system

A system converts voice data to text, uses AI to analyze and respond to suspicious calls, and blocks fraudulent numbers, effectively protecting the elderly from phone scams by immediate identification and user notification.

JP2026074994APending Publication Date: 2026-05-07SOFTBANK GROUP CORP
View PDF 1 Cites 0 Cited by

Patent Information

Authority / Receiving Office
JP · JP
Patent Type
Applications
Current Assignee / Owner
SOFTBANK GROUP CORP
Filing Date
2024-10-21
Publication Date
2026-05-07

AI Technical Summary

Technical Problem

Special fraud targeting the elderly uses sophisticated tactics to deceive personal information and money via phone calls, and existing methods fail to block suspicious calls in advance, posing a risk of contacting fraud even once.

Method used

A system that converts voice data into text in real time, analyzes the content using a generative AI model for pattern recognition, generates natural-sounding responses to handle suspicious calls, records suspicious numbers in a database for automatic blocking, and notifies users of fraud prevention information.

Benefits of technology

Reduces the risk of elderly individuals becoming victims of fraud by identifying and responding to potentially fraudulent calls in real time, blocking repeat calls, and increasing user awareness of potential scams.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure 2026074994000001_ABST
    Figure 2026074994000001_ABST
Patent Text Reader

Abstract

We provide the system. [Solution] A means of converting audio data into text data, A means for analyzing the text data using a generative AI model and evaluating the suspiciousness of the communication, A means of generating a natural response when it is determined that there is a high probability of fraud, A method to record suspicious phone numbers in a database and update the blacklist, A system that includes means for notifying users of fraud prevention information.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0002] , , , , , ,

[0004] , , , , , ,

[0005] , , , , ,

[0003] , , , ,

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

Background Art

[0002] Patent Document 1 discloses a method for controlling a persona chatbot, which is performed by at least one processor, including the steps of receiving a user utterance, adding the user utterance to a prompt including an instruction sentence related to an explanation of the chatbot character, encoding the prompt, and inputting the encoded prompt into a language model to generate a chatbot utterance in response to the user utterance.

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] Special fraud targeting the elderly uses sophisticated tactics to deceive personal information and money via phone calls. Such fraud may expand due to the decline in the judgment of the elderly and the risk of dementia. As a means of protecting the elderly, such as parents living in a remote area, from these frauds, a mechanism that can identify fraud in real time and respond appropriately is required. With the conventional method, it is difficult to block suspicious calls in advance, and there is a risk of contacting fraud even once.

Means for Solving the Problems

[0005] This invention provides a system that determines the likelihood of fraud in real time by converting voice data into text and analyzing the content of a call using a generation AI model. The system converts voice to text in real time, and then the AI ​​performs pattern recognition to identify suspicious communications. If a call is deemed highly likely to be fraudulent, it generates a natural-sounding response and handles the situation on behalf of the elderly. Furthermore, suspicious phone numbers are recorded in a database, and automatic blocking is implemented by updating a blacklist. Users are regularly notified of fraud prevention information, contributing to increased safety awareness. In this way, the system aims to reduce the risk of elderly people becoming victims of fraud.

[0006] "Audio data" refers to data that records sound in digital format and serves as basic input information for processing in computer systems.

[0007] "Text data" refers to strings of information converted from sources such as speech, and is in a format suitable for processing by generative AI models and analysis engines.

[0008] A "generative AI model" is an artificial intelligence model that learns from large amounts of data to perform specific tasks, and it plays a role in identifying patterns and methods of fraud.

[0009] A "natural response" refers to a response created using language generation technology that allows for responses that sound natural and believable in human conversation.

[0010] A "database" is a system that stores information systematically and has a structure that makes it easy to search and update, and it is used for recording and managing fraudulent phone numbers.

[0011] A "blacklist" is a list of information (such as phone numbers) that has been deemed suspicious or malicious, and is used to block calls or connections.

[0012] A "user" refers to a person or family member who uses the system and is the entity that receives information notifications and warnings.

[0013] "Real-time" refers to the property of information processing being performed immediately and the results being reflected without delay, enabling a rapid response. [Brief explanation of the drawing]

[0014] [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] This is a conceptual diagram showing an example of the configuration of a data processing system according to the second embodiment. [Figure 4] This 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] This is a conceptual diagram showing an example of the configuration of a data processing system according to the third embodiment. [Figure 6] This 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] This is a conceptual diagram showing an example of the configuration of a data processing system according to the fourth embodiment. [Figure 8] This 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] This shows an emotion map where multiple emotions are mapped. [Figure 10] This shows an emotion map where multiple emotions are mapped. [Figure 11] This is a sequence diagram showing the processing flow of the data processing system in Example 1. [Figure 12] This 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 Embodiment 2 when the 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 the emotion engine is combined.

Mode for Carrying Out the Invention

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

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

[0017] In the following embodiments, the numbered processor (hereinafter simply referred to as "processor") may be one arithmetic unit or a combination of a plurality of arithmetic units. Also, the processor may be one type of arithmetic unit or a combination of a plurality of 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.

[0018] 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.

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

[0020] In the following embodiments, the signed communication interface (I / F) is an interface that includes a communication processor and an antenna, etc. The communication interface manages communication between multiple computers. Examples of communication standards applicable to the communication interface include wireless communication standards such as 5G (5th Generation Mobile Communication System), Wi-Fi (registered trademark), or Bluetooth (registered trademark).

[0021] 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."

[0022] [First Embodiment]

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

[0024] 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.

[0025] 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).

[0026] 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.

[0027] 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.

[0028] The output device 40 includes a display 40A and a speaker 40B, and presents data to the user 20 by outputting the data in a form 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.

[0029] 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.

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

[0031] 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.

[0032] 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.

[0033] 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.

[0034] 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".

[0035] As an embodiment of this invention, a system having the following functions is constructed: The server monitors the user's voice calls in real time and acquires voice data. The acquired voice data is converted into text data on the server by a speech recognition engine. This text is input into a generative AI model to analyze whether the call content is suspicious and evaluate the possibility of fraud.

[0036] The server receives feedback from the generative AI model and generates a natural response if it determines that the call is likely to be fraudulent. This allows the server to handle suspicious calls on behalf of the user. For example, the server can generate an appropriate response such as "Please wait while I check the invoice" to calm the caller.

[0037] On the other hand, the server records suspicious phone numbers in a database and periodically updates the blacklist. Each time the update is performed, the system is configured to block future incoming calls from the registered suspicious phone numbers. This prevents secondary damage from the same fraudsters.

[0038] Users will receive notifications via their devices when potentially fraudulent calls are detected, as well as periodic notifications about fraud prevention. These notifications allow users to review suspicious call content, helping to raise their security awareness.

[0039] This system can prevent fraudulent activities that target users who are particularly vulnerable to fraud, such as the elderly, and enable safe communication.

[0040] The following describes the processing flow.

[0041] Step 1:

[0042] The server receives user voice calls in real time. A speech recognition engine is used to convert the voice data into text data. This conversion is performed quickly to enable immediate analysis of the call content.

[0043] Step 2:

[0044] The server inputs the converted text data into a generating AI model. The AI ​​model identifies suspicious patterns and phrases from the text and assesses the likelihood of fraud. Based on the assessment results, it quantifies the risk of fraud.

[0045] Step 3:

[0046] If the server detects a high probability of fraud, it prepares a natural-sounding response generated by an AI model. This response follows the flow of human conversation and is intended to delay or mislead the caller of a fraudulent call.

[0047] Step 4:

[0048] The server records phone numbers deemed suspicious in its database. This updates the blacklist, preparing to automatically block future calls from that number.

[0049] Step 5:

[0050] The device will send a notification to the user if it detects a potentially fraudulent call. This notification will include details of the call and a warning about the scam, helping to raise user awareness.

[0051] Step 6:

[0052] Users receive notifications from their devices and take appropriate measures to respond to potentially fraudulent calls. This gives them an opportunity to avoid trouble before it happens.

[0053] (Example 1)

[0054] 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."

[0055] In modern communication technology, users are vulnerable to fraud, and there is a lack of means to quickly identify suspicious calls and identification numbers. Effective measures are needed to reduce fraud, especially for the elderly and tech-savvy users. The challenge lies in achieving a safe communication environment.

[0056] 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.

[0057] In this invention, the server includes means for converting voice information into text information, means for analyzing the text information using a generative information processing model and evaluating the suspiciousness of the communication, and means for recording suspicious identification numbers in an information storage device and updating a restriction list. This makes it possible for users to receive advance warnings about suspicious calls and prevent fraud.

[0058] "Auditory information" refers to data related to sounds such as human speech that are transmitted as sound waves.

[0059] "Textual information" refers to data obtained by converting audio information into text, and is used in natural language processing.

[0060] A "generative information processing model" refers to a trained artificial intelligence technology used for natural language generation and analysis.

[0061] "Means for evaluating suspicious activity" refers to a function that analyzes communication content and patterns to detect suspicious activity that differs from normal communication.

[0062] An "identification number" is a combination of numbers or characters used to identify the caller.

[0063] An "information storage device" is a device or system for recording and storing data temporarily or for a long period of time.

[0064] A "restriction list" is a list used to register suspicious communication partners and restrict or prohibit future contact.

[0065] "User" refers to an individual or organization that uses this system.

[0066] To implement this invention, it is necessary to construct a system having the following elements: The server monitors the user's voice calls in real time using a communication infrastructure. To acquire voice, voice streaming technology is used to collect voice information, and a speech recognition engine is used to convert it into text information. Specifically, a speech recognition API (e.g., a cloud-based service) can be used for speech recognition.

[0067] The server inputs the converted character information into a generative information processing model and uses the model's analytical capabilities to evaluate the suspiciousness of the communication. This generative information processing model uses a natural language processing model based on existing technologies (e.g., a large-scale language model). Based on the analysis results, the server generates a natural response and, if necessary, deals with suspicious calls.

[0068] Furthermore, the server has a mechanism to prevent similar suspicious communication activities by storing suspicious identification numbers in an information storage device and updating the restriction list. The terminal has a notification function related to fraud prevention, providing users with security-related information in real time.

[0069] For example, if a user prompts, "Have I had any suspicious calls recently?", the server scans recent call logs and notifies the user of details about any calls suspected of being fraudulent. In this way, users can more securely manage their own calling environment. This system helps combat fraudulent activities and provides a safe and secure communication experience.

[0070] The flow of the specific processing in Example 1 will be explained using Figure 11.

[0071] Step 1:

[0072] The server acquires voice data from a user who initiates a voice call in real time. An audio stream is provided as input. The server transfers this audio data to a speech recognition engine, where it is converted into text data. The output is a string of characters reflecting the content of the call. This process involves accumulating the audio data in a buffer and inputting it to the speech recognition engine in batch processing.

[0073] Step 2:

[0074] The server inputs the obtained text data into the generative AI model. The string generated in the previous step is used as input. The generative AI model analyzes this string and evaluates whether the communication content is suspicious. The results of this analysis are scored, and the evaluation result is obtained as output. Specifically, this involves the model analyzing text patterns and quantifying the likelihood of fraud.

[0075] Step 3:

[0076] The server generates a natural response if the evaluation result exceeds a threshold. In this process, it receives an evaluation score from a generative AI model as input and constructs a response using a response generation engine. The output is the response that should be sent to the caller. The operation includes creating the optimal response by referring to past conversation data and preset templates.

[0077] Step 4:

[0078] The server records suspicious identification numbers in a database. The input is the identification numbers detected during a call, which are stored in the information storage device. The output is an updated restricted list. The operation includes a mechanism where new identification numbers are added to the existing list, and this information is periodically transmitted to other user terminals.

[0079] Step 5:

[0080] The device notifies the user of fraud prevention information. Here, it receives a warning message from the server as input and displays it in the user interface. The output is a warning message that the user can review. Specifically, this includes actions such as a notification popping up on the screen and providing detailed information as needed.

[0081] (Application Example 1)

[0082] 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."

[0083] In recent years, fraudulent activities have become a serious social problem, particularly telephone scams targeting the elderly, and effective preventative measures are needed. The problem is that many users continue calls without fully understanding the risks of fraud. Furthermore, a system is needed that can immediately detect communications with a high probability of being fraudulent and warn users.

[0084] 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.

[0085] In this invention, the server includes means for converting acoustic information into text information, means for analyzing the text information using a generation AI model and evaluating the suspiciousness of the communication, and means for displaying a warning for suspicious communication and providing an option to terminate the call. This makes it possible to immediately detect communications that are highly likely to be fraudulent and to help users make safe decisions.

[0086] "Means for converting acoustic information into text information" refers to a device or program that performs the process of converting audio data into text data in real time.

[0087] "Means for analyzing the aforementioned textual information using a generative AI model and evaluating the suspiciousness of the communication" refers to a device or program that uses artificial intelligence to analyze textual information and determine whether the content of the communication is fraudulent.

[0088] "Means for generating an automated response when a communication is deemed highly likely to be fraudulent" refers to a device or program that automatically generates an appropriate response when it detects a communication that is strongly suspected of being fraudulent.

[0089] "Means for recording suspicious identification information in a data storage device and updating the restriction list" refers to a device or program that stores suspicious phone numbers or identification information in a database and periodically updates the restriction list.

[0090] "Means of notifying users of fraud prevention information" refers to a device or program that notifies the user's terminal of information regarding the risk of fraud.

[0091] "Means for displaying a warning about suspicious communications and providing an option to terminate a call" refers to a device or program that, upon detecting suspicious communications, presents a warning to the user and provides an interface for choosing whether to continue or terminate the call.

