System

Abstract

A system is provided.SOLUTION: The system includes a voice collection means for collecting voice data at all times, a generation model means for analyzing the collected voice data and evaluating the possibility of cash card fraud in real time, an alert generation means for transmitting an alert when it is determined that the possibility of cash card fraud is high by the generation model means, and a notification means for notifying a family member or the police of the transmitted alert.SELECTED DRAWING: Figure 1

Description

[Technical Field]

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

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

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

[0004] Cash card fraud, a crime that particularly targets seniors aged 65 and over, has become a serious social problem. Current countermeasures rely primarily on advance warnings and education, making it difficult to respond in real time. This makes it difficult to prevent fraud before it occurs, and many seniors fall victim to it. The present invention aims to address this cash card fraud problem by providing a system that can detect fraud in real time and quickly take appropriate countermeasures, thereby ensuring the safety of seniors and preventing fraud. [Means for solving the problem]

[0005] The present invention is a system that includes a voice collection means that constantly collects voice data, a generative model means that analyzes the collected voice data and evaluates the possibility of cash card fraud in real time, an alert generation means that issues an alert when the generative model means determines that there is a high possibility of cash card fraud, and a notification means that notifies family members and the police of the issued alert. Furthermore, the alert sent by the notification means includes location information of the point of origin, and the generative model means works in conjunction with a database of past cash card fraud cases, thereby enabling more accurate detection of fraud. This allows for the prevention of fraud and, in emergencies, allows relevant parties to be quickly notified and appropriate measures to be taken.

[0006] The "audio collection means" is a device or software that has the function of constantly collecting surrounding audio data and supplying it for subsequent processing.

[0007] "Generative model means" refers to the generative AI model used to analyze collected voice data and, in particular, assess the likelihood of bank card fraud in real time.

[0008] The "alert generation means" is a device or software that has the function of generating a warning or notification signal when the generative model means determines that there is a high possibility of cash card fraud.

[0009] The "notification means" is a device or software that has the function of transmitting the warning or notification signal generated by the alert generation means to registered related parties such as family members or the police.

[0010] "Location Information" refers to data included in an alert sent by a notification means for identifying the geographic location of the point of origin.

[0011] The "database of past cash card fraud cases" is a database that accumulates data on cash card fraud cases that have occurred in the past, and is referenced by the generative model means when evaluating fraud. [Brief explanation of the drawings]

[0012] [Figure 1] 1 is a conceptual diagram showing an example of the configuration of a data processing system according to a first embodiment. [Figure 2] 1 is a conceptual diagram showing an example of main functions of a data processing device and a smart device according to a first embodiment. [Figure 3] FIG. 10 is a conceptual diagram showing an example of the configuration of a data processing system according to a second embodiment. [Figure 4] FIG. 10 is a conceptual diagram showing an example of main functions of a data processing device and smart glasses according to a second embodiment. [Figure 5] FIG. 10 is a conceptual diagram showing an example of the configuration of a data processing system according to a third embodiment. [Figure 6] FIG. 11 is a conceptual diagram showing an example of main functions of a data processing device and a headset-type terminal according to a third embodiment. [Figure 7] FIG. 10 is a conceptual diagram showing an example of the configuration of a data processing system according to a fourth embodiment. [Figure 8] FIG. 10 is a conceptual diagram showing an example of main functions of a data processing device and a robot according to a fourth embodiment. [Figure 9] 1 shows an emotion map onto which multiple emotions are mapped. [Figure 10] 1 shows an emotion map onto which multiple emotions are mapped. [Figure 11] FIG. 3 is a sequence diagram showing a processing flow of the data processing system according to the first embodiment. [Figure 12] FIG. 10 is a sequence diagram showing the flow of processing in the data processing system in Application Example 1. [Figure 13] FIG. 10 is a sequence diagram showing the flow of processing in the data processing system according to the second embodiment when an emotion engine is combined. [Figure 14] FIG. 10 is a sequence diagram showing the flow of processing in the data processing system in Application Example 2 when an emotion engine is combined. DETAILED DESCRIPTION OF THE INVENTION

[0013] An example of an embodiment of a system according to the technology of the present disclosure will be described below with reference to the accompanying drawings.

[0014] First, the terms used in the following description will be explained.

[0015] In the following embodiments, a coded processor (hereinafter simply referred to as a "processor") may be a single arithmetic device or a combination of multiple arithmetic devices. Furthermore, a processor may be a single type of arithmetic device or a combination of multiple types of arithmetic devices. Examples of arithmetic devices include a CPU (Central Processing Unit), a GPU (Graphics Processing Unit), a GPGPU (General-Purpose computing on Graphics Processing Units), and an APU (Accelerated Processing Unit).

[0016] In the following embodiments, a coded RAM (Random Access Memory) is a memory in which information is temporarily stored and is used as a working memory by a processor.

[0017] In the following embodiments, the coded storage is one or more non-volatile storage devices that store various programs, various parameters, etc. Examples of non-volatile storage devices include flash memory (SSD (Solid State Drive)), magnetic disks (e.g., hard disks), and magnetic tapes.

[0018] In the following embodiments, a communication I / F (Interface) with a symbol is an interface including a communication processor, an antenna, etc. The communication I / F controls communication between multiple computers. Examples of communication standards applied to the communication I / F include wireless communication standards including 5G (5th Generation Mobile Communication System), Wi-Fi (registered trademark), Bluetooth (registered trademark), etc.

[0019] In the following embodiments, "A and / or B" is synonymous with "at least one of A and B." In other words, "A and / or B" means that it may be only A, only B, or a combination of A and B. Furthermore, in this specification, the same concept as "A and / or B" is also applied when three or more things are expressed connected by "and / or."

[0020] [First embodiment]

[0021] FIG. 1 shows an example of the configuration of a data processing system 10 according to the first embodiment.

[0022] 1, a data processing system 10 includes a data processing device 12 and a smart device 14. An example of the data processing device 12 is a server.

[0023] The data processing device 12 includes a computer 22, a database 24, and a communication I / F 26. The computer 22 is an example of a "computer" according to the technology of the present disclosure. The computer 22 includes a processor 28, a RAM 30, and a storage 32. The processor 28, the RAM 30, and the storage 32 are connected to a bus 34. The database 24 and the communication I / F 26 are also connected to the bus 34. The communication I / F 26 is connected to a network 54. Examples of the network 54 include a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0024] The smart device 14 includes a computer 36, a reception device 38, an output device 40, a camera 42, and a communication I / F 44. The computer 36 includes a processor 46, a RAM 48, and a storage 50. The processor 46, the RAM 48, and the storage 50 are connected to a bus 52. The reception device 38, the output device 40, and the camera 42 are also connected to the bus 52.

[0025] The reception device 38 includes a touch panel 38A, a microphone 38B, and the like, and receives user input. The touch panel 38A detects contact with an indicator (for example, a pen or a finger) to receive user input by the touch of the indicator. The microphone 38B detects the user's voice to receive user input by voice. The control unit 46A transmits data indicating the user input received by the touch panel 38A and the microphone 38B to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the data indicating the user input.

[0026] The output device 40 includes a display 40A and a speaker 40B, and presents data to the user 20 by outputting the data in a form of expression that the user 20 can perceive (for example, audio and / or text). The display 40A displays visible information such as text and images in accordance with instructions from the processor 46. The speaker 40B outputs audio in accordance with instructions from the processor 46. The camera 42 is a compact digital camera equipped with an optical system including a lens, aperture, and shutter, and an imaging element such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor.

[0027] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 control the exchange of various information between the processor 46 and the processor 28 via the network 54.

[0028] FIG. 2 shows an example of the main functions of the data processing device 12 and the smart device 14.

[0029] 2, in the data processing device 12, a specific process is performed by the processor 28. A specific processing program 56 is stored in the storage 32. The specific processing program 56 is an example of a "program" according to the technology of the present disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific process is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific processing program 56 executed on the RAM 30.

[0030] The storage 32 stores a data generation model 58 and an emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[0031] In the smart device 14, the processor 46 performs the reception output process. The storage 50 stores a reception output program 60. The reception output program 60 is used in conjunction with the specific processing program 56 by the data processing system 10. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output process is realized by the processor 46 operating as the control unit 46A in accordance with the reception output program 60 executed on the RAM 48.

[0032] Next, a description will be given of the specific processing performed by the specific processing unit 290 of the data processing device 12. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0033] The system of the present invention is an advanced crime prevention system for detecting and preventing cash card fraud in real time. The detailed operation of the system and specific methods of use are described below.

[0034] System configuration

[0035] 1. Audio collection method

[0036] The device has a built-in microphone that constantly collects sounds around the user, and the collected sound data is processed digitally in real time.

[0037] 2. Generative Modeling Methods

[0038] To analyze the collected voice data, the device incorporates a generative AI model that uses pre-trained machine learning algorithms to detect specific phrases and patterns in speech.

[0039] 3. Alert Generation Methods

[0040] If the generative AI model determines that there is a high probability of bank card fraud, the device will immediately generate an alert, which will include specific details of the suspected fraudulent conversation and its location.

[0041] 4. Means of notification

[0042] The alert generated on the device is sent to a server, which then notifies registered family members and the police via a dedicated app, SMS, or email.

[0043] Specific examples

[0044] 1. If you receive a fraudulent phone call

[0045] The elderly user receives a call from an unknown number. The caller claims to be a bank employee and asks for confirmation of their cash card number and PIN.

[0046] The device collects this conversation using a voice capture means and analyzes it using a generative model means. The generative AI model detects dangerous phrases such as "cash card" and "PIN number."

[0047] The generative model means determines that there is a high suspicion of fraud and activates the alert generation means, which immediately generates an alert containing detailed information about the fraud and location information.

[0048] The generated alert is sent to a server by a notification means, and the server sends notifications to registered family members and the police in real time.

[0049] Family members and police receive alerts through a dedicated app and can immediately contact the elderly person or go directly to the scene.

[0050] 2. Normal everyday conversation

[0051] When a user has a daily conversation with a family member over the phone, the terminal similarly collects voice data and analyzes it using the generative model means.

[0052] If the generative AI model does not detect any particularly dangerous phrases, the audio data is securely stored in the system as a log and no alert is generated.

[0053] The server periodically collects log data and uses the analysis results to strengthen security.

[0054] Program processing

[0055] To make it easier to understand, the specific flow of the system's program processing is shown below:

[0056] 1. Initialize the device

[0057] When the device boots up, it loads the generative model means, activates the audio collection means, establishes an internet connection, and synchronizes with the server.

[0058] 2. Audio collection and analysis

[0059] The device constantly collects audio from the user's surroundings and feeds it into a generative AI model in real time, which then analyzes the audio data to detect phrases that could be fraudulent.

[0060] 3. Alert generation and notification

[0061] If the generative model means highly evaluates the possibility of cash card fraud, the alert generation means immediately generates an alert and transmits it to the server.

[0062] The server receives the alert and sends notifications to registered family members and police.

[0063] This will prevent fraud and ensure the safety of the elderly.

[0064] The processing flow will be explained below.

[0065] Program processing flow

[0066] Step 1: Initialization

[0067] Device:

[0068] When the terminal is powered on, it loads the generative model means and enables the audio collection means.

[0069] The device establishes an Internet connection and begins communicating with the server.

[0070] server:

[0071] The server accepts the connection request from the terminal and checks the authentication information.

[0072] User:

[0073] Users install a dedicated app and enter the necessary information (personal information, emergency contact information).

[0074] Step 2: Continuously collect audio data

[0075] Device:

[0076] The device constantly collects surrounding sounds through a microphone, and this collected audio data is processed digitally in real time.

[0077] Step 3: Analyzing the audio data

[0078] Device:

[0079] The device feeds the collected voice data into a generative AI model for real-time analysis.

[0080] The generative modeling means detects specific phrases and patterns in the audio data that are indicative of fraud.

[0081] Step 4: Generate an alert if fraud is suspected

[0082] Device:

[0083] If the generative AI model determines that fraud is highly suspected, the device will generate an alert.

[0084] The alert will include detailed information about the scam and its location.

[0085] Step 5: Sending an alert

[0086] Device:

[0087] The generated alert is sent to the server via a notification means.

[0088] server:

[0089] The server receives the alerts and records them in a database.

[0090] The server will then send notifications to registered family members and the police.

[0091] Step 6: Receive and respond to notifications

[0092] User (family / police):

[0093] Family members and police will receive alerts via a dedicated app, SMS or email.

[0094] Family members and police will check the alert and take necessary measures.

[0095] Step 7: Handling normal everyday conversations

[0096] Device:

[0097] The device also collects speech data from everyday conversations and analyzes it using generative modeling techniques.

[0098] If no dangerous phrases are detected, the data is not processed as an alert but is saved as a log.

[0099] server:

[0100] The server periodically collects the stored log data and uses it, along with the analysis results, to help maintain a safe environment.

[0101] Step 8: Update the generative model

[0102] server:

[0103] The server periodically updates the generative AI model to accommodate new fraud patterns.

[0104] Device:

[0105] The device downloads and applies the updated generative AI model from the server.

[0106] This allows the system to detect cash card fraud in real time and ensure the safety of the elderly.

[0107] Example 1

[0108] Next, a description will be given of Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0109] In modern society, cash card fraud targeting the elderly is on the rise. This not only causes significant financial losses for these individuals, but also psychological damage. However, current security systems have difficulty detecting fraudulent activity in real time and taking immediate appropriate action. Therefore, a more advanced and rapid system for detecting and notifying fraud is needed.

[0110] The specific processing by the specific processing unit 290 of the data processing device 12 in the first embodiment is realized by the following means.

[0111] In this invention, the server includes an audio recording means for constantly collecting voice data, a generative model means for analyzing the collected voice data and evaluating the possibility of fraud in real time, a warning generation means for issuing an alert when the generative model means determines that there is a high possibility of fraud, and an information transmission means for notifying family members and the police of the issued alert. This makes it possible to quickly and accurately detect fraudulent acts targeting the elderly and to immediately notify family members and the police.

[0112] The "sound recording means" is a device for constantly collecting sound data around the user.

[0113] The "generative model means" is a device or software that analyzes collected voice data and evaluates the likelihood of fraud in real time in conjunction with a database of past fraud cases.

[0114] The "alert generation means" is a device for issuing an alert when the generative model means determines that there is a high possibility of fraud.

[0115] "Information transmission means" refers to a communication device or system for notifying family members or police of an alert that has been issued.

[0116] The "fraud case database" is a storage device that accumulates data on fraud cases that have occurred in the past, and that the generative model means refers to and analyzes this data.

[0117] An "alert" is warning information generated when the generative model means determines that there is a high possibility of fraud, and includes detailed information about the fraud and location information.

[0118] "Location information" is data indicating the geographic location of the point of origin, and is obtained using technology such as GPS.

[0119] The present invention is a system for detecting fraudulent activities targeting elderly people in real time and responding promptly. The system consists of an acoustic recording means, a generative model means, a warning generation means, and an information transmission means.

[0120] System Components

[0121] 1. Acoustic Recording Means

[0122] The device has a built-in microphone that constantly collects sounds around the user. This collected sound data is stored in a digital format. For example, the built-in high-sensitivity microphone records sound in WAV format.

