleted, new steps added, or the processing order rearranged, as long as it does not deviate from the main purpose.

[0786] The descriptions and illustrations presented above are detailed explanations of the technical aspects of this disclosure and are merely examples of the technical aspects. For example, the above descriptions of the structure, function, operation, and effect are examples of the structure, function, operation, and effect of the technical aspects of this disclosure. Therefore, it goes without saying that you may delete unnecessary parts, add new elements, or replace elements in the descriptions and illustrations presented above, as long as you do not deviate from the essence of the technical aspects of this disclosure. Furthermore, in order to avoid confusion and facilitate understanding of the technical aspects of this disclosure, explanations of common technical knowledge and the like that do not require special explanation to enable the implementation of the technical aspects of this disclosure have been omitted from the descriptions and illustrations presented above.

[0787] All documents, patent applications, and technical standards described herein are incorporated by reference to the same extent as if each individual document, patent application, and technical standard were specifically and individually noted to be incorporated by reference.

[0788] The following is further disclosed regarding the embodiments described above.

[0789] (Claim 1)

[0790] A means of analyzing acoustic data to detect fraudulent activity,

[0791] A means for automatically generating a response based on the above analysis,

[0792] A means for outputting the generated response as audio,

[0793] A means of sending a notification to a designated contact in the event of suspected fraudulent activity,

[0794] A system that includes this.

[0795] (Claim 2)

[0796] The system according to claim 1, wherein the analysis means detects fraudulent activity using multiple language processing methods.

[0797] (Claim 3)

[0798] The system according to claim 1, wherein the response generation means generates conversations that mimic the characteristics of an elderly person.

[0799] "Example 1"

[0800] (Claim 1)

[0801] A means of detecting fraudulent activity by analyzing voice information input from a communication device,

[0802] A means for automatically generating response information based on the above analysis,

[0803] A means of outputting the generated response information as sound,

[0804] A means of sending notifications to pre-configured recipients in the event of potential fraud,

[0805] A system that includes this.

[0806] (Claim 2)

[0807] The system according to claim 1, wherein the analysis means analyzes speech information using a natural language processing device.

[0808] (Claim 3)

[0809] The system according to claim 1, wherein the response generation means creates a conversation that mimics the characteristics of a specific person.

[0810] "Application Example 1"

[0811] (Claim 1)

[0812] A method for detecting fraudulent activity by analyzing acoustic information,

[0813] A means for automatically generating a reaction based on the above analysis,

[0814] A means of outputting the generated response as sound,

[0815] A means of sending notifications to registered contacts if there are signs of fraud,

[0816] A means of conducting natural conversations using AI agents,

[0817] A system that includes this.

[0818] (Claim 2)

[0819] The system according to claim 1, wherein the analysis means detects fraudulent activity using a number of natural language processing techniques.

[0820] (Claim 3)

[0821] The system according to claim 1, wherein the response generating means generates a dialogue that imitates the way an elderly person speaks.

[0822] "Example 2 of combining an emotion engine"

[0823] (Claim 1)

[0824] A means of analyzing acoustic information to evaluate fraudulent activity,

[0825] A means for automatically generating a reaction based on the aforementioned analysis,

[0826] A means of outputting the generated response as sound,

[0827] A means of sending a notification to registered contacts when there is a high possibility of fraudulent activity,

[0828] A means of identifying the user's emotional state in real time,

[0829] A means of dynamically adjusting responses according to the user's emotional state,

[0830] A system that includes this.

[0831] (Claim 2)

[0832] The system according to claim 1, wherein the analysis means evaluates fraudulent activity using multiple language processing techniques.

[0833] (Claim 3)

[0834] The system according to claim 1, wherein the response generating means generates conversations that mimic the characteristics of a specific population group.

[0835] "Application example 2 when combining with an emotional engine"

[0836] (Claim 1)

[0837] A device that analyzes acoustic information to detect fraudulent activity,

[0838] A device that automatically generates a response based on the above analysis,

[0839] A device that outputs the generated response as sound,

[0840] A device that sends a notification to a registered communication recipient in the event of potential fraudulent activity,

[0841] A device that analyzes acoustic information to understand the user's emotional state,

[0842] A device that dynamically adjusts its response based on the user's emotions,

[0843] A system that includes this.

[0844] (Claim 2)

[0845] The system according to claim 1, wherein the analysis device detects fraudulent activity using various language processing methods.

[0846] (Claim 3)

[0847] The system according to claim 1, wherein the response generation device generates conversations that mimic the characteristics of elderly people, thereby reducing the burden on the user. [Explanation of Symbols]

[0848] 10, 210, 310, 410 Data Processing Systems 12 Data Processing Devices 14 Smart Devices 214 Smart Glasses 314 Headset-type terminal 414 Robots< / url:> < / url:> < / url:> < / url:>

Claims

1. A method for detecting fraudulent activity by analyzing acoustic information, A means for automatically generating a reaction based on the above analysis, A means of outputting the generated response as sound, A means of sending notifications to registered contacts if there are signs of fraud, A means of conducting natural conversations using AI agents, A system that includes this.

2. The system according to claim 1, wherein the analysis means detects fraudulent activity using a number of natural language processing techniques.

3. The system according to claim 1, wherein the response generating means generates a dialogue that imitates the way an elderly person speaks.

Images