[0092] The system for implementing this invention primarily involves the processes of acquiring, analyzing, responding to, and recording voice data. First, the terminal immediately transmits the received call audio as acoustic information to the server. The server converts this acoustic information into text information using the Google® Cloud Speech-to-Text API. The converted text information is then analyzed by a generative AI model, in this case OpenAI® GPT, to assess the likelihood of fraud.

[0093] If the server determines that there is a high probability of fraud, it will automatically generate an appropriate response message. This response message can instruct the other party, such as "We are checking, please wait a moment," in a way that does not make the call seem unnatural.

[0094] The server also records suspicious identification information, such as phone numbers, in its database. This data is added to a restricted list, and settings are made to block future communications. Furthermore, users are displayed a warning about the risk of fraud through their device and given the option to continue or end the call. This feature allows users to ensure their own safety through their own judgment.

[0095] For example, if an elderly user receives a phone call from a stranger requesting personal information, this system will respond quickly. After the audio is converted into text on the server, the user will be warned that "this call may be a scam," allowing for a swift response. An example of a prompt to the generating AI model is, "Does the following text contain signs of fraud? Please tell me if it is a scam and why."

[0096] The flow of a specific process in Application Example 1 will be explained using Figure 12.

[0097] Step 1:

[0098] The terminal instantly transmits the call audio received by the user to the server. The input is the audio data obtained from the call, and the output is the transmission of this audio data to the server. The terminal's operation involves digitizing the received audio signal and transferring it to the server using a secure communication protocol.

[0099] Step 2:

[0100] The server uses the Google Cloud Speech-to-Text API to convert the received audio data into text. The input is the audio data received in step 1, and the output is the converted text. In this step, a speech recognition algorithm is applied, the audio waveform is analyzed within the server, and the corresponding text is converted.

[0101] Step 3:

[0102] The server uses the OpenAI GPT generative AI model to analyze textual information and evaluate the suspiciousness of the communication. The input is the textual information obtained in step 2, and the output is the evaluation result regarding the likelihood of fraud. In this step, the prompt "Does the following text contain signs of fraud? Please tell us the likelihood of fraud and why." is used to input textual information into the generative AI model and quantify the risk of fraud.

[0103] Step 4:

[0104] If the server determines that there is a high probability of fraud, it generates an automated response message. The input is the evaluation result obtained in step 3, and the output is the generated response message. The server uses a generation AI model to generate a calming response such as "We are checking, please wait a moment."

[0105] Step 5:

[0106] The server records suspicious identification information in a data storage device and updates the restriction list. The input is the identification information of a call suspected of being fraudulent, and the output is the updated restriction list. The server writes this identification information to a database and automatically configures settings to block future calls.

[0107] Step 6:

[0108] The user will see a warning message about the risk of fraud through their device. The input is the evaluation result from step 3, and the output is the warning message that the user can see on the screen. The device presents the user with the risks of continuing the call and the option to end it, allowing the user to end the call at their own discretion if necessary.

[0109] 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.

[0110] As an embodiment of this invention, a system incorporating a server, a terminal, and an emotion engine is provided. This system monitors the user's voice calls and processes the data acquired by speech recognition in real time. The server converts the voice data into text data and performs analysis using a generative AI model and an emotion engine.

[0111] The generative AI model analyzes the call content and assesses the likelihood of fraud. The emotion engine recognizes emotions from the user's voice and evaluates their state in real time. Based on this information, the server determines the appropriate course of action, considering not only the risk of fraud but also the psychological impact the call has on the user.

[0112] Specifically, the server generates more cautious responses when it determines there is a high probability of fraud, as well as when it assesses that the user is in a stressful state. For example, it can use phrases such as "Please calm down" or "Please wait while I check the details" to reduce the user's stress while the call progresses.

[0113] Furthermore, the server records suspicious phone numbers in a database. User safety is ensured by continuously updating the blacklist and automatically blocking incoming calls from known fraudulent phone numbers.

[0114] The device will notify users of potentially fraudulent calls and information about the user's emotional state. This allows users to take appropriate action. For example, a notification such as "A potential scam has been detected. Please be careful" will appear on the device.

[0115] This system allows users, including the elderly, to avoid the risk of fraud while reducing psychological burden. By combining it with emotional recognition, it goes beyond mere technical responses and enables activities that are empathetic to the user's feelings.

[0116] The following describes the processing flow.

[0117] Step 1:

[0118] The server receives user voice calls in real time. A speech recognition engine is used to convert the voice data into text data. This process extracts the call content, enabling immediate analysis.

[0119] Step 2:

[0120] The server inputs the converted text data into a generating AI model. This model analyzes the call content and assesses the likelihood of fraud. Specifically, it searches for matches with past fraud patterns and calculates the fraud risk.

[0121] Step 3:

[0122] The server inputs voice data from the user into an emotion engine to analyze the user's emotional state. This engine evaluates the user's stress and anxiety levels based on the tone and pace of their voice.

[0123] Step 4:

[0124] The server integrates the evaluation of the generative AI model and the analysis results of the emotion engine. If the risk of fraud is high and the user's emotional state is unstable, it generates a cautious response. This response includes words that soothe the user's emotions.

[0125] Step 5:

[0126] The server records suspicious phone numbers in its database. It updates the blacklist and configures the system to automatically block incoming calls from those numbers.

[0127] Step 6:

[0128] The device sends notifications to the user based on the possibility of fraud and their emotional state. These notifications may include messages such as, "Fraud has been detected. Please stay calm."

[0129] Step 7:

[0130] Users can check notifications from their devices and take action based on the analysis of call details and emotional state. This helps prevent fraud and promotes emotional stability.

[0131] (Example 2)

[0132] 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".

[0133] In recent years, there has been a growing need for effective systems to protect users from telephone-based fraud. In particular, for users, including the elderly, it is necessary to quickly assess the risk of fraud and provide responses that reduce psychological burden. However, existing technologies make it difficult to provide a comprehensive response that simultaneously considers fraud risk assessment and psychological impact. To solve this problem, a system is needed that evaluates the suspiciousness of communications through real-time voice analysis and generates appropriate responses that take into account the user's emotional state.

[0134] 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.

[0135] In this invention, the server includes means for converting voice data into text data, means for analyzing the text data using a generative AI model and evaluating the suspiciousness of the communication, and means for evaluating the user's emotional state in real time and determining how to respond to the call based on that information. This makes it possible to quickly evaluate communications that are likely to be fraudulent and provide responses that take the user's emotions into consideration, thereby preventing communication troubles and ensuring the user's safety and psychological security.

[0136] "Methods for converting audio data into text data" refers to technologies that use speech recognition technology to convert audio into strings of characters, creating a text format that can be used for subsequent data analysis and processing.

[0137] "Means for analyzing the text data using a generative AI model and evaluating the suspiciousness of the communication" refers to a technology that uses an AI model trained through machine learning to analyze text data and evaluate the risk of fraud based on its content.

[0138] "A means of evaluating a user's emotional state in real time and determining how to respond to a call based on that information" refers to a technology that recognizes emotions from a user's voice data and determines an appropriate response or action based on the results.

[0139] "A means of recording suspicious communication destinations in a database and updating the information list" refers to a technology that identifies potentially fraudulent communication destinations, records them in a database, and updates the list to improve the security of future communications.

[0140] "Means of notifying terminals of fraud prevention information" refers to technology that provides a function to display warnings and alerts on the user's terminal device regarding the possibility of detected fraud.

[0141] "Means of generating careful responses that take into account the impact on the user" refers to technologies that construct responses based on the user's emotional state and fraud risk assessment to provide accurate information while reducing psychological burden.

[0142] In an embodiment of the present invention, the system is designed to analyze voice calls in real time, assess the risk of fraud and the user's emotional state, and generate appropriate responses. The main components of the system include a server, a terminal, a speech recognition engine, a generative AI model, and an emotion engine.

[0143] The server uses a speech recognition engine to convert the user's voice call data into text data. This process utilizes commonly used speech recognition technologies such as the Google Cloud Speech-to-Text API, enabling highly accurate speech-to-text conversion.

[0144] Subsequently, the server analyzes the converted text data using a generative AI model to evaluate the suspiciousness of the communication. It is expected that OpenAI's GPT series will be used as the generative AI model. The generative AI model analyzes the text data based on the prompt message. A specific example of a prompt message is, "Is this call potentially fraudulent? Please assess the risk of fraud."

[0145] Simultaneously, the server uses an emotion engine to evaluate the user's emotional state in real time. This emotion engine utilizes tools such as Affectiva, which provides Emotion AI technology. Through this process, the server determines whether the user is experiencing stress or anxiety, and uses this information to improve call handling.

[0146] The device displays fraud prevention-related information to the user based on notifications from the server. The notifications displayed on the device are warning messages such as "This may be a scam. Please be careful," and are designed to help the user take appropriate action immediately.

[0147] This system protects users from fraudulent activities and allows them to communicate with peace of mind. It also reduces psychological burden and ensures safety for users unfamiliar with digital devices, such as the elderly. By providing support from both a technological and emotional perspective, a high-quality user experience is achieved.

[0148] The flow of the specific processing in Example 2 will be explained using Figure 13.

[0149] Step 1:

[0150] The server receives the user's voice call. It receives voice data as input and analyzes the data using a speech recognition engine. It converts the voice data into text data in real time using APIs such as the Google Cloud Speech-to-Text API. This results in text data as output. Specifically, this process involves the server sending the voice file to the speech recognition API.

[0151] Step 2:

[0152] The server inputs the text data obtained in Step 1 into the generating AI model. The input at this stage is the text data that is the output of Step 1. The generating AI model uses OpenAI's GPT series and performs text analysis based on the prompt sentence. For the purpose of evaluating fraud risk, a prompt sentence such as "Is this call potentially fraudulent? Please evaluate the fraud risk." is used. As a result of the analysis, a fraud risk evaluation score is obtained as output. Specifically, the process involves the server sending the text data to the AI ​​model and performing the analysis.

[0153] Step 3:

[0154] The server passes the audio data to the emotion engine to analyze the user's emotional state. The input at this stage is the audio data handled in step 1. Using emotion recognition technology such as Affectiva, the user's emotions are evaluated in real time. The output is an emotional state report indicating whether the user is stressed or not. The specific operation involves the server inputting audio data into the emotion engine and obtaining the analysis results.

[0155] Step 4:

[0156] The server determines how to respond to the call based on the fraud risk assessment score in Step 2 and the emotional state report in Step 3. In this step, it receives both analysis results as input and generates a careful and appropriate response message as output. For example, messages such as "Please stay calm" or "Please wait while I check the details" may be selected. Specifically, this involves the server generating and outputting a response message based on the analysis results.

[0157] Step 5:

[0158] The device receives a notification from the server and displays fraud prevention information to the user. The input for this step is the response message for step 4. A warning message such as "This may be a scam. Please be careful" is displayed on the device screen. Specifically, this involves a process in which the user visually obtains information via the notification function.

[0159] Step 6:

[0160] The server records suspicious communication destinations in its database and updates the information list. The input for this step is the suspicious communication destination information obtained through the analysis in step 2. The output is an updated blacklist. Specifically, this involves the server adding suspicious phone numbers to the database and strengthening the known list.

[0161] (Application Example 2)

[0162] 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 device 14 will be referred to as the "terminal."

[0163] In voice communication, the challenge lies in early detection of fraudulent activity and protecting users from fraud and stress while considering their emotional state. This technology needs to be used to provide a safer and more reliable communication environment. In addition to fraud detection, it is necessary to quickly detect the fear and anxiety users feel and respond appropriately.

[0164] 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.

[0165] In this invention, the server includes means for converting voice information into text information, means for analyzing the text information using a generative AI model to evaluate the suspiciousness of the communication and the user's emotional state, and means for generating a natural response when there is a high probability of fraud and when the user's stress state is detected. This makes it possible to simultaneously achieve rapid detection of fraudulent activity and ensure the psychological safety of the user.

[0166] "Voice information" refers to digital or analog input data of human voices collected through phone calls or recordings.

[0167] "Textual information" refers to data in text format obtained by processing audio data.

[0168] A "generative AI model" is an artificial intelligence algorithm that learns human language and patterns to derive new results.

[0169] "Suspicious communication" refers to unusual behavior or patterns in phone calls or conversations that suggest the possibility of fraud or illegal activity.

[0170] "User emotional state" refers to the emotional tendencies and psychological state that a user experiences during voice interaction.

[0171] "Stress level" refers to the degree of psychological or emotional pressure a user experiences during their daily tasks or interactions.

[0172] A "natural response" is a response generated to maintain smooth and human-like communication in interactions with users.

[0173] "Suspicious contact information" refers to contact information from sources where fraudulent or dishonest activity has been detected in past communications.

[0174] A "database" is a collection of digital information that systematically organizes information, enabling efficient data storage, management, and retrieval.

[0175] A "blacklist" is a list of registered information that restricts or prohibits certain actions or behaviors.

[0176] "Automatic blocking" is a system function that automatically restricts or prohibits communication based on pre-defined criteria or conditions.

[0177] "Fraud prevention information" refers to information that provides notifications and advice related to the detection and prevention of fraudulent activities.

[0178] "Advice" is a term that refers to instructions or suggestions that propose the best course of action in a particular situation.

[0179] To implement this invention, the server uses a program that converts voice information into text information. This program utilizes voice recognition software such as "Google Speech-to-Text API" and has the function of converting voice data into text data in real time. Next, the text data is passed to a generative AI model such as "OpenAI GPT-3 (registered trademark)" to evaluate the suspiciousness of the communication and the user's emotional state. This generative AI model not only analyzes the content of the call to assess the possibility of fraud, but also uses an emotion analysis algorithm to determine whether the user is feeling stressed.