[0123] 2. Generative Modeling Methods

[0124] The device is equipped with a generative AI model to analyze the collected voice data. This generative AI model converts the voice data into text using natural language processing (NLP) techniques and compares it with a database of past fraud cases. Specific examples include machine learning frameworks such as TensorFlow and PyTorch.

[0125] 3. Warning generation means

[0126] If the generative modeling method assesses the likelihood of fraud, the device immediately generates an alert that includes the text of the suspected fraudulent conversation, the time of occurrence, and location information (e.g., GPS data).

[0127] 4. Means of communication

[0128] The alert generated on the device is sent to a server, which then notifies registered family members and the police via a dedicated app, SMS, email, or other means.

[0129] Specific examples

[0130] Example 1: If you receive a fraudulent phone call

[0131] 1. Situation

[0132] The elderly user receives a call from an unknown number. The caller claims to be a bank employee and asks for confirmation of their cash card number and PIN.

[0133] 2. Operation

[0134] The device collects this conversation using a voice capture means and analyzes it using a generative model means. The generative AI model detects dangerous phrases such as "cash card" and "PIN number."

[0135] The generative model means determines that there is a high suspicion of fraud and activates the alert generation means, which generates and sends out an alert.

[0136] The alert is sent to a server via a communication device, and the server then notifies registered family members or police. Family members or police receive the alert through a dedicated app and can immediately contact the elderly person or call the police.

[0137] Example 2: Normal everyday conversation

[0138] 1. Situation

[0139] The user has everyday conversations with his family.

[0140] 2. Operation

[0141] The device also collects voice data and analyzes it using a generative model. If the generative AI model does not detect any particularly dangerous phrases, the voice data is safely stored in the system as a log and no alert is generated.

[0142] The server periodically collects log data and uses it, along with the analysis results, to improve the accuracy of the system.

[0143] Prompt Sentence Examples

[0144] "If families want to protect their elderly relatives from fraud, please tell us about the development of a system that can detect and notify them of fraudulent phone calls."

[0145] "Give us an example of how we could design a system that leverages generative AI models to prevent bank card fraud."

[0146] The system configuration and operating principles described above make it possible to quickly detect and immediately deal with fraudulent acts targeting the elderly, making this system particularly effective in ensuring the safety of the elderly.

[0147] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0148] System program processing flow

[0149] Step 1:

[0150] Initializing the device

[0151] Input: The device is powered on.

[0152] Processing: The device first loads the generative AI model into memory, using libraries such as TensorFlow or PyTorch.

[0153] Specific operation: The trained model file of the generative AI model is loaded and the model is initialized. Next, the microphone, which is the audio collection means, is enabled and set up so that audio data can be collected in digital format. Next, the system connects to the Internet using Wi-Fi or wired communication and synchronizes with the server.

[0154] Output: An initialized AI model, ready to collect audio.

[0155] Step 2:

[0156] Audio data collection

[0157] Input: Audio around the user.

[0158] Processing: The device continuously collects surrounding audio and stores it in digital format (e.g., WAV format).

[0159] Specific operation: The audio captured by the microphone undergoes pre-processing such as noise removal using digital signal processing (DSP), and is then input into the generative AI model.

[0160] Output: Preprocessed digital audio data.

[0161] Step 3:

[0162] Analysis of audio data

[0163] Input: Digital audio data collected in step 2.

[0164] Processing: The device feeds collected voice data into a generative AI model to detect phrases and patterns.

[0165] How it works: The generative AI model converts digital voice data into text and then uses natural language processing (NLP) techniques to detect potentially fraudulent phrases from a list.

[0166] Output: A fraud likelihood assessment score.

[0167] Step 4:

[0168] Generate alerts

[0169] Input: Analysis results from step 3 (likely fraudulent reputation score).

[0170] Processing: If the generative AI model determines that there is a high possibility of fraud, the alert generation means operates and generates an alert.

[0171] What it does: The alert will include the text of the allegedly fraudulent conversation, the time it occurred, and location information (GPS data). This information will be compiled into a single message.

[0172] Output: The generated alert message.

[0173] Step 5:

[0174] Alert Notification

[0175] Input: The alert message generated in step 4.

[0176] Processing: The terminal sends the generated alert message to the server.

[0177] How it works: Communication is via HTTP / HTTPS protocol and is encrypted for data security. The server analyzes the received alert and notifies registered family members and police.

[0178] Output: Notification message sent to family and police.

[0179] The above is an explanation of the specific operations, inputs, and outputs at each processing step. This system will enable us to quickly and effectively detect fraud targeting the elderly and respond immediately.

[0180] (Application example 1)

[0181] Next, a description will be given of Application Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0182] The risk of elderly people and general users becoming victims of cash card fraud and other fraudulent activities is increasing. In particular, fraudulent activities over the phone are becoming more sophisticated, and many users are becoming victims without realizing it. Therefore, there is a need for an effective system that can detect fraudulent activities in real time and immediately notify relevant parties and crime prevention agencies.

[0183] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 1 is realized by the following means.

[0184] In this invention, the server includes an audio collection means for constantly collecting audio data, an AI model generation means for analyzing the collected audio data and evaluating the possibility of fraudulent activity in real time, an alert generation means for issuing an alert when the AI ​​model generation means determines that there is a high possibility of fraudulent activity, and a notification means for notifying relevant parties and crime prevention agencies of the issued alert, thereby enabling fraudulent activity to be detected in real time and dealt with promptly.

[0185] "Audio collection means" refers to devices or technologies for constantly collecting surrounding audio data.

[0186] "Generative AI model means" refers to technology that uses a generative AI model to analyze collected voice data and assess the possibility of fraudulent activity in real time.

[0187] "Alert generation means" refers to a device or technology for issuing an alert when the generation AI model means determines that there is a high possibility of fraudulent activity.

[0188] "Notification means" refers to devices or technologies for notifying relevant parties and crime prevention agencies of the issued alert.

[0189] "Fraudulent acts" are acts such as cash card fraud that attempt to deceive users and illegally obtain money or personal information.

[0190] The "database of past misconduct cases" is a database that stores information about misconduct that has occurred in the past.

[0191] This invention aims to realize a security system for detecting and preventing fraudulent activities. The system combines a voice collection means, a generative AI model means, an alert generation means, and a notification means. Here, a specific example of this system will be described.

[0192] System Components

[0193] 1. Audio collection method

[0194] The device is equipped with a built-in microphone that constantly collects sounds around the user. This microphone has high sensitivity and noise-canceling capabilities, allowing it to collect clear audio data.

[0195] 2. Generative AI Model Means

[0196] The collected voice data is sent to an on-device generative AI modeling facility, which uses a pre-trained natural language processing model (e.g., GPT-4) to analyze the voice data and detect specific phrases and patterns associated with fraudulent activity. The model works in conjunction with a database of past fraud cases to improve detection accuracy.

[0197] 3. Alert Generation Methods

[0198] If the generation AI model means determines that there is a high possibility of fraudulent activity, the alert generation means immediately issues an alert, which includes the user's location information and details of the dangerous phrase, allowing for a prompt response.

[0199] 4. Means of notification

[0200] The alert is sent from the device to a server, which then notifies relevant parties and crime prevention agencies via SMS, email, a dedicated app, etc.

[0201] Example

[0202] When an elderly person receives a fraudulent phone call

[0203] When an elderly person hears phrases such as "cash card," "PIN number," or "bank employee" over the phone, the device's microphone collects the audio and sends it to a generative AI model. The model analyzes these phrases and, if it determines there is a high possibility of fraud, the alert generation means immediately sends out an alert. This alert also includes location information, allowing relevant parties and crime prevention agencies to respond quickly.

[0204] Use in everyday conversation

[0205] Similarly, when a user has a daily conversation with a family member on the phone, the device collects the audio and sends it to the generative AI model. If the model does not detect any particularly dangerous phrases, the audio data is securely stored as a log within the system. This allows the system to prevent false alarms and provide accurate information when needed.

[0206] Prompt Sentence Examples

[0207] Analyze audio data of conversations containing phrases such as "cash card," "PIN number," "bank employee," and "provided" to assess the likelihood of fraudulent activity.

[0208] In this way, the present invention provides a system that reduces the risk of users becoming involved in fraud or other fraudulent activity and allows for a quick and accurate response.

[0209] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0210] Step 1:

[0211] Initialization:

[0212] When the device boots up, it first loads the generative AI model, activates the audio capture method, prepares a microphone with high sensitivity and noise cancellation, establishes an internet connection, and synchronizes with the server.

[0213] Input: Start terminal

[0214] Output: Loading the generative AI model, enabling the audio capture method, and establishing an internet connection

[0215] Step 2:

[0216] Audio Collection:

[0217] The microphone constantly collects sounds around the user, and the audio data is converted into a digital format in real time and stored on the device.

[0218] Input: Ambient audio

[0219] Output: Digital audio data

[0220] Step 3:

[0221] Audio Analysis:

[0222] The collected voice data is sent to a generative AI model, which analyzes the speech based on prompts to detect specific phrases and patterns associated with fraudulent activity.

[0223] Input: Digital audio data

[0224] Output: The result of assessing the likelihood of fraud.

[0225] Step 4:

[0226] Alert generation:

[0227] If the generative AI model determines that fraud is likely, the alert generator will immediately send an alert, which will include the user's location and details of the dangerous phrase.

[0228] Input: Results of evaluation of potential fraud

[0229] Output: Alert information (location information, dangerous phrases)

[0230] Step 5:

[0231] notification:

[0232] The alert is sent from the device to a server, which then notifies relevant parties and crime prevention agencies via SMS, email, a dedicated app, etc.

[0233] Input: Alert information (location information, dangerous phrases)

[0234] Output: Notification to relevant parties and security agencies

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

[0236] The system of the present invention is an advanced crime prevention system for detecting and preventing cash card fraud in real time, and combines a voice collection means, a generative model means, an alert generation means, a notification means, and an emotion engine that recognizes the user's emotions. The detailed operation of the system and specific usage methods are described below.

[0237] System configuration

[0238] 1. Audio collection method

[0239] The device has a built-in microphone that constantly collects sounds around the user, and this sound data is processed digitally in real time.

[0240] 2. Generative Modeling Methods

[0241] To analyze the collected voice data, the device incorporates a generative AI model that uses machine learning algorithms to detect specific phrases and patterns in conversation.

[0242] 3. Emotion Engine

[0243] The device is equipped with an emotion engine that recognizes the user's emotions from collected voice data, and determines the user's emotional state from the tone, speed, and patterns of the voice.

[0244] 4. Alert Generation Methods

[0245] If the generative model and emotion engine determine that fraud is likely, the device generates an alert, which includes details about the fraud and its location.

[0246] 5. Means of notification

[0247] The generated alert is sent to the server via the notification method. The server receives the alert and notifies registered family members and the police. Notifications are sent via a dedicated app, SMS, or email.

[0248] Specific examples

[0249] 1. If you receive a fraudulent phone call

[0250] The elderly user receives a call from an unknown number. The caller claims to be a bank employee and asks for confirmation of their cash card number and PIN.

[0251] The device collects this conversation using a voice capture means and analyzes it using a generative model means. The generative AI model detects dangerous phrases such as "cash card" and "PIN number."

[0252] At the same time, the emotion engine recognizes changes in emotion from the tone and rate of the user's voice, and if a suspected change in emotion is detected, it feeds this data to the generative model means.

[0253] The generative modeling means takes the sentiment data into account to reassess the likelihood of fraud, and if suspicion of fraud increases, it activates the alert generation means, which immediately generates an alert.

[0254] The generated alert is sent to a server by a notification means, and the server sends notifications to registered family members and the police in real time.

[0255] Family members and police receive alerts through a dedicated app and can immediately contact the elderly person or go directly to the scene.

[0256] 2. Normal everyday conversation

[0257] When a user has a daily conversation with a family member over the phone, the terminal similarly collects voice data and analyzes it using the generative model means.

[0258] If no dangerous phrases are detected, and the results of the emotion engine's analysis of the user's emotions are normal, the data will not be processed as an alert, but will be saved as a log.

[0259] The server periodically collects the stored log data and uses it, along with the analysis results, to strengthen the security of the system.

[0260] Program processing

[0261] To make it easier to understand, the specific flow of the system's program processing is shown below:

[0262] 1. Initialize the device

[0263] When the device starts up, it loads the generative model means and emotion engine, activates the voice collection means, establishes an Internet connection, and starts communication with the server.

[0264] 2. Audio collection and analysis

[0265] The device constantly collects audio from the user's surroundings and feeds it into a generative AI model and emotion engine in real time. The generative model analyzes the audio data to detect specific phrases and patterns that may be indicative of fraud.

[0266] 3. Emotion recognition and evaluation

[0267] The emotion engine recognizes the user's emotions from the collected voice data and supplies changes in emotions to the generative model means.

[0268] 4. Alert Generation and Notification

[0269] If the generative model means assesses the likelihood of bank card fraud highly or if the emotion engine detects a suspicious change in emotion, the terminal generates an alert.

[0270] The alert will include detailed information about the fraud and location information and will be sent to the server via the notification method.

[0271] The server receives the alert and sends notifications to registered family members and police.

[0272] This allows the system to combine voice analysis and emotion recognition to detect cash card fraud in real time and ensure the safety of the elderly more precisely.

[0273] The processing flow will be explained below.

[0274] Processing flow of a system that combines emotion engines

[0275] Step 1: Initialization

[0276] Device:

[0277] When the device is powered on, it loads the generative model means and emotion engine and enables the voice collection means.

[0278] The device establishes an Internet connection and begins communicating with the server.

[0279] server:

[0280] The server accepts the connection request from the terminal and checks the authentication information.

[0281] User:

[0282] Users install a dedicated app and enter the necessary information (personal information, emergency contact information).

[0283] Step 2: Continuously collect audio data

[0284] Device:

[0285] The device constantly collects surrounding sounds through a microphone and processes them digitally in real time.

[0286] Step 3: Analyzing the audio data

[0287] Device:

[0288] The device feeds the collected voice data into a generative AI model for real-time analysis.

[0289] The generative modeling means detects specific phrases and patterns in the audio data that are indicative of fraud.

[0290] Step 4: Recognizing Emotional Data

[0291] Device:

[0292] At the same time, the device analyzes the voice data using an emotion engine to recognize the user's emotional state, which determines emotions based on the tone, speed, and patterns of the voice.

[0293] Step 5: Overall Scam Assessment

[0294] Device:

[0295] Based on the analysis results from the generative model and the recognition results from the emotion engine, the possibility of fraud is comprehensively evaluated.

[0296] The user's emotion data recognized by the emotion engine is supplied to the generative model means and reflected in the fraud probability assessment.

[0297] Step 6: Alert Generation

[0298] Device:

[0299] If the generative model means assesses the likelihood of bank card fraud highly or if the emotion engine detects a suspicious change in emotion, the terminal generates an alert.

[0300] The alert will include detailed information about the scam and its location.

[0301] Step 7: Sending an alert

[0302] Device:

[0303] The generated alert is sent to the server via a notification means.

[0304] server:

[0305] The server receives the alerts and records them in a database.

[0306] The server will then send notifications to registered family members and the police.

[0307] Step 8: Receive and respond to notifications

[0308] User (family / police):

[0309] Family members and police will receive alerts via a dedicated app, SMS or email.

[0310] Family members and police will check the alert and take necessary measures.

[0311] Step 9: Handling normal everyday conversations

[0312] Device:

[0313] The device also collects voice data from everyday conversations and analyzes it using a generative modeling method and emotion engine.

[0314] If no dangerous phrases are detected and the user's sentiment according to the sentiment engine is stable, the data is not processed as an alert but is stored as a log.

[0315] server:

[0316] The server periodically collects the stored log data and uses it, along with the analysis results, to help maintain a safe environment.

[0317] Step 10: Update the generative model

[0318] server:

[0319] The server periodically updates the generative AI model to accommodate new fraud patterns.

[0320] Device:

[0321] The device downloads and applies the updated generative AI model from the server.

[0322] Through these steps, this system combines emotion analysis and voice analysis to detect cash card fraud in real time and ensure the safety of the elderly with greater accuracy.

[0323] Example 2

[0324] Next, a description will be given of Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0325] Cash card fraud is a crime that particularly targets the elderly, and the damage it causes is serious. In addition, fraud methods are becoming more sophisticated, so traditional simple crime prevention systems are no longer sufficient to deal with it. Current technology does not have a system in place to detect signs of fraud in real time and prevent damage before it occurs.

[0326] The specific processing by the specific processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means.

[0327] In this invention, the server includes: a voice collection means for constantly collecting voice data; a generative model means for analyzing the collected voice data and evaluating the possibility of cash card fraud in real time; an emotion engine for analyzing the voice data to recognize emotions; an alert generation means for reevaluating the possibility of cash card fraud based on data from the generative model means and the emotion engine and issuing an alert if it is determined that fraud is highly likely; and a notification means for notifying family members and the police of the issued alert. This makes it possible to detect cash card fraud in real time by combining analysis of voice data and emotion recognition, and to prevent damage before it occurs by taking prompt action.

[0328] The "voice collection means" is hardware or software for constantly collecting voices around the user.

[0329] The "generative model means" is a system that includes a machine learning algorithm for analyzing collected voice data and assessing the likelihood of cash card fraud in real time.

[0330] The "Emotion Engine" is a software module that analyzes voice data to recognize the user's emotions. It determines the emotional state from the tone, rate, and patterns of the voice.

[0331] The "alert generation means" is a system that has the function of reevaluating the possibility of cash card fraud based on data from the generative model means and the emotion engine, and issuing an alert if it is determined that there is a high possibility of fraud.

[0332] "Notification methods" refer to the technology used to notify family members and the police of the alert. Notifications can be sent via a dedicated app, SMS, email, or other means.

[0333] "Location information" refers to geographical data of the point of origin, and this information is included in the alert.

[0334] "Fraud details" refers to specific information about cash card fraud that is included in the alert, such as the time the fraud occurred and the details of the fraud.

[0335] The "database of past cash card fraud cases" is a database that accumulates cases of cash card fraud that have occurred in the past.

[0336] The "user emotion database" is a database that stores past data on user emotions. It is used to operate the emotion engine.

[0337] The system of the present invention is an advanced security system for detecting and preventing cash card fraud in real time, and combines a voice collection means, a generative model means, an alert generation means, a notification means, and an emotion engine that recognizes user emotions. Specific embodiments for implementing the present invention are described in detail below.

[0338] System configuration

[0339] 1. Audio collection method

[0340] The device uses a built-in microphone to continuously collect sounds around the user, and this audio data is processed digitally in real time. For example, this is the case with microphones found on devices such as PCs, smartphones, and tablets.

[0341] 2. Generative Modeling Methods

[0342] The device incorporates a generative AI model (e.g., GPT-4) to analyze the collected voice data. This generative AI model uses machine learning algorithms to detect specific phrases and patterns in conversations, including risky phrases like "cash card" and "PIN number."

[0343] 3. Emotion Engine

[0344] The device is equipped with an emotion engine that recognizes the user's emotions from collected voice data. The emotion engine determines the user's emotional state from the tone, speed, and patterns of the voice, detecting changes such as anxiety or tension.

[0345] 4. Alert Generation Methods

[0346] If the device determines that fraud is likely based on information from the generative model means and emotion engine, it generates an alert. This alert includes detailed information about the fraud and location information. The alert generation means operates when a relevant dangerous phrase or suspicious emotion change is detected.

[0347] 5. Means of notification

[0348] The generated alert is sent to the server via a notification method. The server receives the alert and notifies registered family members and the police. Notifications are sent via a dedicated app, SMS, email, etc. Location information and detailed information about the fraud are included, urging immediate action.

[0349] Specific examples

[0350] 1. If you receive a fraudulent phone call

[0351] The elderly user receives a call from an unknown number. The caller claims to be a bank employee and asks for their cash card and PIN number.

[0352] The device collects this conversation using a voice capture method and analyzes it with a generative AI model, which detects dangerous phrases such as "cash card" and "PIN number."

[0353] At the same time, the emotion engine recognizes changes in emotions from the tone, rate, and patterns of the user's voice, detecting anxiety and tension.

[0354] The generative model determines if there is a high risk of fraud and generates an alert, which includes details of the fraud and location information.

[0355] The alert is sent to a server via the notification method, and the server then sends a real-time notification to family members or the police via a dedicated app, SMS, email, etc., allowing for immediate action.

[0356] Prompt Sentence Examples

[0357] By using prompt sentences like the one below for the generative AI model, you can verify and adjust the system's operation.

[0358] "Please give an example of how you would respond if you were asked to verify your bank card over the phone."

[0359] "How does the emotion engine recognize changes in emotion from the tone and rate of a user's voice?"

[0360] "What information is included in the generated alert?"

[0361] This system combines analysis of user voice data with emotion recognition to detect cash card fraud in real time and promote quick and appropriate responses.

[0362] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0363] Step 1:

[0364] When the device starts up, it loads the generative modeling means and emotion engine, activates the voice collection means, establishes an internet connection, and starts communication with the server, allowing the system to be fully functional.

[0365] Input: Power on

[0366] Output: System ready state

[0367] Step 2:

[0368] The device constantly collects the user's surrounding sounds through a built-in microphone. This audio data is processed digitally in real time and input to the generative modeling means and emotion engine.

[0369] Input: Audio around the user

[0370] Output: Digital audio data

[0371] Step 3:

[0372] The device's generative modeling means analyzes the collected voice data to detect specific phrases and patterns. A generative AI model (e.g., GPT-4) detects risky phrases such as "cash card" and "PIN number." This analysis assesses the likelihood of fraud.

[0373] Input: Digital audio data

[0374] Output: Detection results for specific phrases and fraud likelihood rating

[0375] Step 4:

[0376] The emotion engine analyzes the user's emotions from the voice data, determining the user's emotional state from the tone, rate, and patterns of the voice, and detecting emotional changes such as anxiety or tension.

[0377] Input: Digital audio data

[0378] Output: Evaluation result of the user's emotional state

[0379] Step 5:

[0380] The generative modeling means considers the emotion data from the emotion engine and reassess the risk of fraud. If it determines that fraud is likely, the alert generation means operates to generate an alert that includes detailed information and location information.

[0381] Input: Phrase detection results and emotional state evaluation results

[0382] Output: Alert generation decision and alert data

[0383] Step 6:

[0384] The generated alert is sent to the server via the notification method. The server receives the alert and notifies registered family members and the police. Notifications are sent via a dedicated app, SMS, email, etc.

[0385] Input: Alert data

[0386] Output: Notify family and police

[0387] Step 7:

[0388] The server evaluates responses from family members and police who receive the notification and, in some cases, provides further details or escalates the alert.

[0389] Input: Responses from family and police

[0390] Output: Additional details and escalation instructions

[0391] Through these processing steps, the system combines voice data analysis and emotion recognition to detect cash card fraud in real time and promptly respond.

[0392] (Application example 2)

[0393] Next, a description will be given of Application Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0394] In recent years, there has been an increase in cash card frauds targeting the elderly, and effective measures to prevent such damage are needed. Current security systems have difficulty detecting signs of fraud in real time, and are particularly ineffective against telephone fraud. Furthermore, because it is difficult for elderly people to recognize fraud themselves, more advanced fraud detection systems must be developed. Furthermore, fraud detection systems must be able to recognize users' emotions and take their changes into account for more accurate detection. This requires immediate and appropriate alerts in situations where there is a high possibility of fraud, and prompt notification to family members and the police.

[0395] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 2 is realized by the following means.

[0396] In this invention, the server includes voice collection means for constantly collecting voice data, generative model means for analyzing the collected voice data and evaluating the possibility of cash card fraud in real time, alert generation means for issuing an alert when the generative model means determines that there is a high possibility of cash card fraud, notification means for notifying family members and the police of the issued alert, an emotion engine for recognizing the user's emotions from the voice data, and means for reevaluating the possibility of fraud when the emotion engine determines that a change in emotion is suspicious. This enables highly accurate fraud detection that takes into account changes in the user's emotions and rapid issuance of alerts and notifications.

[0397] The "voice collection means" is a device or system for constantly collecting voices around the user.

[0398] A "generative model means" is a device or system that incorporates a machine learning algorithm for analyzing collected voice data and assessing the likelihood of cash card fraud in real time.

[0399] The "alert generation means" is a device or system that issues an alert when the generative model means determines that there is a high possibility of cash card fraud.

[0400] "Notification means" refers to a communication device or system for notifying family members or the police of an alert that has been issued.

[0401] An "emotion engine" is a device or algorithm that recognizes a user's emotions from voice data and determines changes in emotions.

[0402] A "means for reassessing" is a device or system that reassess the possibility of fraud when the emotion engine determines that a change in emotion is suspicious.

[0403] The system for implementing this invention is broadly composed of the following elements: a voice collection means, a generative model means, an emotion engine, an alert generation means, and a notification means. These elements are realized by combining multiple pieces of software and hardware.

[0404] 1. Audio collection method

[0405] The voice collection means uses a microphone built into a device such as a smartphone. It constantly collects voices around the user and converts them into data in real time. This voice data is then input into the generative model means and emotion engine in the next step.

[0406] 2. Generative Modeling Methods

[0407] The generative model means includes a machine learning algorithm for analyzing the collected voice data. Specifically, the generative AI model is used to detect specific phrases and patterns in the voice data that indicate possible bank card fraud. The model works in conjunction with a pre-trained database (past bank card fraud cases).

[0408] 3. Emotion Engine

[0409] The emotion engine includes algorithms for recognizing a user's emotional state from speech data. The emotion engine analyzes the tone, rate, and patterns of speech to determine changes in the user's emotions. This information is then used to assess the likelihood of fraud by means of a generative model.

[0410] 4. Alert Generation Methods

[0411] The alert generator operates based on information from the generative model and the emotion engine. If a fraud is deemed likely or a suspicious change in emotion is detected, an alert is generated. The alert includes detailed information about the fraud and data about the user's current location.

[0412] 5. Means of notification

[0413] The notification method sends the generated alert to a server, which processes the received alert and sends notifications to registered family members and the police via a dedicated application, SMS, email, etc.

[0414] Hardware and software used

[0415] Hardware: Smartphone (built-in microphone)

[0416] Software: Audio collection library (e.g., some_audio_library), generative AI model library (e.g., some_ml_library), emotion recognition library (e.g., emotion_recognition_library)

[0417] Cloud server: Alert notifications and log data storage

[0418] Adding specific examples

[0419] For example, consider a scenario in which an elderly person receives a scam call from an unknown number. If the generative AI model and emotion recognition engine detect a suspicious phrase or emotional change, the following prompt will be generated:

[0420] Scam Guardian Alert: Call from unknown number may be a bank card scam. Details: Scam details, Location: Current location

[0421] This prompt will alert you and your family immediately so you can take any necessary action.

[0422] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[0423] Step 1:

[0424] Start audio collection

[0425] When the device is started, it activates the built-in microphone and starts the audio collection method. It continuously collects audio data around the user and converts it into a digital format in real time. The input is the surrounding audio, and the output is the digitized audio data.

[0426] Step 2:

[0427] Analysis of audio data

[0428] The device inputs the collected voice data into a generative AI model means. The generative AI model uses a machine learning algorithm to analyze the voice data and detect specific phrases or patterns that may indicate possible bank card fraud. The input is the digitized voice data, and the output is the results of the analyzed phrases and patterns.

[0429] Step 3:

[0430] Emotion recognition

[0431] At the same time, the device also inputs the voice data into the emotion engine, which analyzes the tone, speed, and pattern of the voice to recognize the user's emotional state. The input is digitized voice data, and the output is the result of the user's emotional state assessment.

[0432] Step 4:

[0433] Fraud likelihood assessment

[0434] The terminal integrates information from the generative model means and the emotion engine. The generative model means reevaluates the emotion data provided by the emotion engine and determines the likelihood of fraud. The input is the analyzed phrases, patterns, and emotion data, and the output is the result of the fraud likelihood assessment.

[0435] Step 5:

[0436] Generate alerts

[0437] If a fraud possibility is determined, the device immediately generates an alert using the alert generation means. This alert includes detailed fraud information and location information. The input is the fraud possibility assessment result, and the output is the generated alert.

[0438] Step 6:

[0439] Alert Notification

[0440] The generated alert is sent to the server via the notification means. The server processes the received alert and sends a notification to registered family members or police. Notifications are sent via a dedicated app, SMS, email, etc. The input is the generated alert, and the output is the notification to family members or police.

[0441] Step 7:

[0442] Facilitating user and family support

[0443] The family member can check the received notification and immediately contact the user or arrange for on-site visit if necessary. The input is the received alert notification, and the output is the appropriate response action.

[0444] As a result of this, the system will be able to combine voice data and emotional data to detect cash card fraud with high accuracy and send notifications to users, their families, and the police in real time.

[0445] The specific processing unit 290 transmits the result of the specific processing to the smart device 14. In the smart device 14, the control unit 46A causes the output device 40 to output the result of the specific processing. The microphone 38B acquires audio indicating a user input regarding the result of the specific processing. The control unit 46A transmits audio data indicating the user input acquired by the microphone 38B to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the audio data.

[0446] The data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of the data generation model 58 is ChatGPT (Internet Search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search <url: https: gemini.google.com ?hl="ja">) and other generation AIs. The data generation model 58 is obtained by performing deep learning on a neural network. A prompt including an instruction is input to the data generation model 58, and inference data such as voice data indicating voice, text data indicating text, and image data indicating an image is also input. The data generation model 58 performs inference on the input inference data in accordance with the instruction indicated by the prompt, and outputs the inference result in a data format such as voice data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[0447] In the above embodiment, an example in which the specific process is performed by the data processing device 12 has been given, but the technology of the present disclosure is not limited to this, and the specific process may be performed by the smart device 14.

[0448] [Second embodiment]

[0449] FIG. 3 shows an example of the configuration of a data processing system 210 according to the second embodiment.

[0450] 3, the data processing system 210 includes the data processing device 12 and smart glasses 214. An example of the data processing device 12 is a server.