[0180] The device is equipped with a function that receives notifications from the server and displays fraud prevention information and advice based on the user's emotional state. Notifications alert the user if there is a possibility of fraud or if the user is detected to be experiencing stress. If signs of fraud are detected, the server records suspicious contact information in a database such as "MongoDB" and updates the blacklist. This blacklist is managed through an automatic blocking function to limit further contact.

[0181] As an example of this system, consider a scenario where a user receives a fraudulent phone call impersonating a financial institution. The server analyzes the content of the call in real time, and if it determines that there is a high probability of fraudulent activity, it sends a notification to the device stating, "A potential scam has been detected. Please be careful." Furthermore, if the emotion engine determines that the user is in a high-stress state, it also displays advice such as, "Please stay calm and verify the information." An example of a prompt message to the generative AI model is, "Is the content of this call related to fraud? Please assess the user's stress level, taking into account the data from the emotion analysis engine."

[0182] This allows users to simultaneously achieve rapid detection of fraudulent activity and ensure psychological safety.

[0183] The flow of a specific process in Application Example 2 will be explained using Figure 14.

[0184] Step 1:

[0185] The server receives audio information. The user's phone call audio is provided as input to the server, and the server uses the "Google Speech-to-Text API" to convert this audio information into text information. This process generates text data from the audio data and obtains text-formatted output ready for subsequent analysis processes.

[0186] Step 2:

[0187] The server sends the generated text information to a generative AI model such as "OpenAI GPT-3". The model analyzes the call content and evaluates the suspiciousness of the communication from the input text. In this step, it generates a prompt message to determine whether it is potentially fraudulent, and the AI ​​model processes this information and outputs an evaluation result. For example, the evaluation result may include a judgment such as "Probability of fraud: High".

[0188] Step 3:

[0189] The server simultaneously uses an emotion analysis engine to evaluate the user's emotional state. It passes audio information as input to the emotion analysis engine, which then analyzes the audio data to assess the user's stress level. This step yields an output indicating whether the user is experiencing stress.

[0190] Step 4:

[0191] The server generates natural responses as needed, based on the suspicious nature of the call content and the user's emotional state. Using the evaluation results of the generation AI model and the emotional state as input, it creates response messages such as "A potential scam has been detected. Please be careful" or "Please stay calm and verify the information." These response messages are output using a natural language generation (NLG) algorithm.

[0192] Step 5:

[0193] The device receives a notification from the server and displays a generated response message to the user. Through this notification, the user can receive fraud warnings and stress reduction advice in real time. In this step, the device receives output from the server and displays the message on the screen.

[0194] Step 6:

[0195] The server records suspicious contact information in a database and updates the blacklist. It takes the contact information of the caller of a suspicious call as input and stores it in a database such as "MongoDB," automatically expanding the blacklist. This process outputs a setting that automatically blocks future communications from suspicious contacts.

[0196] 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.

[0197] Data generation model 58 is a 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.

[0198] 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.

[0199] [Second Embodiment]

[0200] Figure 3 shows an example of the configuration of the data processing system 210 according to the second embodiment.

[0201] 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.

[0202] 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).

[0203] 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.

[0204] 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.

[0205] 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).

[0206] 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.

[0207] 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.

[0208] 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.

[0209] 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.

[0210] 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.

[0211] 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".

[0212] As an embodiment of this invention, a system having the following functions is constructed: The server monitors the user's voice calls in real time and acquires voice data. The acquired voice data is converted into text data on the server by a speech recognition engine. This text is input into a generative AI model to analyze whether the call content is suspicious and evaluate the possibility of fraud.

[0213] The server receives feedback from the generative AI model and generates a natural response if it determines that the call is likely to be fraudulent. This allows the server to handle suspicious calls on behalf of the user. For example, the server can generate an appropriate response such as "Please wait while I check the invoice" to calm the caller.

[0214] On the other hand, the server records suspicious phone numbers in a database and periodically updates the blacklist. Each time the update is performed, the system is configured to block future incoming calls from the registered suspicious phone numbers. This prevents secondary damage from the same fraudsters.

[0215] Users will receive notifications via their devices when potentially fraudulent calls are detected, as well as periodic notifications about fraud prevention. These notifications allow users to review suspicious call content, helping to raise their security awareness.

[0216] This system can prevent fraudulent activities that target users who are particularly vulnerable to fraud, such as the elderly, and enable safe communication.

[0217] The following describes the processing flow.

[0218] Step 1:

[0219] The server receives user voice calls in real time. A speech recognition engine is used to convert the voice data into text data. This conversion is performed quickly to enable immediate analysis of the call content.

[0220] Step 2:

[0221] The server inputs the converted text data into a generating AI model. The AI ​​model identifies suspicious patterns and phrases from the text and assesses the likelihood of fraud. Based on the assessment results, it quantifies the risk of fraud.

[0222] Step 3:

[0223] If the server detects a high probability of fraud, it prepares a natural-sounding response generated by an AI model. This response follows the flow of human conversation and is intended to delay or mislead the caller of a fraudulent call.

[0224] Step 4:

[0225] The server records phone numbers deemed suspicious in its database. This updates the blacklist, preparing to automatically block future calls from that number.

[0226] Step 5:

[0227] The device will send a notification to the user if it detects a potentially fraudulent call. This notification will include details of the call and a warning about the scam, helping to raise user awareness.

[0228] Step 6:

[0229] Users receive notifications from their devices and take appropriate measures to respond to potentially fraudulent calls. This gives them an opportunity to avoid trouble before it happens.

[0230] (Example 1)

[0231] 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."

[0232] In modern communication technology, users are vulnerable to fraud, and there is a lack of means to quickly identify suspicious calls and identification numbers. Effective measures are needed to reduce fraud, especially for the elderly and tech-savvy users. The challenge lies in achieving a safe communication environment.

[0233] 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.

[0234] In this invention, the server includes means for converting voice information into text information, means for analyzing the text information using a generative information processing model and evaluating the suspiciousness of the communication, and means for recording suspicious identification numbers in an information storage device and updating a restriction list. This makes it possible for users to receive advance warnings about suspicious calls and prevent fraud.

[0235] "Auditory information" refers to data related to sounds such as human speech that are transmitted as sound waves.

[0236] "Textual information" refers to data obtained by converting audio information into text, and is used in natural language processing.

[0237] A "generative information processing model" refers to a trained artificial intelligence technology used for natural language generation and analysis.

[0238] "Means for evaluating suspicious activity" refers to a function that analyzes communication content and patterns to detect suspicious activity that differs from normal communication.

[0239] An "identification number" is a combination of numbers or characters used to identify the caller.

[0240] An "information storage device" is a device or system for recording and storing data temporarily or for a long period of time.

[0241] A "restriction list" is a list used to register suspicious communication partners and restrict or prohibit future contact.

[0242] "User" refers to an individual or organization that uses this system.

[0243] To implement this invention, it is necessary to construct a system having the following elements: The server monitors the user's voice calls in real time using a communication infrastructure. To acquire voice, voice streaming technology is used to collect voice information, and a speech recognition engine is used to convert it into text information. Specifically, a speech recognition API (e.g., a cloud-based service) can be used for speech recognition.

[0244] The server inputs the converted character information into a generative information processing model and uses the model's analytical capabilities to evaluate the suspiciousness of the communication. This generative information processing model uses a natural language processing model based on existing technologies (e.g., a large-scale language model). Based on the analysis results, the server generates a natural response and, if necessary, deals with suspicious calls.

[0245] Furthermore, the server has a mechanism to prevent similar suspicious communication activities by storing suspicious identification numbers in an information storage device and updating the restriction list. The terminal has a notification function related to fraud prevention, providing users with security-related information in real time.

[0246] For example, if a user prompts, "Have I had any suspicious calls recently?", the server scans recent call logs and notifies the user of details about any calls suspected of being fraudulent. In this way, users can more securely manage their own calling environment. This system helps combat fraudulent activities and provides a safe and secure communication experience.

[0247] The flow of the specific processing in Example 1 will be explained using Figure 11.

[0248] Step 1:

[0249] The server acquires voice data from a user who initiates a voice call in real time. An audio stream is provided as input. The server transfers this audio data to a speech recognition engine, where it is converted into text data. The output is a string of characters reflecting the content of the call. This process involves accumulating the audio data in a buffer and inputting it to the speech recognition engine in batch processing.

[0250] Step 2:

[0251] The server inputs the obtained text data into the generative AI model. The string generated in the previous step is used as input. The generative AI model analyzes this string and evaluates whether the communication content is suspicious. The results of this analysis are scored, and the evaluation result is obtained as output. Specifically, this involves the model analyzing text patterns and quantifying the likelihood of fraud.

[0252] Step 3:

[0253] The server generates a natural response if the evaluation result exceeds a threshold. In this process, it receives an evaluation score from a generative AI model as input and constructs a response using a response generation engine. The output is the response that should be sent to the caller. The operation includes creating the optimal response by referring to past conversation data and preset templates.

[0254] Step 4:

[0255] The server records suspicious identification numbers in a database. The input is the identification numbers detected during a call, which are stored in the information storage device. The output is an updated restricted list. The operation includes a mechanism where new identification numbers are added to the existing list, and this information is periodically transmitted to other user terminals.

[0256] Step 5:

[0257] The device notifies the user of fraud prevention information. Here, it receives a warning message from the server as input and displays it in the user interface. The output is a warning message that the user can review. Specifically, this includes actions such as a notification popping up on the screen and providing detailed information as needed.

[0258] (Application Example 1)

[0259] 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."

[0260] In recent years, fraudulent activities have become a serious social problem, particularly telephone scams targeting the elderly, and effective preventative measures are needed. The problem is that many users continue calls without fully understanding the risks of fraud. Furthermore, a system is needed that can immediately detect communications with a high probability of being fraudulent and warn users.

[0261] 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.

[0262] In this invention, the server includes means for converting acoustic information into text information, means for analyzing the text information using a generation AI model and evaluating the suspiciousness of the communication, and means for displaying a warning for suspicious communication and providing an option to terminate the call. This makes it possible to immediately detect communications that are highly likely to be fraudulent and to help users make safe decisions.

[0263] "Means for converting acoustic information into text information" refers to a device or program that performs the process of converting audio data into text data in real time.

[0264] "Means for analyzing the aforementioned textual information using a generative AI model and evaluating the suspiciousness of the communication" refers to a device or program that uses artificial intelligence to analyze textual information and determine whether the content of the communication is fraudulent.

[0265] "Means for generating an automated response when a communication is deemed highly likely to be fraudulent" refers to a device or program that automatically generates an appropriate response when it detects a communication that is strongly suspected of being fraudulent.

[0266] "Means for recording suspicious identification information in a data storage device and updating the restriction list" refers to a device or program that stores suspicious phone numbers or identification information in a database and periodically updates the restriction list.

[0267] "Means of notifying users of fraud prevention information" refers to a device or program that notifies the user's terminal of information regarding the risk of fraud.

[0268] "Means for displaying a warning about suspicious communications and providing an option to terminate a call" refers to a device or program that, upon detecting suspicious communications, presents a warning to the user and provides an interface for choosing whether to continue or terminate the call.

[0269] The system for implementing this invention primarily involves the processes of acquiring, analyzing, responding to, and recording voice data. First, the terminal immediately transmits the received call audio as acoustic information to the server. The server converts this acoustic information into text information using the Google Cloud Speech-to-Text API. The converted text information is then analyzed by a generative AI model, in this case OpenAI GPT, to assess the likelihood of fraud.

[0270] If the server determines that there is a high probability of fraud, it will automatically generate an appropriate response message. This response message can instruct the other party, such as "We are checking, please wait a moment," in a way that does not make the call seem unnatural.

[0271] The server also records suspicious identification information, such as phone numbers, in its database. This data is added to a restricted list, and settings are made to block future communications. Furthermore, users are displayed a warning about the risk of fraud through their device and given the option to continue or end the call. This feature allows users to ensure their own safety through their own judgment.

[0272] For example, if an elderly user receives a phone call from a stranger requesting personal information, this system will respond quickly. After the audio is converted into text on the server, the user will be warned that "this call may be a scam," allowing for a swift response. An example of a prompt to the generating AI model is, "Does the following text contain signs of fraud? Please tell me if it is a scam and why."

[0273] The flow of a specific process in Application Example 1 will be explained using Figure 12.

[0274] Step 1:

[0275] The terminal instantly transmits the call audio received by the user to the server. The input is the audio data obtained from the call, and the output is the transmission of this audio data to the server. The terminal's operation involves digitizing the received audio signal and transferring it to the server using a secure communication protocol.

[0276] Step 2:

[0277] The server uses the Google Cloud Speech-to-Text API to convert the received audio data into text. The input is the audio data received in step 1, and the output is the converted text. In this step, a speech recognition algorithm is applied, the audio waveform is analyzed within the server, and the corresponding text is converted.

[0278] Step 3:

[0279] The server uses the generative AI model OpenAI GPT to analyze the text information and evaluate the suspiciousness of the communication. The input is the text information obtained in Step 2, and the output is the evaluation result regarding the likelihood of fraud. In this step, using the prompt sentence "Are there any signs of fraud in the following text? Please tell me the likelihood of fraud and the reasons.", the text information is input into the generative AI model to quantify the risk of fraud.

[0280] Step 4:

[0281] If it is determined that the likelihood of fraud is high, the server generates an automated response message. The input is the evaluation result obtained in Step 3, and the output is the generated response message. The server uses the generative AI model to generate a response sentence such as "Please wait a moment while we are checking" to calm down the other party.

[0282] Step 5:

[0283] The server records the suspicious identification information in the data storage device and updates the restricted list. The input is the identification information of the call suspected of fraud, and the output is the updated restricted list. The server writes this identification information into the database and automatically sets to block future calls.