[0451] The data processing device 12 includes a computer 22, a database 24, and a communication I / F 26. The computer 22 is an example of a "computer" according to the technology of the present disclosure. The computer 22 includes a processor 28, a RAM 30, and a storage 32. The processor 28, the RAM 30, and the storage 32 are connected to a bus 34. The database 24 and the communication I / F 26 are also connected to the bus 34. The communication I / F 26 is connected to a network 54. Examples of the network 54 include a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0452] The smart glasses 214 include a computer 36, a microphone 238, a speaker 240, a camera 42, and a communication I / F 44. The computer 36 includes a processor 46, a RAM 48, and a storage 50. The processor 46, the RAM 48, and the storage 50 are connected to a bus 52. The microphone 238, the speaker 240, and the camera 42 are also connected to the bus 52.

[0453] The microphone 238 receives instructions and the like from the user 20 by receiving voice uttered by the user 20. The microphone 238 captures the voice uttered by the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio in accordance with instructions from the processor 46.

[0454] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an imaging element such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the surroundings of user 20 (for example, an imaging range defined by an angle of view equivalent to the field of vision of a typical healthy person).

[0455] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 are responsible for the exchange of various information between the processor 46 and the processor 28 via the network 54. The exchange of various information between the processor 46 and the processor 28 using the communication I / Fs 44 and 26 is carried out in a secure state.

[0456] Fig. 4 shows an example of the main functions of the data processing device 12 and the smart glasses 214. As shown in Fig. 4, in the data processing device 12, a specific process is performed by the processor 28. A specific process program 56 is stored in the storage 32.

[0457] The specific processing program 56 is an example of a "program" according to the technology of the present disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific processing is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific processing program 56 executed on the RAM 30.

[0458] The storage 32 stores a data generation model 58 and an emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[0459] In the smart glasses 214, the reception output process is performed by the processor 46. A reception output program 60 is stored in the storage 50. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output process is realized by the processor 46 operating as the control unit 46A in accordance with the reception output program 60 executed on the RAM 48.

[0460] Next, a description will be given of the identification process performed by the identification processing unit 290 of the data processing device 12. In the following description, the data processing device 12 will be referred to as the "server" and the smart glasses 214 will be referred to as the "terminal."

[0461] The system of the present invention is an advanced crime prevention system for detecting and preventing cash card fraud in real time. The detailed operation of the system and specific methods of use are described below.

[0462] System configuration

[0463] 1. Audio collection method

[0464] The device has a built-in microphone that constantly collects sounds around the user, and the collected sound data is processed digitally in real time.

[0465] 2. Generative Modeling Methods

[0466] To analyze the collected voice data, the device incorporates a generative AI model that uses pre-trained machine learning algorithms to detect specific phrases and patterns in speech.

[0467] 3. Alert Generation Methods

[0468] If the generative AI model determines that there is a high probability of bank card fraud, the device will immediately generate an alert, which will include specific details of the suspected fraudulent conversation and its location.

[0469] 4. Means of notification

[0470] The alert generated on the device is sent to a server, which then notifies registered family members and the police via a dedicated app, SMS, or email.

[0471] Specific examples

[0472] 1. If you receive a fraudulent phone call

[0473] The elderly user receives a call from an unknown number. The caller claims to be a bank employee and asks for confirmation of their cash card number and PIN.

[0474] The device collects this conversation using a voice capture means and analyzes it using a generative model means. The generative AI model detects dangerous phrases such as "cash card" and "PIN number."

[0475] The generative model means determines that there is a high suspicion of fraud and activates the alert generation means, which immediately generates an alert containing detailed information about the fraud and location information.

[0476] The generated alert is sent to a server by a notification means, and the server sends notifications to registered family members and the police in real time.

[0477] Family members and police receive alerts through a dedicated app and can immediately contact the elderly person or go directly to the scene.

[0478] 2. Normal everyday conversation

[0479] When a user has a daily conversation with a family member over the phone, the terminal similarly collects voice data and analyzes it using the generative model means.

[0480] If the generative AI model does not detect any particularly dangerous phrases, the audio data is securely stored in the system as a log and no alert is generated.

[0481] The server periodically collects log data and uses the analysis results to strengthen security.

[0482] Program processing

[0483] To make it easier to understand, the specific flow of the system's program processing is shown below:

[0484] 1. Initialize the device

[0485] When the device boots up, it loads the generative model means, activates the audio collection means, establishes an internet connection, and synchronizes with the server.

[0486] 2. Audio collection and analysis

[0487] The device constantly collects audio from the user's surroundings and feeds it into a generative AI model in real time, which then analyzes the audio data to detect phrases that could be fraudulent.

[0488] 3. Alert generation and notification

[0489] If the generative model means highly evaluates the possibility of cash card fraud, the alert generation means immediately generates an alert and transmits it to the server.

[0490] The server receives the alert and sends notifications to registered family members and police.

[0491] This will prevent fraud and ensure the safety of the elderly.

[0492] The processing flow will be explained below.

[0493] Program processing flow

[0494] Step 1: Initialization

[0495] Device:

[0496] When the terminal is powered on, it loads the generative model means and enables the audio collection means.

[0497] The device establishes an Internet connection and begins communicating with the server.

[0498] server:

[0499] The server accepts the connection request from the terminal and checks the authentication information.

[0500] User:

[0501] Users install a dedicated app and enter the necessary information (personal information, emergency contact information).

[0502] Step 2: Continuously collect audio data

[0503] Device:

[0504] The device constantly collects surrounding sounds through a microphone, and this collected audio data is processed digitally in real time.

[0505] Step 3: Analyzing the audio data

[0506] Device:

[0507] The device feeds the collected voice data into a generative AI model for real-time analysis.

[0508] The generative modeling means detects specific phrases and patterns in the audio data that are indicative of fraud.

[0509] Step 4: Generate an alert if fraud is suspected

[0510] Device:

[0511] If the generative AI model determines that fraud is highly suspected, the device will generate an alert.

[0512] The alert will include detailed information about the scam and its location.

[0513] Step 5: Sending an alert

[0514] Device:

[0515] The generated alert is sent to the server via a notification means.

[0516] server:

[0517] The server receives the alerts and records them in a database.

[0518] The server will then send notifications to registered family members and the police.

[0519] Step 6: Receive and respond to notifications

[0520] User (family / police):

[0521] Family members and police will receive alerts via a dedicated app, SMS or email.

[0522] Family members and police will check the alert and take necessary measures.

[0523] Step 7: Handling normal everyday conversations

[0524] Device:

[0525] The device also collects speech data from everyday conversations and analyzes it using generative modeling techniques.

[0526] If no dangerous phrases are detected, the data is not processed as an alert but is saved as a log.

[0527] server:

[0528] The server periodically collects the stored log data and uses it, along with the analysis results, to help maintain a safe environment.

[0529] Step 8: Update the generative model

[0530] server:

[0531] The server periodically updates the generative AI model to accommodate new fraud patterns.

[0532] Device:

[0533] The device downloads and applies the updated generative AI model from the server.

[0534] This allows the system to detect cash card fraud in real time and ensure the safety of the elderly.

[0535] Example 1

[0536] Next, a description will be given of Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[0537] In modern society, cash card fraud targeting the elderly is on the rise. This not only causes significant financial losses for these individuals, but also psychological damage. However, current security systems have difficulty detecting fraudulent activity in real time and taking immediate appropriate action. Therefore, a more advanced and rapid system for detecting and notifying fraud is needed.

[0538] The specific processing by the specific processing unit 290 of the data processing device 12 in the first embodiment is realized by the following means.

[0539] In this invention, the server includes an audio recording means for constantly collecting voice data, a generative model means for analyzing the collected voice data and evaluating the possibility of fraud in real time, a warning generation means for issuing an alert when the generative model means determines that there is a high possibility of fraud, and an information transmission means for notifying family members and the police of the issued alert. This makes it possible to quickly and accurately detect fraudulent acts targeting the elderly and to immediately notify family members and the police.

[0540] The "sound recording means" is a device for constantly collecting sound data around the user.

[0541] The "generative model means" is a device or software that analyzes collected voice data and evaluates the likelihood of fraud in real time in conjunction with a database of past fraud cases.

[0542] The "alert generation means" is a device for issuing an alert when the generative model means determines that there is a high possibility of fraud.

[0543] "Information transmission means" refers to a communication device or system for notifying family members or police of an alert that has been issued.

[0544] The "fraud case database" is a storage device that accumulates data on fraud cases that have occurred in the past, and that the generative model means refers to and analyzes this data.

[0545] An "alert" is warning information generated when the generative model means determines that there is a high possibility of fraud, and includes detailed information about the fraud and location information.

[0546] "Location information" is data indicating the geographic location of the point of origin, and is obtained using technology such as GPS.

[0547] The present invention is a system for detecting fraudulent activities targeting elderly people in real time and responding promptly. The system consists of an acoustic recording means, a generative model means, a warning generation means, and an information transmission means.

[0548] System Components

[0549] 1. Acoustic Recording Means

[0550] The device has a built-in microphone that constantly collects sounds around the user. This collected sound data is stored in a digital format. For example, the built-in high-sensitivity microphone records sound in WAV format.

[0551] 2. Generative Modeling Methods

[0552] The device is equipped with a generative AI model to analyze the collected voice data. This generative AI model converts the voice data into text using natural language processing (NLP) techniques and compares it with a database of past fraud cases. Specific examples include machine learning frameworks such as TensorFlow and PyTorch.

[0553] 3. Warning generation means

[0554] If the generative modeling method assesses the likelihood of fraud, the device immediately generates an alert that includes the text of the suspected fraudulent conversation, the time of occurrence, and location information (e.g., GPS data).

[0555] 4. Means of communication

[0556] The alert generated on the device is sent to a server, which then notifies registered family members and the police via a dedicated app, SMS, email, or other means.

[0557] Specific examples

[0558] Example 1: If you receive a fraudulent phone call

[0559] 1. Situation

[0560] The elderly user receives a call from an unknown number. The caller claims to be a bank employee and asks for confirmation of their cash card number and PIN.

[0561] 2. Operation

[0562] The device collects this conversation using a voice capture means and analyzes it using a generative model means. The generative AI model detects dangerous phrases such as "cash card" and "PIN number."

[0563] The generative model means determines that there is a high suspicion of fraud and activates the alert generation means, which generates and sends out an alert.

[0564] The alert is sent to a server via a communication device, and the server then notifies registered family members or police. Family members or police receive the alert through a dedicated app and can immediately contact the elderly person or call the police.

[0565] Example 2: Normal everyday conversation

[0566] 1. Situation

[0567] The user has everyday conversations with his family.

[0568] 2. Operation

[0569] The device also collects voice data and analyzes it using a generative model. If the generative AI model does not detect any particularly dangerous phrases, the voice data is safely stored in the system as a log and no alert is generated.

[0570] The server periodically collects log data and uses it, along with the analysis results, to improve the accuracy of the system.

[0571] Prompt Sentence Examples

[0572] "If families want to protect their elderly relatives from fraud, please tell us about the development of a system that can detect and notify them of fraudulent phone calls."

[0573] "Give us an example of how we could design a system that leverages generative AI models to prevent bank card fraud."

[0574] The system configuration and operating principles described above make it possible to quickly detect and immediately deal with fraudulent acts targeting the elderly, making this system particularly effective in ensuring the safety of the elderly.

[0575] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0576] System program processing flow

[0577] Step 1:

[0578] Initializing the device

[0579] Input: The device is powered on.

[0580] Processing: The device first loads the generative AI model into memory, using libraries such as TensorFlow or PyTorch.

[0581] Specific operation: The trained model file of the generative AI model is loaded and the model is initialized. Next, the microphone, which is the audio collection means, is enabled and set up so that audio data can be collected in digital format. Next, the system connects to the Internet using Wi-Fi or wired communication and synchronizes with the server.

[0582] Output: An initialized AI model, ready to collect audio.

[0583] Step 2:

[0584] Audio data collection

[0585] Input: Audio around the user.

[0586] Processing: The device continuously collects surrounding audio and stores it in digital format (e.g., WAV format).

[0587] Specific operation: The audio captured by the microphone undergoes pre-processing such as noise removal using digital signal processing (DSP), and is then input into the generative AI model.

[0588] Output: Preprocessed digital audio data.

[0589] Step 3:

[0590] Analysis of audio data

[0591] Input: Digital audio data collected in step 2.

[0592] Processing: The device feeds collected voice data into a generative AI model to detect phrases and patterns.

[0593] How it works: The generative AI model converts digital voice data into text and then uses natural language processing (NLP) techniques to detect potentially fraudulent phrases from a list.

[0594] Output: A fraud likelihood assessment score.

[0595] Step 4:

[0596] Generate alerts

[0597] Input: Analysis results from step 3 (likely fraudulent reputation score).

[0598] Processing: If the generative AI model determines that there is a high possibility of fraud, the alert generation means operates and generates an alert.

[0599] What it does: The alert will include the text of the allegedly fraudulent conversation, the time it occurred, and location information (GPS data). This information will be compiled into a single message.

[0600] Output: The generated alert message.

[0601] Step 5:

[0602] Alert Notification

[0603] Input: The alert message generated in step 4.

[0604] Processing: The terminal sends the generated alert message to the server.

[0605] How it works: Communication is via HTTP / HTTPS protocol and is encrypted for data security. The server analyzes the received alert and notifies registered family members and police.

[0606] Output: Notification message sent to family and police.

[0607] The above is an explanation of the specific operations, inputs, and outputs at each processing step. This system will enable us to quickly and effectively detect fraud targeting the elderly and respond immediately.

[0608] (Application example 1)

[0609] Next, a description will be given of Application Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[0610] The risk of elderly people and general users becoming victims of cash card fraud and other fraudulent activities is increasing. In particular, fraudulent activities over the phone are becoming more sophisticated, and many users are becoming victims without realizing it. Therefore, there is a need for an effective system that can detect fraudulent activities in real time and immediately notify relevant parties and crime prevention agencies.

[0611] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 1 is realized by the following means.

[0612] In this invention, the server includes an audio collection means for constantly collecting audio data, an AI model generation means for analyzing the collected audio data and evaluating the possibility of fraudulent activity in real time, an alert generation means for issuing an alert when the AI ​​model generation means determines that there is a high possibility of fraudulent activity, and a notification means for notifying relevant parties and crime prevention agencies of the issued alert, thereby enabling fraudulent activity to be detected in real time and dealt with promptly.

[0613] "Audio collection means" refers to devices or technologies for constantly collecting surrounding audio data.

[0614] "Generative AI model means" refers to technology that uses a generative AI model to analyze collected voice data and assess the possibility of fraudulent activity in real time.

[0615] "Alert generation means" refers to a device or technology for issuing an alert when the generation AI model means determines that there is a high possibility of fraudulent activity.

[0616] "Notification means" refers to devices or technologies for notifying relevant parties and crime prevention agencies of the issued alert.

[0617] "Fraudulent acts" are acts such as cash card fraud that attempt to deceive users and illegally obtain money or personal information.

[0618] The "database of past misconduct cases" is a database that stores information about misconduct that has occurred in the past.

[0619] This invention aims to realize a security system for detecting and preventing fraudulent activities. The system combines a voice collection means, a generative AI model means, an alert generation means, and a notification means. Here, a specific example of this system will be described.

[0620] System Components

[0621] 1. Audio collection method

[0622] The device is equipped with a built-in microphone that constantly collects sounds around the user. This microphone has high sensitivity and noise-canceling capabilities, allowing it to collect clear audio data.