[0284] Step 6:

[0285] A warning message regarding the risk of fraud is displayed to the user through the terminal. The input is the evaluation result from Step 3, and the output is a warning message that the user can view on the screen. The terminal presents the risk of continuing the call and the termination option to the user, and enables the user to end the call at will if necessary.

[0286] 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 recognition model 59 and perform specific processing using the user's emotion.

[0287] As an embodiment of this invention, a system incorporating a server, a terminal, and an emotion engine is provided. This system monitors the user's voice calls and processes the data acquired by speech recognition in real time. The server converts the voice data into text data and performs analysis using a generative AI model and an emotion engine.

[0288] The generative AI model analyzes the call content and assesses the likelihood of fraud. The emotion engine recognizes emotions from the user's voice and evaluates their state in real time. Based on this information, the server determines the appropriate course of action, considering not only the risk of fraud but also the psychological impact the call has on the user.

[0289] Specifically, the server generates more cautious responses when it determines there is a high probability of fraud, as well as when it assesses that the user is in a stressful state. For example, it can use phrases such as "Please calm down" or "Please wait while I check the details" to reduce the user's stress while the call progresses.

[0290] Furthermore, the server records suspicious phone numbers in a database. User safety is ensured by continuously updating the blacklist and automatically blocking incoming calls from known fraudulent phone numbers.

[0291] The device will notify users of potentially fraudulent calls and information about the user's emotional state. This allows users to take appropriate action. For example, a notification such as "A potential scam has been detected. Please be careful" will appear on the device.

[0292] This system allows users, including the elderly, to avoid the risk of fraud while reducing psychological burden. By combining it with emotional recognition, it goes beyond mere technical responses and enables activities that are empathetic to the user's feelings.

[0293] The following describes the processing flow.

[0294] Step 1:

[0295] The server receives user voice calls in real time. A speech recognition engine is used to convert the voice data into text data. This process extracts the call content, enabling immediate analysis.

[0296] Step 2:

[0297] The server inputs the converted text data into a generating AI model. This model analyzes the call content and assesses the likelihood of fraud. Specifically, it searches for matches with past fraud patterns and calculates the fraud risk.

[0298] Step 3:

[0299] The server inputs voice data from the user into an emotion engine to analyze the user's emotional state. This engine evaluates the user's stress and anxiety levels based on the tone and pace of their voice.

[0300] Step 4:

[0301] The server integrates the evaluation of the generative AI model and the analysis results of the emotion engine. If the risk of fraud is high and the user's emotional state is unstable, it generates a cautious response. This response includes words that soothe the user's emotions.

[0302] Step 5:

[0303] The server records suspicious phone numbers in its database. It updates the blacklist and configures the system to automatically block incoming calls from those numbers.

[0304] Step 6:

[0305] The device sends notifications to the user based on the possibility of fraud and their emotional state. These notifications may include messages such as, "Fraud has been detected. Please stay calm."

[0306] Step 7:

[0307] The user can check the notification from the terminal and take measures according to the details of the call content and the analysis results of the emotional state. This can prevent fraud damage and achieve emotional stability.

[0308] (Example 2)

[0309] Next, Example 2 will be described. In the following description, the data processing device 12 is referred to as a "server", and the smart glasses 214 are referred to as a "terminal".

[0310] In recent years, an effective system for protecting users from increasing fraud through telephone calls has been demanded. In particular, for users including the elderly, it is necessary to quickly evaluate the risk of fraud and provide a response that reduces the psychological burden. However, with existing technologies, it is difficult to provide a comprehensive response that simultaneously considers fraud risk assessment and psychological impact. To solve such problems, a system that evaluates the suspiciousness of communication through real-time voice analysis and generates an appropriate response considering the emotional state is required.

[0311] The specific processing by the specific processing unit 290 of the data processing device 12 in Example 2 is realized by the following means.

[0312] In this invention, the server includes means for converting voice data into text data, means for analyzing the text data using a generated AI model to evaluate the suspiciousness of communication, and means for evaluating the user's emotional state in real time and determining a call response method based on that information. This makes it possible to quickly evaluate communications with a high likelihood of fraud and provide a response that also takes into account the user's emotions, thereby preventing communication troubles and ensuring the safety and psychological comfort of the user.

[0313] "Methods for converting audio data into text data" refers to technologies that use speech recognition technology to convert audio into strings of characters, creating a text format that can be used for subsequent data analysis and processing.

[0314] "Means for analyzing the text data using a generative AI model and evaluating the suspiciousness of the communication" refers to a technology that uses an AI model trained through machine learning to analyze text data and evaluate the risk of fraud based on its content.

[0315] "A means of evaluating a user's emotional state in real time and determining how to respond to a call based on that information" refers to a technology that recognizes emotions from a user's voice data and determines an appropriate response or action based on the results.

[0316] "A means of recording suspicious communication destinations in a database and updating the information list" refers to a technology that identifies potentially fraudulent communication destinations, records them in a database, and updates the list to improve the security of future communications.

[0317] "Means of notifying terminals of fraud prevention information" refers to technology that provides a function to display warnings and alerts on the user's terminal device regarding the possibility of detected fraud.

[0318] "Means of generating careful responses that take into account the impact on the user" refers to technologies that construct responses based on the user's emotional state and fraud risk assessment to provide accurate information while reducing psychological burden.

[0319] In an embodiment of the present invention, the system is designed to analyze voice calls in real time, assess the risk of fraud and the user's emotional state, and generate appropriate responses. The main components of the system include a server, a terminal, a speech recognition engine, a generative AI model, and an emotion engine.

[0320] The server uses a speech recognition engine to convert the user's voice call data into text data. This process utilizes commonly used speech recognition technologies such as the Google Cloud Speech-to-Text API, enabling highly accurate speech-to-text conversion.

[0321] Subsequently, the server analyzes the converted text data using a generative AI model to evaluate the suspiciousness of the communication. It is expected that OpenAI's GPT series will be used as the generative AI model. The generative AI model analyzes the text data based on the prompt message. A specific example of a prompt message is, "Is this call potentially fraudulent? Please assess the risk of fraud."

[0322] Simultaneously, the server uses an emotion engine to evaluate the user's emotional state in real time. This emotion engine utilizes tools such as Affectiva, which provides Emotion AI technology. Through this process, the server determines whether the user is experiencing stress or anxiety, and uses this information to improve call handling.

[0323] The device displays fraud prevention-related information to the user based on notifications from the server. The notifications displayed on the device are warning messages such as "This may be a scam. Please be careful," and are designed to help the user take appropriate action immediately.

[0324] This system protects users from fraudulent activities and allows them to communicate with peace of mind. It also reduces psychological burden and ensures safety for users unfamiliar with digital devices, such as the elderly. By providing support from both a technological and emotional perspective, a high-quality user experience is achieved.

[0325] The flow of the specific processing in Example 2 will be explained using Figure 13.

[0326] Step 1:

[0327] The server receives the user's voice call. It receives voice data as input and analyzes the data using a speech recognition engine. It converts the voice data into text data in real time using APIs such as the Google Cloud Speech-to-Text API. This results in text data as output. Specifically, this process involves the server sending the voice file to the speech recognition API.

[0328] Step 2:

[0329] The server inputs the text data obtained in Step 1 into the generating AI model. The input at this stage is the text data that is the output of Step 1. The generating AI model uses OpenAI's GPT series and performs text analysis based on the prompt sentence. For the purpose of evaluating fraud risk, a prompt sentence such as "Is this call potentially fraudulent? Please evaluate the fraud risk." is used. As a result of the analysis, a fraud risk evaluation score is obtained as output. Specifically, the process involves the server sending the text data to the AI ​​model and performing the analysis.

[0330] Step 3:

[0331] The server passes the audio data to the emotion engine to analyze the user's emotional state. The input at this stage is the audio data handled in step 1. Using emotion recognition technology such as Affectiva, the user's emotions are evaluated in real time. The output is an emotional state report indicating whether the user is stressed or not. The specific operation involves the server inputting audio data into the emotion engine and obtaining the analysis results.

[0332] Step 4:

[0333] The server determines how to respond to the call based on the fraud risk assessment score in Step 2 and the emotional state report in Step 3. In this step, it receives both analysis results as input and generates a careful and appropriate response message as output. For example, messages such as "Please stay calm" or "Please wait while I check the details" may be selected. Specifically, this involves the server generating and outputting a response message based on the analysis results.

[0334] Step 5:

[0335] The device receives a notification from the server and displays fraud prevention information to the user. The input for this step is the response message for step 4. A warning message such as "This may be a scam. Please be careful" is displayed on the device screen. Specifically, this involves a process in which the user visually obtains information via the notification function.

[0336] Step 6:

[0337] The server records suspicious communication destinations in its database and updates the information list. The input for this step is the suspicious communication destination information obtained through the analysis in step 2. The output is an updated blacklist. Specifically, this involves the server adding suspicious phone numbers to the database and strengthening the known list.

[0338] (Application Example 2)

[0339] 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."

[0340] In voice communication, the challenge lies in early detection of fraudulent activity and protecting users from fraud and stress while considering their emotional state. This technology needs to be used to provide a safer and more reliable communication environment. In addition to fraud detection, it is necessary to quickly detect the fear and anxiety users feel and respond appropriately.

[0341] 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.

[0342] In this invention, the server includes means for converting voice information into text information, means for analyzing the text information using a generative AI model to evaluate the suspiciousness of the communication and the user's emotional state, and means for generating a natural response when there is a high probability of fraud and when the user's stress state is detected. This makes it possible to simultaneously achieve rapid detection of fraudulent activity and ensure the psychological safety of the user.

[0343] "Voice information" refers to digital or analog input data of human voices collected through phone calls or recordings.

[0344] "Textual information" refers to data in text format obtained by processing audio data.

[0345] A "generative AI model" is an artificial intelligence algorithm that learns human language and patterns to derive new results.

[0346] "Suspicious communication" refers to unusual behavior or patterns in phone calls or conversations that suggest the possibility of fraud or illegal activity.

[0347] "User emotional state" refers to the emotional tendencies and psychological state that a user experiences during voice interaction.

[0348] "Stress level" refers to the degree of psychological or emotional pressure a user experiences during their daily tasks or interactions.

[0349] A "natural response" is a response generated to maintain smooth and human-like communication in interactions with users.

[0350] "Suspicious contact information" refers to contact information from sources where fraudulent or dishonest activity has been detected in past communications.

[0351] A "database" is a collection of digital information that systematically organizes information, enabling efficient data storage, management, and retrieval.

[0352] A "blacklist" is a list of registered information that restricts or prohibits certain actions or behaviors.

[0353] "Automatic blocking" is a system function that automatically restricts or prohibits communication based on pre-defined criteria or conditions.

[0354] "Fraud prevention information" refers to information that provides notifications and advice related to the detection and prevention of fraudulent activities.

[0355] "Advice" is a term that refers to instructions or suggestions that propose the best course of action in a particular situation.

[0356] To implement this invention, the server uses a program that converts voice information into text information. This program utilizes voice recognition software such as "Google Speech-to-Text API" and has the function of converting voice data into text data in real time. Next, the text data is passed to a generative AI model such as "OpenAI GPT-3" to evaluate the suspiciousness of the communication and the user's emotional state. This generative AI model not only analyzes the content of the call to assess the possibility of fraud, but also uses an emotion analysis algorithm to determine whether the user is feeling stressed.

[0357] The device is equipped with a function that receives notifications from the server and displays fraud prevention information and advice based on the user's emotional state. Notifications alert the user if there is a possibility of fraud or if the user is detected to be experiencing stress. If signs of fraud are detected, the server records suspicious contact information in a database such as "MongoDB" and updates the blacklist. This blacklist is managed through an automatic blocking function to limit further contact.

[0358] As an example of this system, consider a scenario where a user receives a fraudulent phone call impersonating a financial institution. The server analyzes the content of the call in real time, and if it determines that there is a high probability of fraudulent activity, it sends a notification to the device stating, "A potential scam has been detected. Please be careful." Furthermore, if the emotion engine determines that the user is in a high-stress state, it also displays advice such as, "Please stay calm and verify the information." An example of a prompt message to the generative AI model is, "Is the content of this call related to fraud? Please assess the user's stress level, taking into account the data from the emotion analysis engine."

[0359] This allows users to simultaneously achieve rapid detection of fraudulent activity and ensure psychological safety.

[0360] The flow of a specific process in Application Example 2 will be explained using Figure 14.

[0361] Step 1:

[0362] The server receives audio information. The user's phone call audio is provided as input to the server, and the server uses the "Google Speech-to-Text API" to convert this audio information into text information. This process generates text data from the audio data and obtains text-formatted output ready for subsequent analysis processes.

[0363] Step 2:

[0364] The server sends the generated text information to a generative AI model such as "OpenAI GPT-3". The model analyzes the call content and evaluates the suspiciousness of the communication from the input text. In this step, it generates a prompt message to determine whether it is potentially fraudulent, and the AI ​​model processes this information and outputs an evaluation result. For example, the evaluation result may include a judgment such as "Probability of fraud: High".

[0365] Step 3:

[0366] The server simultaneously uses an emotion analysis engine to evaluate the user's emotional state. It passes audio information as input to the emotion analysis engine, which then analyzes the audio data to assess the user's stress level. This step yields an output indicating whether the user is experiencing stress.

[0367] Step 4:

[0368] The server generates natural responses as needed, based on the suspicious nature of the call content and the user's emotional state. Using the evaluation results of the generation AI model and the emotional state as input, it creates response messages such as "A potential scam has been detected. Please be careful" or "Please stay calm and verify the information." These response messages are output using a natural language generation (NLG) algorithm.