[0623] 2. Generative AI Model Means

[0624] The collected voice data is sent to an on-device generative AI modeling facility, which uses a pre-trained natural language processing model (e.g., GPT-4) to analyze the voice data and detect specific phrases and patterns associated with fraudulent activity. The model works in conjunction with a database of past fraud cases to improve detection accuracy.

[0625] 3. Alert Generation Methods

[0626] If the generation AI model means determines that there is a high possibility of fraudulent activity, the alert generation means immediately issues an alert, which includes the user's location information and details of the dangerous phrase, allowing for a prompt response.

[0627] 4. Means of notification

[0628] The alert is sent from the device to a server, which then notifies relevant parties and crime prevention agencies via SMS, email, a dedicated app, etc.

[0629] Example

[0630] When an elderly person receives a fraudulent phone call

[0631] When an elderly person hears phrases such as "cash card," "PIN number," or "bank employee" over the phone, the device's microphone collects the audio and sends it to a generative AI model. The model analyzes these phrases and, if it determines there is a high possibility of fraud, the alert generation means immediately sends out an alert. This alert also includes location information, allowing relevant parties and crime prevention agencies to respond quickly.

[0632] Use in everyday conversation

[0633] Similarly, when a user has a daily conversation with a family member on the phone, the device collects the audio and sends it to the generative AI model. If the model does not detect any particularly dangerous phrases, the audio data is securely stored as a log within the system. This allows the system to prevent false alarms and provide accurate information when needed.

[0634] Prompt Sentence Examples

[0635] Analyze audio data of conversations containing phrases such as "cash card," "PIN number," "bank employee," and "provided" to assess the likelihood of fraudulent activity.

[0636] In this way, the present invention provides a system that reduces the risk of users becoming involved in fraud or other fraudulent activity and allows for a quick and accurate response.

[0637] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0638] Step 1:

[0639] Initialization:

[0640] When the device boots up, it first loads the generative AI model, activates the audio capture method, prepares a microphone with high sensitivity and noise cancellation, establishes an internet connection, and synchronizes with the server.

[0641] Input: Start terminal

[0642] Output: Loading the generative AI model, enabling the audio capture method, and establishing an internet connection

[0643] Step 2:

[0644] Audio Collection:

[0645] The microphone constantly collects sounds around the user, and the audio data is converted into a digital format in real time and stored on the device.

[0646] Input: Ambient audio

[0647] Output: Digital audio data

[0648] Step 3:

[0649] Audio Analysis:

[0650] The collected voice data is sent to a generative AI model, which analyzes the speech based on prompts to detect specific phrases and patterns associated with fraudulent activity.

[0651] Input: Digital audio data

[0652] Output: The result of assessing the likelihood of fraud.

[0653] Step 4:

[0654] Alert generation:

[0655] If the generative AI model determines that fraud is likely, the alert generator will immediately send an alert, which will include the user's location and details of the dangerous phrase.

[0656] Input: Results of evaluation of potential fraud

[0657] Output: Alert information (location information, dangerous phrases)

[0658] Step 5:

[0659] notification:

[0660] The alert is sent from the device to a server, which then notifies relevant parties and crime prevention agencies via SMS, email, a dedicated app, etc.

[0661] Input: Alert information (location information, dangerous phrases)

[0662] Output: Notification to relevant parties and security agencies

[0663] Furthermore, an emotion engine that estimates the user's emotion may be further combined. That is, the identification processing unit 290 may estimate the user's emotion using the emotion identification model 59, and perform identification processing using the user's emotion.

[0664] The system of the present invention is an advanced crime prevention system for detecting and preventing cash card fraud in real time, and combines a voice collection means, a generative model means, an alert generation means, a notification means, and an emotion engine that recognizes the user's emotions. The detailed operation of the system and specific usage methods are described below.

[0665] System configuration

[0666] 1. Audio collection method

[0667] The device has a built-in microphone that constantly collects sounds around the user, and this sound data is processed digitally in real time.

[0668] 2. Generative Modeling Methods

[0669] To analyze the collected voice data, the device incorporates a generative AI model that uses machine learning algorithms to detect specific phrases and patterns in conversation.

[0670] 3. Emotion Engine

[0671] The device is equipped with an emotion engine that recognizes the user's emotions from collected voice data, and determines the user's emotional state from the tone, speed, and patterns of the voice.

[0672] 4. Alert Generation Methods

[0673] If the generative model and emotion engine determine that fraud is likely, the device generates an alert, which includes details about the fraud and its location.

[0674] 5. Means of notification

[0675] The generated alert is sent to the server via the notification method. The server receives the alert and notifies registered family members and the police. Notifications are sent via a dedicated app, SMS, or email.

[0676] Specific examples

[0677] 1. If you receive a fraudulent phone call

[0678] The elderly user receives a call from an unknown number. The caller claims to be a bank employee and asks for confirmation of their cash card number and PIN.

[0679] The device collects this conversation using a voice capture means and analyzes it using a generative model means. The generative AI model detects dangerous phrases such as "cash card" and "PIN number."

[0680] At the same time, the emotion engine recognizes changes in emotion from the tone and rate of the user's voice, and if a suspected change in emotion is detected, it feeds this data to the generative model means.

[0681] The generative modeling means takes the sentiment data into account to reassess the likelihood of fraud, and if suspicion of fraud increases, it activates the alert generation means, which immediately generates an alert.

[0682] The generated alert is sent to a server by a notification means, and the server sends notifications to registered family members and the police in real time.

[0683] Family members and police receive alerts through a dedicated app and can immediately contact the elderly person or go directly to the scene.

[0684] 2. Normal everyday conversation

[0685] When a user has a daily conversation with a family member over the phone, the terminal similarly collects voice data and analyzes it using the generative model means.

[0686] If no dangerous phrases are detected, and the results of the emotion engine's analysis of the user's emotions are normal, the data will not be processed as an alert, but will be saved as a log.

[0687] The server periodically collects the stored log data and uses it, along with the analysis results, to strengthen the security of the system.

[0688] Program processing

[0689] To make it easier to understand, the specific flow of the system's program processing is shown below:

[0690] 1. Initialize the device

[0691] When the device starts up, it loads the generative model means and emotion engine, activates the voice collection means, establishes an Internet connection, and starts communication with the server.

[0692] 2. Audio collection and analysis

[0693] The device constantly collects audio from the user's surroundings and feeds it into a generative AI model and emotion engine in real time. The generative model analyzes the audio data to detect specific phrases and patterns that may be indicative of fraud.

[0694] 3. Emotion recognition and evaluation

[0695] The emotion engine recognizes the user's emotions from the collected voice data and supplies changes in emotions to the generative model means.

[0696] 4. Alert Generation and Notification

[0697] If the generative model means assesses the likelihood of bank card fraud highly or if the emotion engine detects a suspicious change in emotion, the terminal generates an alert.

[0698] The alert will include detailed information about the fraud and location information and will be sent to the server via the notification method.

[0699] The server receives the alert and sends notifications to registered family members and police.

[0700] This allows the system to combine voice analysis and emotion recognition to detect cash card fraud in real time and ensure the safety of the elderly more precisely.

[0701] The processing flow will be explained below.

[0702] Processing flow of a system that combines emotion engines

[0703] Step 1: Initialization

[0704] Device:

[0705] When the device is powered on, it loads the generative model means and emotion engine and enables the voice collection means.

[0706] The device establishes an Internet connection and begins communicating with the server.

[0707] server:

[0708] The server accepts the connection request from the terminal and checks the authentication information.

[0709] User:

[0710] Users install a dedicated app and enter the necessary information (personal information, emergency contact information).

[0711] Step 2: Continuously collect audio data

[0712] Device:

[0713] The device constantly collects surrounding sounds through a microphone and processes them digitally in real time.

[0714] Step 3: Analyzing the audio data

[0715] Device:

[0716] The device feeds the collected voice data into a generative AI model for real-time analysis.

[0717] The generative modeling means detects specific phrases and patterns in the audio data that are indicative of fraud.

[0718] Step 4: Recognizing Emotional Data

[0719] Device:

[0720] At the same time, the device analyzes the voice data using an emotion engine to recognize the user's emotional state, which determines emotions based on the tone, speed, and patterns of the voice.

[0721] Step 5: Overall Scam Assessment

[0722] Device:

[0723] Based on the analysis results from the generative model and the recognition results from the emotion engine, the possibility of fraud is comprehensively evaluated.

[0724] The user's emotion data recognized by the emotion engine is supplied to the generative model means and reflected in the fraud probability assessment.

[0725] Step 6: Alert Generation

[0726] Device:

[0727] If the generative model means assesses the likelihood of bank card fraud highly or if the emotion engine detects a suspicious change in emotion, the terminal generates an alert.

[0728] The alert will include detailed information about the scam and its location.

[0729] Step 7: Sending an alert

[0730] Device:

[0731] The generated alert is sent to the server via a notification means.

[0732] server:

[0733] The server receives the alerts and records them in a database.

[0734] The server will then send notifications to registered family members and the police.

[0735] Step 8: Receive and respond to notifications

[0736] User (family / police):

[0737] Family members and police will receive alerts via a dedicated app, SMS or email.

[0738] Family members and police will check the alert and take necessary measures.

[0739] Step 9: Handling normal everyday conversations

[0740] Device:

[0741] The device also collects voice data from everyday conversations and analyzes it using a generative modeling method and emotion engine.

[0742] If no dangerous phrases are detected and the user's sentiment according to the sentiment engine is stable, the data is not processed as an alert but is stored as a log.

[0743] server:

[0744] The server periodically collects the stored log data and uses it, along with the analysis results, to help maintain a safe environment.

[0745] Step 10: Update the generative model

[0746] server:

[0747] The server periodically updates the generative AI model to accommodate new fraud patterns.

[0748] Device:

[0749] The device downloads and applies the updated generative AI model from the server.

[0750] Through these steps, this system combines emotion analysis and voice analysis to detect cash card fraud in real time and ensure the safety of the elderly with greater accuracy.

[0751] Example 2

[0752] Next, a description will be given of Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[0753] Cash card fraud is a crime that particularly targets the elderly, and the damage it causes is serious. In addition, fraud methods are becoming more sophisticated, so traditional simple crime prevention systems are no longer sufficient to deal with it. Current technology does not have a system in place to detect signs of fraud in real time and prevent damage before it occurs.

[0754] The specific processing by the specific processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means.

[0755] In this invention, the server includes: a voice collection means for constantly collecting voice data; a generative model means for analyzing the collected voice data and evaluating the possibility of cash card fraud in real time; an emotion engine for analyzing the voice data to recognize emotions; an alert generation means for reevaluating the possibility of cash card fraud based on data from the generative model means and the emotion engine and issuing an alert if it is determined that fraud is highly likely; and a notification means for notifying family members and the police of the issued alert. This makes it possible to detect cash card fraud in real time by combining analysis of voice data and emotion recognition, and to prevent damage before it occurs by taking prompt action.

[0756] The "voice collection means" is hardware or software for constantly collecting voices around the user.

[0757] The "generative model means" is a system that includes a machine learning algorithm for analyzing collected voice data and assessing the likelihood of cash card fraud in real time.

[0758] The "Emotion Engine" is a software module that analyzes voice data to recognize the user's emotions. It determines the emotional state from the tone, rate, and patterns of the voice.

[0759] The "alert generation means" is a system that has the function of reevaluating the possibility of cash card fraud based on data from the generative model means and the emotion engine, and issuing an alert if it is determined that there is a high possibility of fraud.

[0760] "Notification methods" refer to the technology used to notify family members and the police of the alert. Notifications can be sent via a dedicated app, SMS, email, or other means.

[0761] "Location information" refers to geographical data of the point of origin, and this information is included in the alert.

[0762] "Fraud details" refers to specific information about cash card fraud that is included in the alert, such as the time the fraud occurred and the details of the fraud.

[0763] The "database of past cash card fraud cases" is a database that accumulates cases of cash card fraud that have occurred in the past.

[0764] The "user emotion database" is a database that stores past data on user emotions. It is used to operate the emotion engine.

[0765] The system of the present invention is an advanced security system for detecting and preventing cash card fraud in real time, and combines a voice collection means, a generative model means, an alert generation means, a notification means, and an emotion engine that recognizes user emotions. Specific embodiments for implementing the present invention are described in detail below.

[0766] System configuration

[0767] 1. Audio collection method

[0768] The device uses a built-in microphone to continuously collect sounds around the user, and this audio data is processed digitally in real time. For example, this is the case with microphones found on devices such as PCs, smartphones, and tablets.

[0769] 2. Generative Modeling Methods

[0770] The device incorporates a generative AI model (e.g., GPT-4) to analyze the collected voice data. This generative AI model uses machine learning algorithms to detect specific phrases and patterns in conversations, including risky phrases like "cash card" and "PIN number."

[0771] 3. Emotion Engine

[0772] The device is equipped with an emotion engine that recognizes the user's emotions from collected voice data. The emotion engine determines the user's emotional state from the tone, speed, and patterns of the voice, detecting changes such as anxiety or tension.

[0773] 4. Alert Generation Methods

[0774] If the device determines that fraud is likely based on information from the generative model means and emotion engine, it generates an alert. This alert includes detailed information about the fraud and location information. The alert generation means operates when a relevant dangerous phrase or suspicious emotion change is detected.

[0775] 5. Means of notification

[0776] The generated alert is sent to the server via a notification method. The server receives the alert and notifies registered family members and the police. Notifications are sent via a dedicated app, SMS, email, etc. Location information and detailed information about the fraud are included, urging immediate action.

[0777] Specific examples

[0778] 1. If you receive a fraudulent phone call

[0779] The elderly user receives a call from an unknown number. The caller claims to be a bank employee and asks for their cash card and PIN number.

[0780] The device collects this conversation using a voice capture method and analyzes it with a generative AI model, which detects dangerous phrases such as "cash card" and "PIN number."

[0781] At the same time, the emotion engine recognizes changes in emotions from the tone, rate, and patterns of the user's voice, detecting anxiety and tension.

[0782] The generative model determines if there is a high risk of fraud and generates an alert, which includes details of the fraud and location information.

[0783] The alert is sent to a server via the notification method, and the server then sends a real-time notification to family members or the police via a dedicated app, SMS, email, etc., allowing for immediate action.

[0784] Prompt Sentence Examples

[0785] By using prompt sentences like the one below for the generative AI model, you can verify and adjust the system's operation.

[0786] "Please give an example of how you would respond if you were asked to verify your bank card over the phone."

[0787] "How does the emotion engine recognize changes in emotion from the tone and rate of a user's voice?"

[0788] "What information is included in the generated alert?"

[0789] This system combines analysis of user voice data with emotion recognition to detect cash card fraud in real time and promote quick and appropriate responses.

[0790] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0791] Step 1:

[0792] When the device starts up, it loads the generative modeling means and emotion engine, activates the voice collection means, establishes an internet connection, and starts communication with the server, allowing the system to be fully functional.

[0793] Input: Power on

[0794] Output: System ready state

[0795] Step 2:

[0796] The device constantly collects the user's surrounding sounds through a built-in microphone. This audio data is processed digitally in real time and input to the generative modeling means and emotion engine.

[0797] Input: Audio around the user

[0798] Output: Digital audio data

[0799] Step 3:

[0800] The device's generative modeling means analyzes the collected voice data to detect specific phrases and patterns. A generative AI model (e.g., GPT-4) detects risky phrases such as "cash card" and "PIN number." This analysis assesses the likelihood of fraud.

[0801] Input: Digital audio data

[0802] Output: Detection results for specific phrases and fraud likelihood rating

[0803] Step 4:

[0804] The emotion engine analyzes the user's emotions from the voice data, determining the user's emotional state from the tone, rate, and patterns of the voice, and detecting emotional changes such as anxiety or tension.

[0805] Input: Digital audio data

[0806] Output: Evaluation result of the user's emotional state

[0807] Step 5:

[0808] The generative modeling means considers the emotion data from the emotion engine and reassess the risk of fraud. If it determines that fraud is likely, the alert generation means operates to generate an alert that includes detailed information and location information.

[0809] Input: Phrase detection results and emotional state evaluation results

[0810] Output: Alert generation decision and alert data

[0811] Step 6:

[0812] The generated alert is sent to the server via the notification method. The server receives the alert and notifies registered family members and the police. Notifications are sent via a dedicated app, SMS, email, etc.

[0813] Input: Alert data

[0814] Output: Notify family and police

[0815] Step 7:

[0816] The server evaluates responses from family members and police who receive the notification and, in some cases, provides further details or escalates the alert.

[0817] Input: Responses from family and police

[0818] Output: Additional details and escalation instructions

[0819] Through these processing steps, the system combines voice data analysis and emotion recognition to detect cash card fraud in real time and promptly respond.

[0820] (Application example 2)

[0821] Next, a description will be given of Application Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[0822] In recent years, there has been an increase in cash card frauds targeting the elderly, and effective measures to prevent such damage are needed. Current security systems have difficulty detecting signs of fraud in real time, and are particularly ineffective against telephone fraud. Furthermore, because it is difficult for elderly people to recognize fraud themselves, more advanced fraud detection systems must be developed. Furthermore, fraud detection systems must be able to recognize users' emotions and take their changes into account for more accurate detection. This requires immediate and appropriate alerts in situations where there is a high possibility of fraud, and prompt notification to family members and the police.

[0823] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 2 is realized by the following means.

[0824] In this invention, the server includes voice collection means for constantly collecting voice data, generative model means for analyzing the collected voice data and evaluating the possibility of cash card fraud in real time, alert generation means for issuing an alert when the generative model means determines that there is a high possibility of cash card fraud, notification means for notifying family members and the police of the issued alert, an emotion engine for recognizing the user's emotions from the voice data, and means for reevaluating the possibility of fraud when the emotion engine determines that a change in emotion is suspicious. This enables highly accurate fraud detection that takes into account changes in the user's emotions and rapid issuance of alerts and notifications.

[0825] The "voice collection means" is a device or system for constantly collecting voices around the user.

[0826] A "generative model means" is a device or system that incorporates a machine learning algorithm for analyzing collected voice data and assessing the likelihood of cash card fraud in real time.

[0827] The "alert generation means" is a device or system that issues an alert when the generative model means determines that there is a high possibility of cash card fraud.

[0828] "Notification means" refers to a communication device or system for notifying family members or the police of an alert that has been issued.

[0829] An "emotion engine" is a device or algorithm that recognizes a user's emotions from voice data and determines changes in emotions.

[0830] A "means for reassessing" is a device or system that reassess the possibility of fraud when the emotion engine determines that a change in emotion is suspicious.

[0831] The system for implementing this invention is broadly composed of the following elements: a voice collection means, a generative model means, an emotion engine, an alert generation means, and a notification means. These elements are realized by combining multiple pieces of software and hardware.

[0832] 1. Audio collection method

[0833] The voice collection means uses a microphone built into a device such as a smartphone. It constantly collects voices around the user and converts them into data in real time. This voice data is then input into the generative model means and emotion engine in the next step.

[0834] 2. Generative Modeling Methods

[0835] The generative model means includes a machine learning algorithm for analyzing the collected voice data. Specifically, the generative AI model is used to detect specific phrases and patterns in the voice data that indicate possible bank card fraud. The model works in conjunction with a pre-trained database (past bank card fraud cases).

[0836] 3. Emotion Engine

[0837] The emotion engine includes algorithms for recognizing a user's emotional state from speech data. The emotion engine analyzes the tone, rate, and patterns of speech to determine changes in the user's emotions. This information is then used to assess the likelihood of fraud by means of a generative model.

[0838] 4. Alert Generation Methods

[0839] The alert generator operates based on information from the generative model and the emotion engine. If a fraud is deemed likely or a suspicious change in emotion is detected, an alert is generated. The alert includes detailed information about the fraud and data about the user's current location.

[0840] 5. Means of notification

[0841] The notification method sends the generated alert to a server, which processes the received alert and sends notifications to registered family members and the police via a dedicated application, SMS, email, etc.

[0842] Hardware and software used

[0843] Hardware: Smartphone (built-in microphone)

[0844] Software: Audio collection library (e.g., some_audio_library), generative AI model library (e.g., some_ml_library), emotion recognition library (e.g., emotion_recognition_library)

[0845] Cloud server: Alert notifications and log data storage

[0846] Adding specific examples

[0847] For example, consider a scenario in which an elderly person receives a scam call from an unknown number. If the generative AI model and emotion recognition engine detect a suspicious phrase or emotional change, the following prompt will be generated:

[0848] Scam Guardian Alert: Call from unknown number may be a bank card scam. Details: Scam details, Location: Current location

[0849] This prompt will alert you and your family immediately so you can take any necessary action.

[0850] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[0851] Step 1:

[0852] Start audio collection

[0853] When the device is started, it activates the built-in microphone and starts the audio collection method. It continuously collects audio data around the user and converts it into a digital format in real time. The input is the surrounding audio, and the output is the digitized audio data.

[0854] Step 2:

[0855] Analysis of audio data

[0856] The device inputs the collected voice data into a generative AI model means. The generative AI model uses a machine learning algorithm to analyze the voice data and detect specific phrases or patterns that may indicate possible bank card fraud. The input is the digitized voice data, and the output is the results of the analyzed phrases and patterns.

[0857] Step 3:

[0858] Emotion recognition

[0859] At the same time, the device also inputs the voice data into the emotion engine, which analyzes the tone, speed, and pattern of the voice to recognize the user's emotional state. The input is digitized voice data, and the output is the result of the user's emotional state assessment.

[0860] Step 4:

[0861] Fraud likelihood assessment

[0862] The terminal integrates information from the generative model means and the emotion engine. The generative model means reevaluates the emotion data provided by the emotion engine and determines the likelihood of fraud. The input is the analyzed phrases, patterns, and emotion data, and the output is the result of the fraud likelihood assessment.

[0863] Step 5:

[0864] Generate alerts

[0865] If a fraud possibility is determined, the device immediately generates an alert using the alert generation means. This alert includes detailed fraud information and location information. The input is the fraud possibility assessment result, and the output is the generated alert.

[0866] Step 6:

[0867] Alert Notification

[0868] The generated alert is sent to the server via the notification means. The server processes the received alert and sends a notification to registered family members or police. Notifications are sent via a dedicated app, SMS, email, etc. The input is the generated alert, and the output is the notification to family members or police.

[0869] Step 7:

[0870] Facilitating user and family support

[0871] The family member can check the received notification and immediately contact the user or arrange for on-site visit if necessary. The input is the received alert notification, and the output is the appropriate response action.

[0872] As a result of this, the system will be able to combine voice data and emotional data to detect cash card fraud with high accuracy and send notifications to users, their families, and the police in real time.

[0873] The specific processing unit 290 transmits the result of the specific processing to the smart glasses 214. In the smart glasses 214, the control unit 46A causes the speaker 240 to output the result of the specific processing. The microphone 238 acquires audio indicating a user input regarding the result of the specific processing. The control unit 46A transmits audio data indicating the user input acquired by the microphone 238 to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the audio data.

[0874] The data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of the data generation model 58 is ChatGPT (Internet Search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search <url: https: gemini.google.com ?hl="ja">) and other generation AIs. The data generation model 58 is obtained by performing deep learning on a neural network. A prompt including an instruction is input to the data generation model 58, and inference data such as voice data indicating voice, text data indicating text, and image data indicating an image is also input. The data generation model 58 performs inference on the input inference data in accordance with the instruction indicated by the prompt, and outputs the inference result in a data format such as voice data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[0875] In the above embodiment, an example in which the specific processing is performed by the data processing device 12 has been given, but the technology of the present disclosure is not limited to this, and the specific processing may be performed by the smart glasses 214.

[0876] [Third embodiment]

[0877] FIG. 5 shows an example of the configuration of a data processing system 310 according to the third embodiment.

[0878] 5, the data processing system 310 includes the data processing device 12 and a headset type terminal 314. An example of the data processing device 12 is a server.

[0879] The data processing device 12 includes a computer 22, a database 24, and a communication I / F 26. The computer 22 is an example of a "computer" according to the technology of the present disclosure. The computer 22 includes a processor 28, a RAM 30, and a storage 32. The processor 28, the RAM 30, and the storage 32 are connected to a bus 34. The database 24 and the communication I / F 26 are also connected to the bus 34. The communication I / F 26 is connected to a network 54. Examples of the network 54 include a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0880] The headset type terminal 314 includes a computer 36, a microphone 238, a speaker 240, a camera 42, a communication I / F 44, and a display 343. The computer 36 includes a processor 46, a RAM 48, and a storage 50. The processor 46, the RAM 48, and the storage 50 are connected to a bus 52. The microphone 238, the speaker 240, the camera 42, and the display 343 are also connected to the bus 52.

[0881] The microphone 238 receives instructions and the like from the user 20 by receiving voice uttered by the user 20. The microphone 238 captures the voice uttered by the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio in accordance with instructions from the processor 46.

[0882] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an imaging element such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the surroundings of user 20 (for example, an imaging range defined by an angle of view equivalent to the field of vision of a typical healthy person).

[0883] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 are responsible for the exchange of various information between the processor 46 and the processor 28 via the network 54. The exchange of various information between the processor 46 and the processor 28 using the communication I / Fs 44 and 26 is carried out in a secure state.

[0884] Fig. 6 shows an example of the main functions of the data processing device 12 and the headset type terminal 314. As shown in Fig. 6, in the data processing device 12, a specific process is performed by the processor 28. A specific process program 56 is stored in the storage 32.

[0885] The specific processing program 56 is an example of a "program" according to the technology of the present disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific processing is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific processing program 56 executed on the RAM 30.

[0886] The storage 32 stores a data generation model 58 and an emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[0887] In the headset type terminal 314, a reception output process is performed by the processor 46. A reception output program 60 is stored in the storage 50. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output process is realized by the processor 46 operating as the control unit 46A in accordance with the reception output program 60 executed on the RAM 48.

[0888] Next, a description will be given of the identification process performed by the identification processing unit 290 of the data processing device 12. In the following description, the data processing device 12 will be referred to as the "server" and the headset type terminal 314 will be referred to as the "terminal."

[0889] The system of the present invention is an advanced crime prevention system for detecting and preventing cash card fraud in real time. The detailed operation of the system and specific methods of use are described below.

[0890] System configuration

[0891] 1. Audio collection method

[0892] The device has a built-in microphone that constantly collects sounds around the user, and the collected sound data is processed digitally in real time.

[0893] 2. Generative Modeling Methods

[0894] To analyze the collected voice data, the device incorporates a generative AI model that uses pre-trained machine learning algorithms to detect specific phrases and patterns in speech.

[0895] 3. Alert Generation Methods

[0896] If the generative AI model determines that there is a high probability of bank card fraud, the device will immediately generate an alert, which will include specific details of the suspected fraudulent conversation and its location.

[0897] 4. Means of notification

[0898] The alert generated on the device is sent to a server, which then notifies registered family members and the police via a dedicated app, SMS, or email.

[0899] Specific examples

[0900] 1. If you receive a fraudulent phone call

[0901] The elderly user receives a call from an unknown number. The caller claims to be a bank employee and asks for confirmation of their cash card number and PIN.

[0902] The device collects this conversation using a voice capture means and analyzes it using a generative model means. The generative AI model detects dangerous phrases such as "cash card" and "PIN number."

[0903] The generative model means determines that there is a high suspicion of fraud and activates the alert generation means, which immediately generates an alert containing detailed information about the fraud and location information.

[0904] The generated alert is sent to a server by a notification means, and the server sends notifications to registered family members and the police in real time.

[0905] Family members and police receive alerts through a dedicated app and can immediately contact the elderly person or go directly to the scene.

[0906] 2. Normal everyday conversation

[0907] When a user has a daily conversation with a family member over the phone, the terminal similarly collects voice data and analyzes it using the generative model means.

[0908] If the generative AI model does not detect any particularly dangerous phrases, the audio data is securely stored in the system as a log and no alert is generated.

[0909] The server periodically collects log data and uses the analysis results to strengthen security.

[0910] Program processing

[0911] To make it easier to understand, the specific flow of the system's program processing is shown below:

[0912] 1. Initialize the device

[0913] When the device boots up, it loads the generative model means, activates the audio collection means, establishes an internet connection, and synchronizes with the server.

[0914] 2. Audio collection and analysis

[0915] The device constantly collects audio from the user's surroundings and feeds it into a generative AI model in real time, which then analyzes the audio data to detect phrases that could be fraudulent.

[0916] 3. Alert generation and notification

[0917] If the generative model means highly evaluates the possibility of cash card fraud, the alert generation means immediately generates an alert and transmits it to the server.

[0918] The server receives the alert and sends notifications to registered family members and police.

[0919] This will prevent fraud and ensure the safety of the elderly.

[0920] The processing flow will be explained below.

[0921] Program processing flow

[0922] Step 1: Initialization

[0923] Device:

[0924] When the terminal is powered on, it loads the generative model means and enables the audio collection means.

[0925] The device establishes an Internet connection and begins communicating with the server.

[0926] server:

[0927] The server accepts the connection request from the terminal and checks the authentication information.

[0928] User:

[0929] Users install a dedicated app and enter the necessary information (personal information, emergency contact information).

[0930] Step 2: Continuously collect audio data

[0931] Device:

[0932] The device constantly collects surrounding sounds through a microphone, and this collected audio data is processed digitally in real time.

[0933] Step 3: Analyzing the audio data

[0934] Device:

[0935] The device feeds the collected voice data into a generative AI model for real-time analysis.

[0936] The generative modeling means detects specific phrases and patterns in the audio data that are indicative of fraud.

[0937] Step 4: Generate an alert if fraud is suspected

[0938] Device:

[0939] If the generative AI model determines that fraud is highly suspected, the device will generate an alert.

[0940] The alert will include detailed information about the scam and its location.

[0941] Step 5: Sending an alert

[0942] Device:

[0943] The generated alert is sent to the server via a notification means.

[0944] server:

[0945] The server receives the alerts and records them in a database.

[0946] The server will then send notifications to registered family members and the police.

[0947] Step 6: Receive and respond to notifications

[0948] User (family / police):

[0949] Family members and police will receive alerts via a dedicated app, SMS or email.

[0950] Family members and police will check the alert and take necessary measures.

[0951] Step 7: Handling normal everyday conversations

[0952] Device:

[0953] The device also collects speech data from everyday conversations and analyzes it using generative modeling techniques.