[0369] Step 5:

[0370] The device receives a notification from the server and displays a generated response message to the user. Through this notification, the user can receive fraud warnings and stress reduction advice in real time. In this step, the device receives output from the server and displays the message on the screen.

[0371] Step 6:

[0372] The server records suspicious contact information in a database and updates the blacklist. It takes the contact information of the caller of a suspicious call as input and stores it in a database such as "MongoDB," automatically expanding the blacklist. This process outputs a setting that automatically blocks future communications from suspicious contacts.

[0373] 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.

[0374] 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.

[0375] 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.

[0376] [Third Embodiment]

[0377] Figure 5 shows an example of the configuration of the data processing system 310 according to the third embodiment.

[0378] 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.

[0379] 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).

[0380] 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.

[0381] 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.

[0382] 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).

[0383] 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.

[0384] 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.

[0385] 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.

[0386] 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.

[0387] 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.

[0388] 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".

[0389] As an embodiment of this invention, a system having the following functions is constructed: The server monitors the user's voice calls in real time and acquires voice data. The acquired voice data is converted into text data on the server by a speech recognition engine. This text is input into a generative AI model to analyze whether the call content is suspicious and evaluate the possibility of fraud.

[0390] The server receives feedback from the generative AI model and generates a natural response if it determines that the call is likely to be fraudulent. This allows the server to handle suspicious calls on behalf of the user. For example, the server can generate an appropriate response such as "Please wait while I check the invoice" to calm the caller.

[0391] On the other hand, the server records suspicious phone numbers in a database and periodically updates the blacklist. Each time the update is performed, the system is configured to block future incoming calls from the registered suspicious phone numbers. This prevents secondary damage from the same fraudsters.

[0392] Users will receive notifications via their devices when potentially fraudulent calls are detected, as well as periodic notifications about fraud prevention. These notifications allow users to review suspicious call content, helping to raise their security awareness.

[0393] This system can prevent fraudulent activities that target users who are particularly vulnerable to fraud, such as the elderly, and enable safe communication.

[0394] The following describes the processing flow.

[0395] Step 1:

[0396] The server receives user voice calls in real time. A speech recognition engine is used to convert the voice data into text data. This conversion is performed quickly to enable immediate analysis of the call content.

[0397] Step 2:

[0398] The server inputs the converted text data into a generating AI model. The AI ​​model identifies suspicious patterns and phrases from the text and assesses the likelihood of fraud. Based on the assessment results, it quantifies the risk of fraud.

[0399] Step 3:

[0400] If the server detects a high probability of fraud, it prepares a natural-sounding response generated by an AI model. This response follows the flow of human conversation and is intended to delay or mislead the caller of a fraudulent call.

[0401] Step 4:

[0402] The server records phone numbers deemed suspicious in its database. This updates the blacklist, preparing to automatically block future calls from that number.

[0403] Step 5:

[0404] The device will send a notification to the user if it detects a potentially fraudulent call. This notification will include details of the call and a warning about the scam, helping to raise user awareness.

[0405] Step 6:

[0406] Users receive notifications from their devices and take appropriate measures to respond to potentially fraudulent calls. This gives them an opportunity to avoid trouble before it happens.

[0407] (Example 1)

[0408] 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."

[0409] In modern communication technology, users are vulnerable to fraud, and there is a lack of means to quickly identify suspicious calls and identification numbers. Effective measures are needed to reduce fraud, especially for the elderly and tech-savvy users. The challenge lies in achieving a safe communication environment.

[0410] 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.

[0411] In this invention, the server includes means for converting voice information into text information, means for analyzing the text information using a generative information processing model and evaluating the suspiciousness of the communication, and means for recording suspicious identification numbers in an information storage device and updating a restriction list. This makes it possible for users to receive advance warnings about suspicious calls and prevent fraud.

[0412] "Auditory information" refers to data related to sounds such as human speech that are transmitted as sound waves.

[0413] "Textual information" refers to data obtained by converting audio information into text, and is used in natural language processing.

[0414] A "generative information processing model" refers to a trained artificial intelligence technology used for natural language generation and analysis.

[0415] "Means for evaluating suspicious activity" refers to a function that analyzes communication content and patterns to detect suspicious activity that differs from normal communication.

[0416] An "identification number" is a combination of numbers or characters used to identify the caller.

[0417] An "information storage device" is a device or system for recording and storing data temporarily or for a long period of time.

[0418] A "restriction list" is a list used to register suspicious communication partners and restrict or prohibit future contact.

[0419] "User" refers to an individual or organization that uses this system.

[0420] To implement this invention, it is necessary to construct a system having the following elements: The server monitors the user's voice calls in real time using a communication infrastructure. To acquire voice, voice streaming technology is used to collect voice information, and a speech recognition engine is used to convert it into text information. Specifically, a speech recognition API (e.g., a cloud-based service) can be used for speech recognition.

[0421] The server inputs the converted character information into a generative information processing model and uses the model's analytical capabilities to evaluate the suspiciousness of the communication. This generative information processing model uses a natural language processing model based on existing technologies (e.g., a large-scale language model). Based on the analysis results, the server generates a natural response and, if necessary, deals with suspicious calls.

[0422] Furthermore, the server has a mechanism to prevent similar suspicious communication activities by storing suspicious identification numbers in an information storage device and updating the restriction list. The terminal has a notification function related to fraud prevention, providing users with security-related information in real time.

[0423] For example, if a user prompts, "Have I had any suspicious calls recently?", the server scans recent call logs and notifies the user of details about any calls suspected of being fraudulent. In this way, users can more securely manage their own calling environment. This system helps combat fraudulent activities and provides a safe and secure communication experience.

[0424] The flow of the specific processing in Example 1 will be explained using Figure 11.

[0425] Step 1:

[0426] The server acquires voice data from a user who initiates a voice call in real time. An audio stream is provided as input. The server transfers this audio data to a speech recognition engine, where it is converted into text data. The output is a string of characters reflecting the content of the call. This process involves accumulating the audio data in a buffer and inputting it to the speech recognition engine in batch processing.

[0427] Step 2:

[0428] The server inputs the obtained text data into the generative AI model. The string generated in the previous step is used as input. The generative AI model analyzes this string and evaluates whether the communication content is suspicious. The results of this analysis are scored, and the evaluation result is obtained as output. Specifically, this involves the model analyzing text patterns and quantifying the likelihood of fraud.

[0429] Step 3:

[0430] The server generates a natural response if the evaluation result exceeds a threshold. In this process, it receives an evaluation score from a generative AI model as input and constructs a response using a response generation engine. The output is the response that should be sent to the caller. The operation includes creating the optimal response by referring to past conversation data and preset templates.

[0431] Step 4:

[0432] The server records suspicious identification numbers in a database. The input is the identification numbers detected during a call, which are stored in the information storage device. The output is an updated restricted list. The operation includes a mechanism where new identification numbers are added to the existing list, and this information is periodically transmitted to other user terminals.

[0433] Step 5:

[0434] The device notifies the user of fraud prevention information. Here, it receives a warning message from the server as input and displays it in the user interface. The output is a warning message that the user can review. Specifically, this includes actions such as a notification popping up on the screen and providing detailed information as needed.

[0435] (Application Example 1)

[0436] 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."

[0437] In recent years, fraudulent activities have become a serious social problem, particularly telephone scams targeting the elderly, and effective preventative measures are needed. The problem is that many users continue calls without fully understanding the risks of fraud. Furthermore, a system is needed that can immediately detect communications with a high probability of being fraudulent and warn users.

[0438] 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.

[0439] In this invention, the server includes means for converting acoustic information into text information, means for analyzing the text information using a generation AI model and evaluating the suspiciousness of the communication, and means for displaying a warning for suspicious communication and providing an option to terminate the call. This makes it possible to immediately detect communications that are highly likely to be fraudulent and to help users make safe decisions.

[0440] "Means for converting acoustic information into text information" refers to a device or program that performs the process of converting audio data into text data in real time.

[0441] "Means for analyzing the aforementioned textual information using a generative AI model and evaluating the suspiciousness of the communication" refers to a device or program that uses artificial intelligence to analyze textual information and determine whether the content of the communication is fraudulent.

[0442] "Means for generating an automated response when a communication is deemed highly likely to be fraudulent" refers to a device or program that automatically generates an appropriate response when it detects a communication that is strongly suspected of being fraudulent.

[0443] "Means for recording suspicious identification information in a data storage device and updating the restriction list" refers to a device or program that stores suspicious phone numbers or identification information in a database and periodically updates the restriction list.

[0444] "Means of notifying users of fraud prevention information" refers to a device or program that notifies the user's terminal of information regarding the risk of fraud.

[0445] "Means for displaying a warning about suspicious communications and providing an option to terminate a call" refers to a device or program that, upon detecting suspicious communications, presents a warning to the user and provides an interface for choosing whether to continue or terminate the call.

[0446] The system for implementing this invention primarily involves the processes of acquiring, analyzing, responding to, and recording voice data. First, the terminal immediately transmits the received call audio as acoustic information to the server. The server converts this acoustic information into text information using the Google Cloud Speech-to-Text API. The converted text information is then analyzed by a generative AI model, in this case OpenAI GPT, to assess the likelihood of fraud.

[0447] If the server determines that there is a high probability of fraud, it will automatically generate an appropriate response message. This response message can instruct the other party, such as "We are checking, please wait a moment," in a way that does not make the call seem unnatural.

[0448] The server also records suspicious identification information, such as phone numbers, in its database. This data is added to a restricted list, and settings are made to block future communications. Furthermore, users are displayed a warning about the risk of fraud through their device and given the option to continue or end the call. This feature allows users to ensure their own safety through their own judgment.

[0449] For example, if an elderly user receives a phone call from a stranger requesting personal information, this system will respond quickly. After the audio is converted into text on the server, the user will be warned that "this call may be a scam," allowing for a swift response. An example of a prompt to the generating AI model is, "Does the following text contain signs of fraud? Please tell me if it is a scam and why."

[0450] The flow of a specific process in Application Example 1 will be explained using Figure 12.

[0451] Step 1:

[0452] The terminal instantly transmits the call audio received by the user to the server. The input is the audio data obtained from the call, and the output is the transmission of this audio data to the server. The terminal's operation involves digitizing the received audio signal and transferring it to the server using a secure communication protocol.

[0453] Step 2:

[0454] The server uses the Google Cloud Speech-to-Text API to convert the received audio data into text. The input is the audio data received in step 1, and the output is the converted text. In this step, a speech recognition algorithm is applied, the audio waveform is analyzed within the server, and the corresponding text is converted.

[0455] Step 3:

[0456] The server uses the OpenAI GPT generative AI model to analyze textual information and evaluate the suspiciousness of the communication. The input is the textual information obtained in step 2, and the output is the evaluation result regarding the likelihood of fraud. In this step, the prompt "Does the following text contain signs of fraud? Please tell us the likelihood of fraud and why." is used to input textual information into the generative AI model and quantify the risk of fraud.

[0457] Step 4:

[0458] If the server determines that there is a high probability of fraud, it generates an automated response message. The input is the evaluation result obtained in step 3, and the output is the generated response message. The server uses a generation AI model to generate a calming response such as "We are checking, please wait a moment."

[0459] Step 5:

[0460] The server records suspicious identification information in a data storage device and updates the restriction list. The input is the identification information of a call suspected of being fraudulent, and the output is the updated restriction list. The server writes this identification information to a database and automatically configures settings to block future calls.

[0461] Step 6:

[0462] The user will see a warning message about the risk of fraud through their device. The input is the evaluation result from step 3, and the output is the warning message that the user can see on the screen. The device presents the user with the risks of continuing the call and the option to end it, allowing the user to end the call at their own discretion if necessary.

[0463] 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.

[0464] As an embodiment of this invention, a system incorporating a server, a terminal, and an emotion engine is provided. This system monitors the user's voice calls and processes the data acquired by speech recognition in real time. The server converts the voice data into text data and performs analysis using a generative AI model and an emotion engine.

[0465] The generative AI model analyzes the call content and assesses the likelihood of fraud. The emotion engine recognizes emotions from the user's voice and evaluates their state in real time. Based on this information, the server determines the appropriate course of action, considering not only the risk of fraud but also the psychological impact the call has on the user.

[0466] Specifically, the server generates more cautious responses when it determines there is a high probability of fraud, as well as when it assesses that the user is in a stressful state. For example, it can use phrases such as "Please calm down" or "Please wait while I check the details" to reduce the user's stress while the call progresses.

[0467] Furthermore, the server records suspicious phone numbers in a database. User safety is ensured by continuously updating the blacklist and automatically blocking incoming calls from known fraudulent phone numbers.

[0468] The device will notify users of potentially fraudulent calls and information about the user's emotional state. This allows users to take appropriate action. For example, a notification such as "A potential scam has been detected. Please be careful" will appear on the device.

[0469] This system allows users, including the elderly, to avoid the risk of fraud while reducing psychological burden. By combining it with emotional recognition, it goes beyond mere technical responses and enables activities that are empathetic to the user's feelings.

[0470] The following describes the processing flow.

[0471] Step 1:

[0472] The server receives user voice calls in real time. A speech recognition engine is used to convert the voice data into text data. This process extracts the call content, enabling immediate analysis.

[0473] Step 2:

[0474] The server inputs the converted text data into a generating AI model. This model analyzes the call content and assesses the likelihood of fraud. Specifically, it searches for matches with past fraud patterns and calculates the fraud risk.

[0475] Step 3:

[0476] The server inputs voice data from the user into an emotion engine to analyze the user's emotional state. This engine evaluates the user's stress and anxiety levels based on the tone and pace of their voice.

[0477] Step 4:

[0478] The server integrates the evaluation of the generative AI model and the analysis results of the emotion engine. If the risk of fraud is high and the user's emotional state is unstable, it generates a cautious response. This response includes words that soothe the user's emotions.