[0954] If no dangerous phrases are detected, the data is not processed as an alert but is saved as a log.

[0955] server:

[0956] The server periodically collects the stored log data and uses it, along with the analysis results, to help maintain a safe environment.

[0957] Step 8: Update the generative model

[0958] server:

[0959] The server periodically updates the generative AI model to accommodate new fraud patterns.

[0960] Device:

[0961] The device downloads and applies the updated generative AI model from the server.

[0962] This allows the system to detect cash card fraud in real time and ensure the safety of the elderly.

[0963] Example 1

[0964] Next, a description will be given of Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal."

[0965] In modern society, cash card fraud targeting the elderly is on the rise. This not only causes significant financial losses for these individuals, but also psychological damage. However, current security systems have difficulty detecting fraudulent activity in real time and taking immediate appropriate action. Therefore, a more advanced and rapid system for detecting and notifying fraud is needed.

[0966] The specific processing by the specific processing unit 290 of the data processing device 12 in the first embodiment is realized by the following means.

[0967] In this invention, the server includes an audio recording means for constantly collecting voice data, a generative model means for analyzing the collected voice data and evaluating the possibility of fraud in real time, a warning generation means for issuing an alert when the generative model means determines that there is a high possibility of fraud, and an information transmission means for notifying family members and the police of the issued alert. This makes it possible to quickly and accurately detect fraudulent acts targeting the elderly and to immediately notify family members and the police.

[0968] The "sound recording means" is a device for constantly collecting sound data around the user.

[0969] The "generative model means" is a device or software that analyzes collected voice data and evaluates the likelihood of fraud in real time in conjunction with a database of past fraud cases.

[0970] The "alert generation means" is a device for issuing an alert when the generative model means determines that there is a high possibility of fraud.

[0971] "Information transmission means" refers to a communication device or system for notifying family members or police of an alert that has been issued.

[0972] The "fraud case database" is a storage device that accumulates data on fraud cases that have occurred in the past, and that the generative model means refers to and analyzes this data.

[0973] An "alert" is warning information generated when the generative model means determines that there is a high possibility of fraud, and includes detailed information about the fraud and location information.

[0974] "Location information" is data indicating the geographic location of the point of origin, and is obtained using technology such as GPS.

[0975] The present invention is a system for detecting fraudulent activities targeting elderly people in real time and responding promptly. The system consists of an acoustic recording means, a generative model means, a warning generation means, and an information transmission means.

[0976] System Components

[0977] 1. Acoustic Recording Means

[0978] The device has a built-in microphone that constantly collects sounds around the user. This collected sound data is stored in a digital format. For example, the built-in high-sensitivity microphone records sound in WAV format.

[0979] 2. Generative Modeling Methods

[0980] The device is equipped with a generative AI model to analyze the collected voice data. This generative AI model converts the voice data into text using natural language processing (NLP) techniques and compares it with a database of past fraud cases. Specific examples include machine learning frameworks such as TensorFlow and PyTorch.

[0981] 3. Warning generation means

[0982] If the generative modeling method assesses the likelihood of fraud, the device immediately generates an alert that includes the text of the suspected fraudulent conversation, the time of occurrence, and location information (e.g., GPS data).

[0983] 4. Means of communication

[0984] The alert generated on the device is sent to a server, which then notifies registered family members and the police via a dedicated app, SMS, email, or other means.

[0985] Specific examples

[0986] Example 1: If you receive a fraudulent phone call

[0987] 1. Situation

[0988] The elderly user receives a call from an unknown number. The caller claims to be a bank employee and asks for confirmation of their cash card number and PIN.

[0989] 2. Operation

[0990] The device collects this conversation using a voice capture means and analyzes it using a generative model means. The generative AI model detects dangerous phrases such as "cash card" and "PIN number."

[0991] The generative model means determines that there is a high suspicion of fraud and activates the alert generation means, which generates and sends out an alert.

[0992] The alert is sent to a server via a communication device, and the server then notifies registered family members or police. Family members or police receive the alert through a dedicated app and can immediately contact the elderly person or call the police.

[0993] Example 2: Normal everyday conversation

[0994] 1. Situation

[0995] The user has everyday conversations with his family.

[0996] 2. Operation

[0997] The device also collects voice data and analyzes it using a generative model. If the generative AI model does not detect any particularly dangerous phrases, the voice data is safely stored in the system as a log and no alert is generated.

[0998] The server periodically collects log data and uses it, along with the analysis results, to improve the accuracy of the system.

[0999] Prompt Sentence Examples

[1000] "If families want to protect their elderly relatives from fraud, please tell us about the development of a system that can detect and notify them of fraudulent phone calls."

[1001] "Give us an example of how we could design a system that leverages generative AI models to prevent bank card fraud."

[1002] The system configuration and operating principles described above make it possible to quickly detect and immediately deal with fraudulent acts targeting the elderly, making this system particularly effective in ensuring the safety of the elderly.

[1003] The flow of the identification process in the first embodiment will be described with reference to FIG.

[1004] System program processing flow

[1005] Step 1:

[1006] Initializing the device

[1007] Input: The device is powered on.

[1008] Processing: The device first loads the generative AI model into memory, using libraries such as TensorFlow or PyTorch.

[1009] Specific operation: The trained model file of the generative AI model is loaded and the model is initialized. Next, the microphone, which is the audio collection means, is enabled and set up so that audio data can be collected in digital format. Next, the system connects to the Internet using Wi-Fi or wired communication and synchronizes with the server.

[1010] Output: An initialized AI model, ready to collect audio.

[1011] Step 2:

[1012] Audio data collection

[1013] Input: Audio around the user.

[1014] Processing: The device continuously collects surrounding audio and stores it in digital format (e.g., WAV format).

[1015] Specific operation: The audio captured by the microphone undergoes pre-processing such as noise removal using digital signal processing (DSP), and is then input into the generative AI model.

[1016] Output: Preprocessed digital audio data.

[1017] Step 3:

[1018] Analysis of audio data

[1019] Input: Digital audio data collected in step 2.

[1020] Processing: The device feeds collected voice data into a generative AI model to detect phrases and patterns.

[1021] How it works: The generative AI model converts digital voice data into text and then uses natural language processing (NLP) techniques to detect potentially fraudulent phrases from a list.

[1022] Output: A fraud likelihood assessment score.

[1023] Step 4:

[1024] Generate alerts

[1025] Input: Analysis results from step 3 (likely fraudulent reputation score).

[1026] Processing: If the generative AI model determines that there is a high possibility of fraud, the alert generation means operates and generates an alert.

[1027] What it does: The alert will include the text of the allegedly fraudulent conversation, the time it occurred, and location information (GPS data). This information will be compiled into a single message.

[1028] Output: The generated alert message.

[1029] Step 5:

[1030] Alert Notification

[1031] Input: The alert message generated in step 4.

[1032] Processing: The terminal sends the generated alert message to the server.

[1033] How it works: Communication is via HTTP / HTTPS protocol and is encrypted for data security. The server analyzes the received alert and notifies registered family members and police.

[1034] Output: Notification message sent to family and police.

[1035] The above is an explanation of the specific operations, inputs, and outputs at each processing step. This system will enable us to quickly and effectively detect fraud targeting the elderly and respond immediately.

[1036] (Application example 1)

[1037] Next, a description will be given of Application Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal."

[1038] The risk of elderly people and general users becoming victims of cash card fraud and other fraudulent activities is increasing. In particular, fraudulent activities over the phone are becoming more sophisticated, and many users are becoming victims without realizing it. Therefore, there is a need for an effective system that can detect fraudulent activities in real time and immediately notify relevant parties and crime prevention agencies.

[1039] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 1 is realized by the following means.

[1040] In this invention, the server includes an audio collection means for constantly collecting audio data, an AI model generation means for analyzing the collected audio data and evaluating the possibility of fraudulent activity in real time, an alert generation means for issuing an alert when the AI ​​model generation means determines that there is a high possibility of fraudulent activity, and a notification means for notifying relevant parties and crime prevention agencies of the issued alert, thereby enabling fraudulent activity to be detected in real time and dealt with promptly.

[1041] "Audio collection means" refers to devices or technologies for constantly collecting surrounding audio data.

[1042] "Generative AI model means" refers to technology that uses a generative AI model to analyze collected voice data and assess the possibility of fraudulent activity in real time.

[1043] "Alert generation means" refers to a device or technology for issuing an alert when the generation AI model means determines that there is a high possibility of fraudulent activity.

[1044] "Notification means" refers to devices or technologies for notifying relevant parties and crime prevention agencies of the issued alert.

[1045] "Fraudulent acts" are acts such as cash card fraud that attempt to deceive users and illegally obtain money or personal information.

[1046] The "database of past misconduct cases" is a database that stores information about misconduct that has occurred in the past.

[1047] This invention aims to realize a security system for detecting and preventing fraudulent activities. The system combines a voice collection means, a generative AI model means, an alert generation means, and a notification means. Here, a specific example of this system will be described.

[1048] System Components

[1049] 1. Audio collection method

[1050] The device is equipped with a built-in microphone that constantly collects sounds around the user. This microphone has high sensitivity and noise-canceling capabilities, allowing it to collect clear audio data.

[1051] 2. Generative AI Model Means

[1052] The collected voice data is sent to an on-device generative AI modeling facility, which uses a pre-trained natural language processing model (e.g., GPT-4) to analyze the voice data and detect specific phrases and patterns associated with fraudulent activity. The model works in conjunction with a database of past fraud cases to improve detection accuracy.

[1053] 3. Alert Generation Methods

[1054] If the generation AI model means determines that there is a high possibility of fraudulent activity, the alert generation means immediately issues an alert, which includes the user's location information and details of the dangerous phrase, allowing for a prompt response.

[1055] 4. Means of notification

[1056] The alert is sent from the device to a server, which then notifies relevant parties and crime prevention agencies via SMS, email, a dedicated app, etc.

[1057] Example

[1058] When an elderly person receives a fraudulent phone call

[1059] When an elderly person hears phrases such as "cash card," "PIN number," or "bank employee" over the phone, the device's microphone collects the audio and sends it to a generative AI model. The model analyzes these phrases and, if it determines there is a high possibility of fraud, the alert generation means immediately sends out an alert. This alert also includes location information, allowing relevant parties and crime prevention agencies to respond quickly.

[1060] Use in everyday conversation

[1061] Similarly, when a user has a daily conversation with a family member on the phone, the device collects the audio and sends it to the generative AI model. If the model does not detect any particularly dangerous phrases, the audio data is securely stored as a log within the system. This allows the system to prevent false alarms and provide accurate information when needed.

[1062] Prompt Sentence Examples

[1063] Analyze audio data of conversations containing phrases such as "cash card," "PIN number," "bank employee," and "provided" to assess the likelihood of fraudulent activity.

[1064] In this way, the present invention provides a system that reduces the risk of users becoming involved in fraud or other fraudulent activity and allows for a quick and accurate response.

[1065] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[1066] Step 1:

[1067] Initialization:

[1068] When the device boots up, it first loads the generative AI model, activates the audio capture method, prepares a microphone with high sensitivity and noise cancellation, establishes an internet connection, and synchronizes with the server.

[1069] Input: Start terminal

[1070] Output: Loading the generative AI model, enabling the audio capture method, and establishing an internet connection

[1071] Step 2:

[1072] Audio Collection:

[1073] The microphone constantly collects sounds around the user, and the audio data is converted into a digital format in real time and stored on the device.

[1074] Input: Ambient audio

[1075] Output: Digital audio data

[1076] Step 3:

[1077] Audio Analysis:

[1078] The collected voice data is sent to a generative AI model, which analyzes the speech based on prompts to detect specific phrases and patterns associated with fraudulent activity.

[1079] Input: Digital audio data

[1080] Output: The result of assessing the likelihood of fraud.

[1081] Step 4:

[1082] Alert generation:

[1083] If the generative AI model determines that fraud is likely, the alert generator will immediately send an alert, which will include the user's location and details of the dangerous phrase.

[1084] Input: Results of evaluation of potential fraud

[1085] Output: Alert information (location information, dangerous phrases)

[1086] Step 5:

[1087] notification:

[1088] The alert is sent from the device to a server, which then notifies relevant parties and crime prevention agencies via SMS, email, a dedicated app, etc.

[1089] Input: Alert information (location information, dangerous phrases)

[1090] Output: Notification to relevant parties and security agencies

[1091] Furthermore, an emotion engine that estimates the user's emotion may be further combined. That is, the identification processing unit 290 may estimate the user's emotion using the emotion identification model 59, and perform identification processing using the user's emotion.

[1092] The system of the present invention is an advanced crime prevention system for detecting and preventing cash card fraud in real time, and combines a voice collection means, a generative model means, an alert generation means, a notification means, and an emotion engine that recognizes the user's emotions. The detailed operation of the system and specific usage methods are described below.

[1093] System configuration

[1094] 1. Audio collection method

[1095] The device has a built-in microphone that constantly collects sounds around the user, and this sound data is processed digitally in real time.

[1096] 2. Generative Modeling Methods

[1097] To analyze the collected voice data, the device incorporates a generative AI model that uses machine learning algorithms to detect specific phrases and patterns in conversation.

[1098] 3. Emotion Engine

[1099] The device is equipped with an emotion engine that recognizes the user's emotions from collected voice data, and determines the user's emotional state from the tone, speed, and patterns of the voice.

[1100] 4. Alert Generation Methods

[1101] If the generative model and emotion engine determine that fraud is likely, the device generates an alert, which includes details about the fraud and its location.

[1102] 5. Means of notification

[1103] The generated alert is sent to the server via the notification method. The server receives the alert and notifies registered family members and the police. Notifications are sent via a dedicated app, SMS, or email.

[1104] Specific examples

[1105] 1. If you receive a fraudulent phone call

[1106] The elderly user receives a call from an unknown number. The caller claims to be a bank employee and asks for confirmation of their cash card number and PIN.

[1107] The device collects this conversation using a voice capture means and analyzes it using a generative model means. The generative AI model detects dangerous phrases such as "cash card" and "PIN number."

[1108] At the same time, the emotion engine recognizes changes in emotion from the tone and rate of the user's voice, and if a suspected change in emotion is detected, it feeds this data to the generative model means.

[1109] The generative modeling means takes the sentiment data into account to reassess the likelihood of fraud, and if suspicion of fraud increases, it activates the alert generation means, which immediately generates an alert.

[1110] The generated alert is sent to a server by a notification means, and the server sends notifications to registered family members and the police in real time.

[1111] Family members and police receive alerts through a dedicated app and can immediately contact the elderly person or go directly to the scene.

[1112] 2. Normal everyday conversation

[1113] When a user has a daily conversation with a family member over the phone, the terminal similarly collects voice data and analyzes it using the generative model means.

[1114] If no dangerous phrases are detected, and the results of the emotion engine's analysis of the user's emotions are normal, the data will not be processed as an alert, but will be saved as a log.

[1115] The server periodically collects the stored log data and uses it, along with the analysis results, to strengthen the security of the system.

[1116] Program processing

[1117] To make it easier to understand, the specific flow of the system's program processing is shown below:

[1118] 1. Initialize the device

[1119] When the device starts up, it loads the generative model means and emotion engine, activates the voice collection means, establishes an Internet connection, and starts communication with the server.