[0479] Step 5:

[0480] The server records suspicious phone numbers in its database. It updates the blacklist and configures the system to automatically block incoming calls from those numbers.

[0481] Step 6:

[0482] The device sends notifications to the user based on the possibility of fraud and their emotional state. These notifications may include messages such as, "Fraud has been detected. Please stay calm."

[0483] Step 7:

[0484] Users can check notifications from their devices and take action based on the analysis of call details and emotional state. This helps prevent fraud and promotes emotional stability.

[0485] (Example 2)

[0486] 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."

[0487] In recent years, there has been a growing need for effective systems to protect users from telephone-based fraud. In particular, for users, including the elderly, it is necessary to quickly assess the risk of fraud and provide responses that reduce psychological burden. However, existing technologies make it difficult to provide a comprehensive response that simultaneously considers fraud risk assessment and psychological impact. To solve this problem, a system is needed that evaluates the suspiciousness of communications through real-time voice analysis and generates appropriate responses that take into account the user's emotional state.

[0488] 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.

[0489] In this invention, the server includes means for converting voice data into text data, means for analyzing the text data using a generative AI model and evaluating the suspiciousness of the communication, and means for evaluating the user's emotional state in real time and determining how to respond to the call based on that information. This makes it possible to quickly evaluate communications that are likely to be fraudulent and provide responses that take the user's emotions into consideration, thereby preventing communication troubles and ensuring the user's safety and psychological security.

[0490] "Methods for converting audio data into text data" refers to technologies that use speech recognition technology to convert audio into strings of characters, creating a text format that can be used for subsequent data analysis and processing.

[0491] "Means for analyzing the text data using a generative AI model and evaluating the suspiciousness of the communication" refers to a technology that uses an AI model trained through machine learning to analyze text data and evaluate the risk of fraud based on its content.

[0492] "A means of evaluating a user's emotional state in real time and determining how to respond to a call based on that information" refers to a technology that recognizes emotions from a user's voice data and determines an appropriate response or action based on the results.

[0493] "A means of recording suspicious communication destinations in a database and updating the information list" refers to a technology that identifies potentially fraudulent communication destinations, records them in a database, and updates the list to improve the security of future communications.

[0494] "Means of notifying terminals of fraud prevention information" refers to technology that provides a function to display warnings and alerts on the user's terminal device regarding the possibility of detected fraud.

[0495] "Means of generating careful responses that take into account the impact on the user" refers to technologies that construct responses based on the user's emotional state and fraud risk assessment to provide accurate information while reducing psychological burden.

[0496] In an embodiment of the present invention, the system is designed to analyze voice calls in real time, assess the risk of fraud and the user's emotional state, and generate appropriate responses. The main components of the system include a server, a terminal, a speech recognition engine, a generative AI model, and an emotion engine.

[0497] The server uses a speech recognition engine to convert the user's voice call data into text data. This process utilizes commonly used speech recognition technologies such as the Google Cloud Speech-to-Text API, enabling highly accurate speech-to-text conversion.

[0498] Subsequently, the server analyzes the converted text data using a generative AI model to evaluate the suspiciousness of the communication. It is expected that OpenAI's GPT series will be used as the generative AI model. The generative AI model analyzes the text data based on the prompt message. A specific example of a prompt message is, "Is this call potentially fraudulent? Please assess the risk of fraud."

[0499] Simultaneously, the server uses an emotion engine to evaluate the user's emotional state in real time. This emotion engine utilizes tools such as Affectiva, which provides Emotion AI technology. Through this process, the server determines whether the user is experiencing stress or anxiety, and uses this information to improve call handling.

[0500] The device displays fraud prevention-related information to the user based on notifications from the server. The notifications displayed on the device are warning messages such as "This may be a scam. Please be careful," and are designed to help the user take appropriate action immediately.

[0501] This system protects users from fraudulent activities and allows them to communicate with peace of mind. It also reduces psychological burden and ensures safety for users unfamiliar with digital devices, such as the elderly. By providing support from both a technological and emotional perspective, a high-quality user experience is achieved.

[0502] The flow of the specific processing in Example 2 will be explained using Figure 13.

[0503] Step 1:

[0504] The server receives the user's voice call. It receives voice data as input and analyzes the data using a speech recognition engine. It converts the voice data into text data in real time using APIs such as the Google Cloud Speech-to-Text API. This results in text data as output. Specifically, this process involves the server sending the voice file to the speech recognition API.

[0505] Step 2:

[0506] The server inputs the text data obtained in Step 1 into the generating AI model. The input at this stage is the text data that is the output of Step 1. The generating AI model uses OpenAI's GPT series and performs text analysis based on the prompt sentence. For the purpose of evaluating fraud risk, a prompt sentence such as "Is this call potentially fraudulent? Please evaluate the fraud risk." is used. As a result of the analysis, a fraud risk evaluation score is obtained as output. Specifically, the process involves the server sending the text data to the AI ​​model and performing the analysis.

[0507] Step 3:

[0508] The server passes the audio data to the emotion engine to analyze the user's emotional state. The input at this stage is the audio data handled in step 1. Using emotion recognition technology such as Affectiva, the user's emotions are evaluated in real time. The output is an emotional state report indicating whether the user is stressed or not. The specific operation involves the server inputting audio data into the emotion engine and obtaining the analysis results.

[0509] Step 4:

[0510] The server determines how to respond to the call based on the fraud risk assessment score in Step 2 and the emotional state report in Step 3. In this step, it receives both analysis results as input and generates a careful and appropriate response message as output. For example, messages such as "Please stay calm" or "Please wait while I check the details" may be selected. Specifically, this involves the server generating and outputting a response message based on the analysis results.

[0511] Step 5:

[0512] The device receives a notification from the server and displays fraud prevention information to the user. The input for this step is the response message for step 4. A warning message such as "This may be a scam. Please be careful" is displayed on the device screen. Specifically, this involves a process in which the user visually obtains information via the notification function.

[0513] Step 6:

[0514] The server records suspicious communication destinations in its database and updates the information list. The input for this step is the suspicious communication destination information obtained through the analysis in step 2. The output is an updated blacklist. Specifically, this involves the server adding suspicious phone numbers to the database and strengthening the known list.

[0515] (Application Example 2)

[0516] 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."

[0517] In voice communication, the challenge lies in early detection of fraudulent activity and protecting users from fraud and stress while considering their emotional state. This technology needs to be used to provide a safer and more reliable communication environment. In addition to fraud detection, it is necessary to quickly detect the fear and anxiety users feel and respond appropriately.

[0518] 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.

[0519] In this invention, the server includes means for converting voice information into text information, means for analyzing the text information using a generative AI model to evaluate the suspiciousness of the communication and the user's emotional state, and means for generating a natural response when there is a high probability of fraud and when the user's stress state is detected. This makes it possible to simultaneously achieve rapid detection of fraudulent activity and ensure the psychological safety of the user.

[0520] "Voice information" refers to digital or analog input data of human voices collected through phone calls or recordings.

[0521] "Textual information" refers to data in text format obtained by processing audio data.

[0522] A "generative AI model" is an artificial intelligence algorithm that learns human language and patterns to derive new results.

[0523] "Suspicious communication" refers to unusual behavior or patterns in phone calls or conversations that suggest the possibility of fraud or illegal activity.

[0524] "User emotional state" refers to the emotional tendencies and psychological state that a user experiences during voice interaction.

[0525] "Stress level" refers to the degree of psychological or emotional pressure a user experiences during their daily tasks or interactions.

[0526] A "natural response" is a response generated to maintain smooth and human-like communication in interactions with users.

[0527] "Suspicious contact information" refers to contact information from sources where fraudulent or dishonest activity has been detected in past communications.

[0528] A "database" is a collection of digital information that systematically organizes information, enabling efficient data storage, management, and retrieval.

[0529] A "blacklist" is a list of registered information that restricts or prohibits certain actions or behaviors.

[0530] "Automatic blocking" is a system function that automatically restricts or prohibits communication based on pre-defined criteria or conditions.

[0531] "Fraud prevention information" refers to information that provides notifications and advice related to the detection and prevention of fraudulent activities.

[0532] "Advice" is a term that refers to instructions or suggestions that propose the best course of action in a particular situation.

[0533] To implement this invention, the server uses a program that converts voice information into text information. This program utilizes voice recognition software such as "Google Speech-to-Text API" and has the function of converting voice data into text data in real time. Next, the text data is passed to a generative AI model such as "OpenAI GPT-3" to evaluate the suspiciousness of the communication and the user's emotional state. This generative AI model not only analyzes the content of the call to assess the possibility of fraud, but also uses an emotion analysis algorithm to determine whether the user is feeling stressed.

[0534] The device is equipped with a function that receives notifications from the server and displays fraud prevention information and advice based on the user's emotional state. Notifications alert the user if there is a possibility of fraud or if the user is detected to be experiencing stress. If signs of fraud are detected, the server records suspicious contact information in a database such as "MongoDB" and updates the blacklist. This blacklist is managed through an automatic blocking function to limit further contact.

[0535] As an example of this system, consider a scenario where a user receives a fraudulent phone call impersonating a financial institution. The server analyzes the content of the call in real time, and if it determines that there is a high probability of fraudulent activity, it sends a notification to the device stating, "A potential scam has been detected. Please be careful." Furthermore, if the emotion engine determines that the user is in a high-stress state, it also displays advice such as, "Please stay calm and verify the information." An example of a prompt message to the generative AI model is, "Is the content of this call related to fraud? Please assess the user's stress level, taking into account the data from the emotion analysis engine."

[0536] This allows users to simultaneously achieve rapid detection of fraudulent activity and ensure psychological safety.

[0537] The flow of a specific process in Application Example 2 will be explained using Figure 14.

[0538] Step 1:

[0539] The server receives audio information. The user's phone call audio is provided as input to the server, and the server uses the "Google Speech-to-Text API" to convert this audio information into text information. This process generates text data from the audio data and obtains text-formatted output ready for subsequent analysis processes.

[0540] Step 2:

[0541] The server sends the generated text information to a generative AI model such as "OpenAI GPT-3". The model analyzes the call content and evaluates the suspiciousness of the communication from the input text. In this step, it generates a prompt message to determine whether it is potentially fraudulent, and the AI ​​model processes this information and outputs an evaluation result. For example, the evaluation result may include a judgment such as "Probability of fraud: High".

[0542] Step 3:

[0543] The server simultaneously uses an emotion analysis engine to evaluate the user's emotional state. It passes audio information as input to the emotion analysis engine, which then analyzes the audio data to assess the user's stress level. This step yields an output indicating whether the user is experiencing stress.

[0544] Step 4:

[0545] The server generates natural responses as needed, based on the suspicious nature of the call content and the user's emotional state. Using the evaluation results of the generation AI model and the emotional state as input, it creates response messages such as "A potential scam has been detected. Please be careful" or "Please stay calm and verify the information." These response messages are output using a natural language generation (NLG) algorithm.

[0546] Step 5:

[0547] The device receives a notification from the server and displays a generated response message to the user. Through this notification, the user can receive fraud warnings and stress reduction advice in real time. In this step, the device receives output from the server and displays the message on the screen.

[0548] Step 6:

[0549] The server records suspicious contact information in a database and updates the blacklist. It takes the contact information of the caller of a suspicious call as input and stores it in a database such as "MongoDB," automatically expanding the blacklist. This process outputs a setting that automatically blocks future communications from suspicious contacts.

[0550] 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.

[0551] 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.

[0552] 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.

[0553] [Fourth Embodiment]

[0554] Figure 7 shows an example of the configuration of the data processing system 410 according to the fourth embodiment.

[0555] 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.

[0556] 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).

[0557] 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.

[0558] 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.

[0559] 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).

[0560] 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.

[0561] 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.

[0562] 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.

[0563] 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.

[0564] 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.

[0565] 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.

[0566] 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".

[0567] As an embodiment of this invention, a system having the following functions is constructed: The server monitors the user's voice calls in real time and acquires voice data. The acquired voice data is converted into text data on the server by a speech recognition engine. This text is input into a generative AI model to analyze whether the call content is suspicious and evaluate the possibility of fraud.

[0568] The server receives feedback from the generative AI model and generates a natural response if it determines that the call is likely to be fraudulent. This allows the server to handle suspicious calls on behalf of the user. For example, the server can generate an appropriate response such as "Please wait while I check the invoice" to calm the caller.

[0569] On the other hand, the server records suspicious phone numbers in a database and periodically updates the blacklist. Each time the update is performed, the system is configured to block future incoming calls from the registered suspicious phone numbers. This prevents secondary damage from the same fraudsters.

[0570] Users will receive notifications via their devices when potentially fraudulent calls are detected, as well as periodic notifications about fraud prevention. These notifications allow users to review suspicious call content, helping to raise their security awareness.

[0571] This system can prevent fraudulent activities that target users who are particularly vulnerable to fraud, such as the elderly, and enable safe communication.

[0572] The following describes the processing flow.

[0573] Step 1:

[0574] The server receives user voice calls in real time. A speech recognition engine is used to convert the voice data into text data. This conversion is performed quickly to enable immediate analysis of the call content.

[0575] Step 2:

[0576] The server inputs the converted text data into a generating AI model. The AI ​​model identifies suspicious patterns and phrases from the text and assesses the likelihood of fraud. Based on the assessment results, it quantifies the risk of fraud.

[0577] Step 3:

[0578] If the server detects a high probability of fraud, it prepares a natural-sounding response generated by an AI model. This response follows the flow of human conversation and is intended to delay or mislead the caller of a fraudulent call.

[0579] Step 4:

[0580] The server records phone numbers deemed suspicious in its database. This updates the blacklist, preparing to automatically block future calls from that number.

[0581] Step 5:

[0582] The device will send a notification to the user if it detects a potentially fraudulent call. This notification will include details of the call and a warning about the scam, helping to raise user awareness.

[0583] Step 6:

[0584] Users receive notifications from their devices and take appropriate measures to respond to potentially fraudulent calls. This gives them an opportunity to avoid trouble before it happens.

[0585] (Example 1)

[0586] 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".

[0587] In modern communication technology, users are vulnerable to fraud, and there is a lack of means to quickly identify suspicious calls and identification numbers. Effective measures are needed to reduce fraud, especially for the elderly and tech-savvy users. The challenge lies in achieving a safe communication environment.

[0588] 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.

[0589] In this invention, the server includes means for converting voice information into text information, means for analyzing the text information using a generative information processing model and evaluating the suspiciousness of the communication, and means for recording suspicious identification numbers in an information storage device and updating a restriction list. This makes it possible for users to receive advance warnings about suspicious calls and prevent fraud.

[0590] "Auditory information" refers to data related to sounds such as human speech that are transmitted as sound waves.

[0591] "Textual information" refers to data obtained by converting audio information into text, and is used in natural language processing.

[0592] A "generative information processing model" refers to a trained artificial intelligence technology used for natural language generation and analysis.

[0593] "Means for evaluating suspicious activity" refers to a function that analyzes communication content and patterns to detect suspicious activity that differs from normal communication.

[0594] An "identification number" is a combination of numbers or characters used to identify the caller.

[0595] An "information storage device" is a device or system for recording and storing data temporarily or for a long period of time.

[0596] A "restriction list" is a list used to register suspicious communication partners and restrict or prohibit future contact.

[0597] "User" refers to an individual or organization that uses this system.

[0598] To implement this invention, it is necessary to construct a system having the following elements: The server monitors the user's voice calls in real time using a communication infrastructure. To acquire voice, voice streaming technology is used to collect voice information, and a speech recognition engine is used to convert it into text information. Specifically, a speech recognition API (e.g., a cloud-based service) can be used for speech recognition.

[0599] The server inputs the converted character information into a generative information processing model and uses the model's analytical capabilities to evaluate the suspiciousness of the communication. This generative information processing model uses a natural language processing model based on existing technologies (e.g., a large-scale language model). Based on the analysis results, the server generates a natural response and, if necessary, deals with suspicious calls.

[0600] Furthermore, the server has a mechanism to prevent similar suspicious communication activities by storing suspicious identification numbers in an information storage device and updating the restriction list. The terminal has a notification function related to fraud prevention, providing users with security-related information in real time.

[0601] For example, if a user prompts, "Have I had any suspicious calls recently?", the server scans recent call logs and notifies the user of details about any calls suspected of being fraudulent. In this way, users can more securely manage their own calling environment. This system helps combat fraudulent activities and provides a safe and secure communication experience.

[0602] The flow of the specific processing in Example 1 will be explained using Figure 11.

[0603] Step 1:

[0604] The server acquires voice data from a user who initiates a voice call in real time. An audio stream is provided as input. The server transfers this audio data to a speech recognition engine, where it is converted into text data. The output is a string of characters reflecting the content of the call. This process involves accumulating the audio data in a buffer and inputting it to the speech recognition engine in batch processing.

[0605] Step 2:

[0606] The server inputs the obtained text data into the generative AI model. The string generated in the previous step is used as input. The generative AI model analyzes this string and evaluates whether the communication content is suspicious. The results of this analysis are scored, and the evaluation result is obtained as output. Specifically, this involves the model analyzing text patterns and quantifying the likelihood of fraud.

[0607] Step 3:

[0608] The server generates a natural response if the evaluation result exceeds a threshold. In this process, it receives an evaluation score from a generative AI model as input and constructs a response using a response generation engine. The output is the response that should be sent to the caller. The operation includes creating the optimal response by referring to past conversation data and preset templates.

[0609] Step 4:

[0610] The server records suspicious identification numbers in a database. The input is the identification numbers detected during a call, which are stored in the information storage device. The output is an updated restricted list. The operation includes a mechanism where new identification numbers are added to the existing list, and this information is periodically transmitted to other user terminals.

[0611] Step 5:

[0612] The device notifies the user of fraud prevention information. Here, it receives a warning message from the server as input and displays it in the user interface. The output is a warning message that the user can review. Specifically, this includes actions such as a notification popping up on the screen and providing detailed information as needed.

[0613] (Application Example 1)

[0614] 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".

[0615] In recent years, fraudulent activities have become a serious social problem, particularly telephone scams targeting the elderly, and effective preventative measures are needed. The problem is that many users continue calls without fully understanding the risks of fraud. Furthermore, a system is needed that can immediately detect communications with a high probability of being fraudulent and warn users.

[0616] 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.

[0617] In this invention, the server includes means for converting acoustic information into text information, means for analyzing the text information using a generation AI model and evaluating the suspiciousness of the communication, and means for displaying a warning for suspicious communication and providing an option to terminate the call. This makes it possible to immediately detect communications that are highly likely to be fraudulent and to help users make safe decisions.

[0618] "Means for converting acoustic information into text information" refers to a device or program that performs the process of converting audio data into text data in real time.

[0619] "Means for analyzing the aforementioned textual information using a generative AI model and evaluating the suspiciousness of the communication" refers to a device or program that uses artificial intelligence to analyze textual information and determine whether the content of the communication is fraudulent.

[0620] "Means for generating an automated response when a communication is deemed highly likely to be fraudulent" refers to a device or program that automatically generates an appropriate response when it detects a communication that is strongly suspected of being fraudulent.

[0621] "Means for recording suspicious identification information in a data storage device and updating the restriction list" refers to a device or program that stores suspicious phone numbers or identification information in a database and periodically updates the restriction list.

[0622] "Means of notifying users of fraud prevention information" refers to a device or program that notifies the user's terminal of information regarding the risk of fraud.

[0623] "Means for displaying a warning about suspicious communications and providing an option to terminate a call" refers to a device or program that, upon detecting suspicious communications, presents a warning to the user and provides an interface for choosing whether to continue or terminate the call.

[0624] The system for implementing this invention primarily involves the processes of acquiring, analyzing, responding to, and recording voice data. First, the terminal immediately transmits the received call audio as acoustic information to the server. The server converts this acoustic information into text information using the Google Cloud Speech-to-Text API. The converted text information is then analyzed by a generative AI model, in this case OpenAI GPT, to assess the likelihood of fraud.

[0625] If the server determines that there is a high probability of fraud, it will automatically generate an appropriate response message. This response message can instruct the other party, such as "We are checking, please wait a moment," in a way that does not make the call seem unnatural.

[0626] The server also records suspicious identification information, such as phone numbers, in its database. This data is added to a restricted list, and settings are made to block future communications. Furthermore, users are displayed a warning about the risk of fraud through their device and given the option to continue or end the call. This feature allows users to ensure their own safety through their own judgment.

[0627] For example, if an elderly user receives a phone call from a stranger requesting personal information, this system will respond quickly. After the audio is converted into text on the server, the user will be warned that "this call may be a scam," allowing for a swift response. An example of a prompt to the generating AI model is, "Does the following text contain signs of fraud? Please tell me if it is a scam and why."

[0628] The flow of a specific process in Application Example 1 will be explained using Figure 12.

[0629] Step 1:

[0630] The terminal instantly transmits the call audio received by the user to the server. The input is the audio data obtained from the call, and the output is the transmission of this audio data to the server. The terminal's operation involves digitizing the received audio signal and transferring it to the server using a secure communication protocol.

[0631] Step 2:

[0632] The server uses the Google Cloud Speech-to-Text API to convert the received audio data into text. The input is the audio data received in step 1, and the output is the converted text. In this step, a speech recognition algorithm is applied, the audio waveform is analyzed within the server, and the corresponding text is converted.

[0633] Step 3:

[0634] The server uses the OpenAI GPT generative AI model to analyze textual information and evaluate the suspiciousness of the communication. The input is the textual information obtained in step 2, and the output is the evaluation result regarding the likelihood of fraud. In this step, the prompt "Does the following text contain signs of fraud? Please tell us the likelihood of fraud and why." is used to input textual information into the generative AI model and quantify the risk of fraud.

[0635] Step 4:

[0636] If the server determines that there is a high probability of fraud, it generates an automated response message. The input is the evaluation result obtained in step 3, and the output is the generated response message. The server uses a generation AI model to generate a calming response such as "We are checking, please wait a moment."

[0637] Step 5:

[0638] The server records suspicious identification information in a data storage device and updates the restriction list. The input is the identification information of a call suspected of being fraudulent, and the output is the updated restriction list. The server writes this identification information to a database and automatically configures settings to block future calls.

[0639] Step 6:

[0640] The user will see a warning message about the risk of fraud through their device. The input is the evaluation result from step 3, and the output is the warning message that the user can see on the screen. The device presents the user with the risks of continuing the call and the option to end it, allowing the user to end the call at their own discretion if necessary.

[0641] 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.

[0642] As an embodiment of this invention, a system incorporating a server, a terminal, and an emotion engine is provided. This system monitors the user's voice calls and processes the data acquired by speech recognition in real time. The server converts the voice data into text data and performs analysis using a generative AI model and an emotion engine.

[0643] The generative AI model analyzes the call content and assesses the likelihood of fraud. The emotion engine recognizes emotions from the user's voice and evaluates their state in real time. Based on this information, the server determines the appropriate course of action, considering not only the risk of fraud but also the psychological impact the call has on the user.

[0644] Specifically, the server generates more cautious responses when it determines there is a high probability of fraud, as well as when it assesses that the user is in a stressful state. For example, it can use phrases such as "Please calm down" or "Please wait while I check the details" to reduce the user's stress while the call progresses.

[0645] Furthermore, the server records suspicious phone numbers in a database. User safety is ensured by continuously updating the blacklist and automatically blocking incoming calls from known fraudulent phone numbers.

[0646] The device will notify users of potentially fraudulent calls and information about the user's emotional state. This allows users to take appropriate action. For example, a notification such as "A potential scam has been detected. Please be careful" will appear on the device.

[0647] This system allows users, including the elderly, to avoid the risk of fraud while reducing psychological burden. By combining it with emotional recognition, it goes beyond mere technical responses and enables activities that are empathetic to the user's feelings.

[0648] The following describes the processing flow.

[0649] Step 1:

[0650] The server receives user voice calls in real time. A speech recognition engine is used to convert the voice data into text data. This process extracts the call content, enabling immediate analysis.

[0651] Step 2:

[0652] The server inputs the converted text data into a generating AI model. This model analyzes the call content and assesses the likelihood of fraud. Specifically, it searches for matches with past fraud patterns and calculates the fraud risk.

[0653] Step 3:

[0654] The server inputs voice data from the user into an emotion engine to analyze the user's emotional state. This engine evaluates the user's stress and anxiety levels based on the tone and pace of their voice.

[0655] Step 4:

[0656] The server integrates the evaluation of the generative AI model and the analysis results of the emotion engine. If the risk of fraud is high and the user's emotional state is unstable, it generates a cautious response. This response includes words that soothe the user's emotions.

[0657] Step 5:

[0658] The server records suspicious phone numbers in its database. It updates the blacklist and configures the system to automatically block incoming calls from those numbers.

[0659] Step 6:

[0660] The device sends notifications to the user based on the possibility of fraud and their emotional state. These notifications may include messages such as, "Fraud has been detected. Please stay calm."

[0661] Step 7:

[0662] Users can check notifications from their devices and take action based on the analysis of call details and emotional state. This helps prevent fraud and promotes emotional stability.

[0663] (Example 2)

[0664] 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".

[0665] In recent years, there has been a growing need for effective systems to protect users from telephone-based fraud. In particular, for users, including the elderly, it is necessary to quickly assess the risk of fraud and provide responses that reduce psychological burden. However, existing technologies make it difficult to provide a comprehensive response that simultaneously considers fraud risk assessment and psychological impact. To solve this problem, a system is needed that evaluates the suspiciousness of communications through real-time voice analysis and generates appropriate responses that take into account the user's emotional state.

[0666] 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.

[0667] In this invention, the server includes means for converting voice data into text data, means for analyzing the text data using a generative AI model and evaluating the suspiciousness of the communication, and means for evaluating the user's emotional state in real time and determining how to respond to the call based on that information. This makes it possible to quickly evaluate communications that are likely to be fraudulent and provide responses that take the user's emotions into consideration, thereby preventing communication troubles and ensuring the user's safety and psychological security.

[0668] "Methods for converting audio data into text data" refers to technologies that use speech recognition technology to convert audio into strings of characters, creating a text format that can be used for subsequent data analysis and processing.

[0669] "Means for analyzing the text data using a generative AI model and evaluating the suspiciousness of the communication" refers to a technology that uses an AI model trained through machine learning to analyze text data and evaluate the risk of fraud based on its content.

[0670] "A means of evaluating a user's emotional state in real time and determining how to respond to a call based on that information" refers to a technology that recognizes emotions from a user's voice data and determines an appropriate response or action based on the results.

[0671] "A means of recording suspicious communication destinations in a database and updating the information list" refers to a technology that identifies potentially fraudulent communication destinations, records them in a database, and updates the list to improve the security of future communications.

[0672] "Means of notifying terminals of fraud prevention information" refers to technology that provides a function to display warnings and alerts on the user's terminal device regarding the possibility of detected fraud.

[0673] "Means of generating careful responses that take into account the impact on the user" refers to technologies that construct responses based on the user's emotional state and fraud risk assessment to provide accurate information while reducing psychological burden.

[0674] In an embodiment of the present invention, the system is designed to analyze voice calls in real time, assess the risk of fraud and the user's emotional state, and generate appropriate responses. The main components of the system include a server, a terminal, a speech recognition engine, a generative AI model, and an emotion engine.

[0675] The server uses a speech recognition engine to convert the user's voice call data into text data. This process utilizes commonly used speech recognition technologies such as the Google Cloud Speech-to-Text API, enabling highly accurate speech-to-text conversion.

[0676] Subsequently, the server analyzes the converted text data using a generative AI model to evaluate the suspiciousness of the communication. It is expected that OpenAI's GPT series will be used as the generative AI model. The generative AI model analyzes the text data based on the prompt message. A specific example of a prompt message is, "Is this call potentially fraudulent? Please assess the risk of fraud."

[0677] Simultaneously, the server uses an emotion engine to evaluate the user's emotional state in real time. This emotion engine utilizes tools such as Affectiva, which provides Emotion AI technology. Through this process, the server determines whether the user is experiencing stress or anxiety, and uses this information to improve call handling.

[0678] The device displays fraud prevention-related information to the user based on notifications from the server. The notifications displayed on the device are warning messages such as "This may be a scam. Please be careful," and are designed to help the user take appropriate action immediately.

[0679] This system protects users from fraudulent activities and allows them to communicate with peace of mind. It also reduces psychological burden and ensures safety for users unfamiliar with digital devices, such as the elderly. By providing support from both a technological and emotional perspective, a high-quality user experience is achieved.

[0680] The flow of the specific processing in Example 2 will be explained using Figure 13.

[0681] Step 1:

[0682] The server receives the user's voice call. It receives voice data as input and analyzes the data using a speech recognition engine. It converts the voice data into text data in real time using APIs such as the Google Cloud Speech-to-Text API. This results in text data as output. Specifically, this process involves the server sending the voice file to the speech recognition API.

[0683] Step 2:

[0684] The server inputs the text data obtained in Step 1 into the generating AI model. The input at this stage is the text data that is the output of Step 1. The generating AI model uses OpenAI's GPT series and performs text analysis based on the prompt sentence. For the purpose of evaluating fraud risk, a prompt sentence such as "Is this call potentially fraudulent? Please evaluate the fraud risk." is used. As a result of the analysis, a fraud risk evaluation score is obtained as output. Specifically, the process involves the server sending the text data to the AI ​​model and performing the analysis.

[0685] Step 3:

[0686] The server passes the audio data to the emotion engine to analyze the user's emotional state. The input at this stage is the audio data handled in step 1. Using emotion recognition technology such as Affectiva, the user's emotions are evaluated in real time. The output is an emotional state report indicating whether the user is stressed or not. The specific operation involves the server inputting audio data into the emotion engine and obtaining the analysis results.

[0687] Step 4:

[0688] The server determines how to respond to the call based on the fraud risk assessment score in Step 2 and the emotional state report in Step 3. In this step, it receives both analysis results as input and generates a careful and appropriate response message as output. For example, messages such as "Please stay calm" or "Please wait while I check the details" may be selected. Specifically, this involves the server generating and outputting a response message based on the analysis results.

[0689] Step 5:

[0690] The device receives a notification from the server and displays fraud prevention information to the user. The input for this step is the response message for step 4. A warning message such as "This may be a scam. Please be careful" is displayed on the device screen. Specifically, this involves a process in which the user visually obtains information via the notification function.

[0691] Step 6:

[0692] The server records suspicious communication destinations in its database and updates the information list. The input for this step is the suspicious communication destination information obtained through the analysis in step 2. The output is an updated blacklist. Specifically, this involves the server adding suspicious phone numbers to the database and strengthening the known list.

[0693] (Application Example 2)

[0694] 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".

[0695] In voice communication, the challenge lies in early detection of fraudulent activity and protecting users from fraud and stress while considering their emotional state. This technology needs to be used to provide a safer and more reliable communication environment. In addition to fraud detection, it is necessary to quickly detect the fear and anxiety users feel and respond appropriately.

[0696] 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.

[0697] In this invention, the server includes means for converting voice information into text information, means for analyzing the text information using a generative AI model to evaluate the suspiciousness of the communication and the user's emotional state, and means for generating a natural response when there is a high probability of fraud and when the user's stress state is detected. This makes it possible to simultaneously achieve rapid detection of fraudulent activity and ensure the psychological safety of the user.

[0698] "Voice information" refers to digital or analog input data of human voices collected through phone calls or recordings.

[0699] "Textual information" refers to data in text format obtained by processing audio data.

[0700] A "generative AI model" is an artificial intelligence algorithm that learns human language and patterns to derive new results.

[0701] "Suspicious communication" refers to unusual behavior or patterns in phone calls or conversations that suggest the possibility of fraud or illegal activity.

[0702] "User emotional state" refers to the emotional tendencies and psychological state that a user experiences during voice interaction.

[0703] "Stress level" refers to the degree of psychological or emotional pressure a user experiences during their daily tasks or interactions.

[0704] A "natural response" is a response generated to maintain smooth and human-like communication in interactions with users.

[0705] "Suspicious contact information" refers to contact information from sources where fraudulent or dishonest activity has been detected in past communications.

[0706] A "database" is a collection of digital information that systematically organizes information, enabling efficient data storage, management, and retrieval.

[0707] A "blacklist" is a list of registered information that restricts or prohibits certain actions or behaviors.

[0708] "Automatic blocking" is a system function that automatically restricts or prohibits communication based on pre-defined criteria or conditions.

[0709] "Fraud prevention information" refers to information that provides notifications and advice related to the detection and prevention of fraudulent activities.

[0710] "Advice" is a term that refers to instructions or suggestions that propose the best course of action in a particular situation.

[0711] To implement this invention, the server uses a program that converts voice information into text information. This program utilizes voice recognition software such as "Google Speech-to-Text API" and has the function of converting voice data into text data in real time. Next, the text data is passed to a generative AI model such as "OpenAI GPT-3" to evaluate the suspiciousness of the communication and the user's emotional state. This generative AI model not only analyzes the content of the call to assess the possibility of fraud, but also uses an emotion analysis algorithm to determine whether the user is feeling stressed.

[0712] The device is equipped with a function that receives notifications from the server and displays fraud prevention information and advice based on the user's emotional state. Notifications alert the user if there is a possibility of fraud or if the user is detected to be experiencing stress. If signs of fraud are detected, the server records suspicious contact information in a database such as "MongoDB" and updates the blacklist. This blacklist is managed through an automatic blocking function to limit further contact.

[0713] As an example of this system, consider a scenario where a user receives a fraudulent phone call impersonating a financial institution. The server analyzes the content of the call in real time, and if it determines that there is a high probability of fraudulent activity, it sends a notification to the device stating, "A potential scam has been detected. Please be careful." Furthermore, if the emotion engine determines that the user is in a high-stress state, it also displays advice such as, "Please stay calm and verify the information." An example of a prompt message to the generative AI model is, "Is the content of this call related to fraud? Please assess the user's stress level, taking into account the data from the emotion analysis engine."

[0714] This allows users to simultaneously achieve rapid detection of fraudulent activity and ensure psychological safety.

[0715] The flow of a specific process in Application Example 2 will be explained using Figure 14.

[0716] Step 1:

[0717] The server receives audio information. The user's phone call audio is provided as input to the server, and the server uses the "Google Speech-to-Text API" to convert this audio information into text information. This process generates text data from the audio data and obtains text-formatted output ready for subsequent analysis processes.

[0718] Step 2:

[0719] The server sends the generated text information to a generative AI model such as "OpenAI GPT-3". The model analyzes the call content and evaluates the suspiciousness of the communication from the input text. In this step, it generates a prompt message to determine whether it is potentially fraudulent, and the AI ​​model processes this information and outputs an evaluation result. For example, the evaluation result may include a judgment such as "Probability of fraud: High".

[0720] Step 3:

[0721] The server simultaneously uses an emotion analysis engine to evaluate the user's emotional state. It passes audio information as input to the emotion analysis engine, which then analyzes the audio data to assess the user's stress level. This step yields an output indicating whether the user is experiencing stress.

[0722] Step 4:

[0723] The server generates natural responses as needed, based on the suspicious nature of the call content and the user's emotional state. Using the evaluation results of the generation AI model and the emotional state as input, it creates response messages such as "A potential scam has been detected. Please be careful" or "Please stay calm and verify the information." These response messages are output using a natural language generation (NLG) algorithm.

[0724] Step 5:

[0725] The device receives a notification from the server and displays a generated response message to the user. Through this notification, the user can receive fraud warnings and stress reduction advice in real time. In this step, the device receives output from the server and displays the message on the screen.

[0726] Step 6:

[0727] The server records suspicious contact information in a database and updates the blacklist. It takes the contact information of the caller of a suspicious call as input and stores it in a database such as "MongoDB," automatically expanding the blacklist. This process outputs a setting that automatically blocks future communications from suspicious contacts.

[0728] 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.

[0729] 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.

[0730] 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.

[0731] 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.

[0732] 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.

[0733] 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.

[0734] 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.

[0735] 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.

[0736] 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."

[0737] 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.

[0738] 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.

[0739] 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.

[0740] 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.

[0741] 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.

[0742] 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.

[0743] 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.

[0744] 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.

[0745] 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.

[0746] 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.

[0747] 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.

[0748] 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.

[0749] The following is further disclosed regarding the embodiments described above.

[0750] (Claim 1)

[0751] A means of converting audio data into text data,

[0752] A means for analyzing the text data using a generative AI model and evaluating the suspiciousness of the communication,

[0753] A means of generating a natural response when it is determined that there is a high probability of fraud,

[0754] A method to record suspicious phone numbers in a database and update the blacklist,

[0755] A system that includes means for notifying users of fraud prevention information.

[0756] (Claim 2)

[0757] The system according to claim 1, wherein the speech-to-text conversion means is configured to operate in real time.

[0758] (Claim 3)

[0759] The system according to claim 1, wherein the generative AI model is configured to continuously learn communication patterns and improve its accuracy.

[0760] "Example 1"

[0761] (Claim 1)

[0762] A means of converting audio information into text information,

[0763] A means for analyzing the character information using a generative information processing model and evaluating the suspiciousness of the communication,

[0764] A means of generating a natural response when it is determined that there is a high probability of fraud,

[0765] A means of recording suspicious identification numbers in an information storage device and updating the restricted list,

[0766] A means of notifying users of fraud prevention information,

[0767] A system that includes means for automatically restricting communication partners based on matching information.

[0768] (Claim 2)

[0769] The system according to claim 1, wherein the speech-to-text conversion means is configured to operate immediately.

[0770] (Claim 3)

[0771] The system according to claim 1, wherein the generative information processing model is configured to continuously learn communication patterns and improve their accuracy.

[0772] "Application Example 1"

[0773] (Claim 1)

[0774] A means of converting acoustic information into textual information,

[0775] A means for analyzing the aforementioned text information using a generative AI model and evaluating the suspiciousness of the communication,

[0776] A means of generating an automated response when it is determined that there is a high probability of fraud,

[0777] A means of recording suspicious identification information in a data storage device and updating the restriction list,

[0778] A means of informing users of fraud prevention information,

[0779] A system that includes means for displaying a warning for suspicious communications and providing an option to terminate the call.

[0780] (Claim 2)

[0781] The system according to claim 1, wherein the acoustic text conversion means is configured to operate in real time.

[0782] (Claim 3)

[0783] The system according to claim 1, wherein the generative AI model is configured to continuously learn communication patterns and improve its performance.

[0784] "Example 2 of combining an emotion engine"

[0785] (Claim 1)

[0786] A means of converting audio data into text data,

[0787] A means for analyzing the text data using a generative AI model and evaluating the suspiciousness of the communication,

[0788] A means for evaluating the user's emotional state in real time and determining how to respond to a call based on that information,

[0789] A means of recording suspicious communication destinations in a database and updating the information list,

[0790] A means of notifying the device of fraud prevention information,

[0791] A means of generating a careful response that takes into account the impact on the user,

[0792] A system that includes this.

[0793] (Claim 2)

[0794] The system according to claim 1, wherein the speech-to-text conversion means is configured to operate in real time.

[0795] (Claim 3)

[0796] The system according to claim 1, wherein the generative AI model is configured to continuously learn communication patterns and improve its accuracy.

[0797] "Application example 2 when combining with an emotional engine"

[0798] (Claim 1)

[0799] A means of converting audio information into text information,

[0800] A means for analyzing the aforementioned textual information using a generative AI model and evaluating the suspiciousness of the communication and the emotional state of the user,

[0801] A means for generating a natural response when there is a high probability of fraud and when a user's stressed state is detected,

[0802] A means of recording suspicious contact information in a database, updating the blacklist, and automatically blocking users,

[0803] A system that includes means of notifying users of fraud prevention information and advice based on their emotional state.

[0804] (Claim 2)

[0805] The system according to claim 1, wherein the speech-to-text conversion means is configured to operate in real time.

[0806] (Claim 3)

[0807] The system according to claim 1, wherein the generative AI model is configured to continuously learn both communication patterns and voice patterns to improve its accuracy. [Explanation of symbols]

[0808] 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 means of converting audio data into text data, A means for analyzing the text data using a generative AI model and evaluating the suspiciousness of the communication, A means of generating a natural response when it is determined that there is a high probability of fraud, A method to record suspicious phone numbers in a database and update the blacklist, A system that includes means for notifying users of fraud prevention information.

2. The system according to claim 1, wherein the speech-to-text conversion means is configured to operate in real time.

3. The system according to claim 1, wherein the generative AI model is configured to continuously learn communication patterns and improve its accuracy.

Citation Information

Patent Citations

  • Persona chatbot control method and system

    JP2022180282A