[1120] 2. Audio collection and analysis

[1121] The device constantly collects audio from the user's surroundings and feeds it into a generative AI model and emotion engine in real time. The generative model analyzes the audio data to detect specific phrases and patterns that may be indicative of fraud.

[1122] 3. Emotion recognition and evaluation

[1123] The emotion engine recognizes the user's emotions from the collected voice data and supplies changes in emotions to the generative model means.

[1124] 4. Alert Generation and Notification

[1125] If the generative model means assesses the likelihood of bank card fraud highly or if the emotion engine detects a suspicious change in emotion, the terminal generates an alert.

[1126] The alert will include detailed information about the fraud and location information and will be sent to the server via the notification method.

[1127] The server receives the alert and sends notifications to registered family members and police.

[1128] This allows the system to combine voice analysis and emotion recognition to detect cash card fraud in real time and ensure the safety of the elderly more precisely.

[1129] The processing flow will be explained below.

[1130] Processing flow of a system that combines emotion engines

[1131] Step 1: Initialization

[1132] Device:

[1133] When the device is powered on, it loads the generative model means and emotion engine and enables the voice collection means.

[1134] The device establishes an Internet connection and begins communicating with the server.

[1135] server:

[1136] The server accepts the connection request from the terminal and checks the authentication information.

[1137] User:

[1138] Users install a dedicated app and enter the necessary information (personal information, emergency contact information).

[1139] Step 2: Continuously collect audio data

[1140] Device:

[1141] The device constantly collects surrounding sounds through a microphone and processes them digitally in real time.

[1142] Step 3: Analyzing the audio data

[1143] Device:

[1144] The device feeds the collected voice data into a generative AI model for real-time analysis.

[1145] The generative modeling means detects specific phrases and patterns in the audio data that are indicative of fraud.

[1146] Step 4: Recognizing Emotional Data

[1147] Device:

[1148] At the same time, the device analyzes the voice data using an emotion engine to recognize the user's emotional state, which determines emotions based on the tone, speed, and patterns of the voice.

[1149] Step 5: Overall Scam Assessment

[1150] Device:

[1151] Based on the analysis results from the generative model and the recognition results from the emotion engine, the possibility of fraud is comprehensively evaluated.

[1152] The user's emotion data recognized by the emotion engine is supplied to the generative model means and reflected in the fraud probability assessment.

[1153] Step 6: Alert Generation

[1154] Device:

[1155] If the generative model means assesses the likelihood of bank card fraud highly or if the emotion engine detects a suspicious change in emotion, the terminal generates an alert.

[1156] The alert will include detailed information about the scam and its location.

[1157] Step 7: Sending an alert

[1158] Device:

[1159] The generated alert is sent to the server via a notification means.

[1160] server:

[1161] The server receives the alerts and records them in a database.

[1162] The server will then send notifications to registered family members and the police.

[1163] Step 8: Receive and respond to notifications

[1164] User (family / police):

[1165] Family members and police will receive alerts via a dedicated app, SMS or email.

[1166] Family members and police will check the alert and take necessary measures.

[1167] Step 9: Handling normal everyday conversations

[1168] Device:

[1169] The device also collects voice data from everyday conversations and analyzes it using a generative modeling method and emotion engine.

[1170] If no dangerous phrases are detected and the user's sentiment according to the sentiment engine is stable, the data is not processed as an alert but is stored as a log.

[1171] server:

[1172] The server periodically collects the stored log data and uses it, along with the analysis results, to help maintain a safe environment.

[1173] Step 10: Update the generative model

[1174] server:

[1175] The server periodically updates the generative AI model to accommodate new fraud patterns.

[1176] Device:

[1177] The device downloads and applies the updated generative AI model from the server.

[1178] Through these steps, this system combines emotion analysis and voice analysis to detect cash card fraud in real time and ensure the safety of the elderly with greater accuracy.

[1179] Example 2

[1180] Next, a description will be given of Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal."

[1181] Cash card fraud is a crime that particularly targets the elderly, and the damage it causes is serious. In addition, fraud methods are becoming more sophisticated, so traditional simple crime prevention systems are no longer sufficient to deal with it. Current technology does not have a system in place to detect signs of fraud in real time and prevent damage before it occurs.

[1182] The specific processing by the specific processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means.

[1183] In this invention, the server includes: a voice collection means for constantly collecting voice data; a generative model means for analyzing the collected voice data and evaluating the possibility of cash card fraud in real time; an emotion engine for analyzing the voice data to recognize emotions; an alert generation means for reevaluating the possibility of cash card fraud based on data from the generative model means and the emotion engine and issuing an alert if it is determined that fraud is highly likely; and a notification means for notifying family members and the police of the issued alert. This makes it possible to detect cash card fraud in real time by combining analysis of voice data and emotion recognition, and to prevent damage before it occurs by taking prompt action.

[1184] The "voice collection means" is hardware or software for constantly collecting voices around the user.

[1185] The "generative model means" is a system that includes a machine learning algorithm for analyzing collected voice data and assessing the likelihood of cash card fraud in real time.

[1186] The "Emotion Engine" is a software module that analyzes voice data to recognize the user's emotions. It determines the emotional state from the tone, rate, and patterns of the voice.

[1187] The "alert generation means" is a system that has the function of reevaluating the possibility of cash card fraud based on data from the generative model means and the emotion engine, and issuing an alert if it is determined that there is a high possibility of fraud.

[1188] "Notification methods" refer to the technology used to notify family members and the police of the alert. Notifications can be sent via a dedicated app, SMS, email, or other means.

[1189] "Location information" refers to geographical data of the point of origin, and this information is included in the alert.

[1190] "Fraud details" refers to specific information about cash card fraud that is included in the alert, such as the time the fraud occurred and the details of the fraud.

[1191] The "database of past cash card fraud cases" is a database that accumulates cases of cash card fraud that have occurred in the past.

[1192] The "user emotion database" is a database that stores past data on user emotions. It is used to operate the emotion engine.

[1193] The system of the present invention is an advanced security system for detecting and preventing cash card fraud in real time, and combines a voice collection means, a generative model means, an alert generation means, a notification means, and an emotion engine that recognizes user emotions. Specific embodiments for implementing the present invention are described in detail below.

[1194] System configuration

[1195] 1. Audio collection method

[1196] The device uses a built-in microphone to continuously collect sounds around the user, and this audio data is processed digitally in real time. For example, this is the case with microphones found on devices such as PCs, smartphones, and tablets.

[1197] 2. Generative Modeling Methods

[1198] The device incorporates a generative AI model (e.g., GPT-4) to analyze the collected voice data. This generative AI model uses machine learning algorithms to detect specific phrases and patterns in conversations, including risky phrases like "cash card" and "PIN number."

[1199] 3. Emotion Engine

[1200] The device is equipped with an emotion engine that recognizes the user's emotions from collected voice data. The emotion engine determines the user's emotional state from the tone, speed, and patterns of the voice, detecting changes such as anxiety or tension.

[1201] 4. Alert Generation Methods

[1202] If the device determines that fraud is likely based on information from the generative model means and emotion engine, it generates an alert. This alert includes detailed information about the fraud and location information. The alert generation means operates when a relevant dangerous phrase or suspicious emotion change is detected.

[1203] 5. Means of notification

[1204] The generated alert is sent to the server via a notification method. The server receives the alert and notifies registered family members and the police. Notifications are sent via a dedicated app, SMS, email, etc. Location information and detailed information about the fraud are included, urging immediate action.

[1205] Specific examples

[1206] 1. If you receive a fraudulent phone call

[1207] The elderly user receives a call from an unknown number. The caller claims to be a bank employee and asks for their cash card and PIN number.

[1208] The device collects this conversation using a voice capture method and analyzes it with a generative AI model, which detects dangerous phrases such as "cash card" and "PIN number."

[1209] At the same time, the emotion engine recognizes changes in emotions from the tone, rate, and patterns of the user's voice, detecting anxiety and tension.

[1210] The generative model determines if there is a high risk of fraud and generates an alert, which includes details of the fraud and location information.

[1211] The alert is sent to a server via the notification method, and the server then sends a real-time notification to family members or the police via a dedicated app, SMS, email, etc., allowing for immediate action.

[1212] Prompt Sentence Examples

[1213] By using prompt sentences like the one below for the generative AI model, you can verify and adjust the system's operation.

[1214] "Please give an example of how you would respond if you were asked to verify your bank card over the phone."

[1215] "How does the emotion engine recognize changes in emotion from the tone and rate of a user's voice?"

[1216] "What information is included in the generated alert?"

[1217] This system combines analysis of user voice data with emotion recognition to detect cash card fraud in real time and promote quick and appropriate responses.

[1218] The flow of the identification process in the second embodiment will be described with reference to FIG.

[1219] Step 1:

[1220] When the device starts up, it loads the generative modeling means and emotion engine, activates the voice collection means, establishes an internet connection, and starts communication with the server, allowing the system to be fully functional.

[1221] Input: Power on

[1222] Output: System ready state

[1223] Step 2:

[1224] The device constantly collects the user's surrounding sounds through a built-in microphone. This audio data is processed digitally in real time and input to the generative modeling means and emotion engine.

[1225] Input: Audio around the user

[1226] Output: Digital audio data

[1227] Step 3:

[1228] The device's generative modeling means analyzes the collected voice data to detect specific phrases and patterns. A generative AI model (e.g., GPT-4) detects risky phrases such as "cash card" and "PIN number." This analysis assesses the likelihood of fraud.

[1229] Input: Digital audio data

[1230] Output: Detection results for specific phrases and fraud likelihood rating

[1231] Step 4:

[1232] The emotion engine analyzes the user's emotions from the voice data, determining the user's emotional state from the tone, rate, and patterns of the voice, and detecting emotional changes such as anxiety or tension.

[1233] Input: Digital audio data

[1234] Output: Evaluation result of the user's emotional state

[1235] Step 5:

[1236] The generative modeling means considers the emotion data from the emotion engine and reassess the risk of fraud. If it determines that fraud is likely, the alert generation means operates to generate an alert that includes detailed information and location information.

[1237] Input: Phrase detection results and emotional state evaluation results

[1238] Output: Alert generation decision and alert data

[1239] Step 6:

[1240] The generated alert is sent to the server via the notification method. The server receives the alert and notifies registered family members and the police. Notifications are sent via a dedicated app, SMS, email, etc.

[1241] Input: Alert data

[1242] Output: Notify family and police

[1243] Step 7:

[1244] The server evaluates responses from family members and police who receive the notification and, in some cases, provides further details or escalates the alert.

[1245] Input: Responses from family and police

[1246] Output: Additional details and escalation instructions

[1247] Through these processing steps, the system combines voice data analysis and emotion recognition to detect cash card fraud in real time and promptly respond.

[1248] (Application example 2)

[1249] Next, a description will be given of Application Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal."

[1250] In recent years, there has been an increase in cash card frauds targeting the elderly, and effective measures to prevent such damage are needed. Current security systems have difficulty detecting signs of fraud in real time, and are particularly ineffective against telephone fraud. Furthermore, because it is difficult for elderly people to recognize fraud themselves, more advanced fraud detection systems must be developed. Furthermore, fraud detection systems must be able to recognize users' emotions and take their changes into account for more accurate detection. This requires immediate and appropriate alerts in situations where there is a high possibility of fraud, and prompt notification to family members and the police.

[1251] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 2 is realized by the following means.

[1252] In this invention, the server includes voice collection means for constantly collecting voice data, generative model means for analyzing the collected voice data and evaluating the possibility of cash card fraud in real time, alert generation means for issuing an alert when the generative model means determines that there is a high possibility of cash card fraud, notification means for notifying family members and the police of the issued alert, an emotion engine for recognizing the user's emotions from the voice data, and means for reevaluating the possibility of fraud when the emotion engine determines that a change in emotion is suspicious. This enables highly accurate fraud detection that takes into account changes in the user's emotions and rapid issuance of alerts and notifications.

[1253] The "voice collection means" is a device or system for constantly collecting voices around the user.

[1254] A "generative model means" is a device or system that incorporates a machine learning algorithm for analyzing collected voice data and assessing the likelihood of cash card fraud in real time.

[1255] The "alert generation means" is a device or system that issues an alert when the generative model means determines that there is a high possibility of cash card fraud.

[1256] "Notification means" refers to a communication device or system for notifying family members or the police of an alert that has been issued.

[1257] An "emotion engine" is a device or algorithm that recognizes a user's emotions from voice data and determines changes in emotions.

[1258] A "means for reassessing" is a device or system that reassess the possibility of fraud when the emotion engine determines that a change in emotion is suspicious.

[1259] The system for implementing this invention is broadly composed of the following elements: a voice collection means, a generative model means, an emotion engine, an alert generation means, and a notification means. These elements are realized by combining multiple pieces of software and hardware.

[1260] 1. Audio collection method

[1261] The voice collection means uses a microphone built into a device such as a smartphone. It constantly collects voices around the user and converts them into data in real time. This voice data is then input into the generative model means and emotion engine in the next step.

[1262] 2. Generative Modeling Methods

[1263] The generative model means includes a machine learning algorithm for analyzing the collected voice data. Specifically, the generative AI model is used to detect specific phrases and patterns in the voice data that indicate possible bank card fraud. The model works in conjunction with a pre-trained database (past bank card fraud cases).

[1264] 3. Emotion Engine

[1265] The emotion engine includes algorithms for recognizing a user's emotional state from speech data. The emotion engine analyzes the tone, rate, and patterns of speech to determine changes in the user's emotions. This information is then used to assess the likelihood of fraud by means of a generative model.

[1266] 4. Alert Generation Methods

[1267] The alert generator operates based on information from the generative model and the emotion engine. If a fraud is deemed likely or a suspicious change in emotion is detected, an alert is generated. The alert includes detailed information about the fraud and data about the user's current location.

[1268] 5. Means of notification

[1269] The notification method sends the generated alert to a server, which processes the received alert and sends notifications to registered family members and the police via a dedicated application, SMS, email, etc.

[1270] Hardware and software used

[1271] Hardware: Smartphone (built-in microphone)

[1272] Software: Audio collection library (e.g., some_audio_library), generative AI model library (e.g., some_ml_library), emotion recognition library (e.g., emotion_recognition_library)

[1273] Cloud server: Alert notifications and log data storage

[1274] Adding specific examples

[1275] For example, consider a scenario in which an elderly person receives a scam call from an unknown number. If the generative AI model and emotion recognition engine detect a suspicious phrase or emotional change, the following prompt will be generated:

[1276] Scam Guardian Alert: Call from unknown number may be a bank card scam. Details: Scam details, Location: Current location

[1277] This prompt will alert you and your family immediately so you can take any necessary action.

[1278] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[1279] Step 1:

[1280] Start audio collection

[1281] When the device is started, it activates the built-in microphone and starts the audio collection method. It continuously collects audio data around the user and converts it into a digital format in real time. The input is the surrounding audio, and the output is the digitized audio data.

[1282] Step 2:

[1283] Analysis of audio data

[1284] The device inputs the collected voice data into a generative AI model means. The generative AI model uses a machine learning algorithm to analyze the voice data and detect specific phrases or patterns that may indicate possible bank card fraud. The input is the digitized voice data, and the output is the results of the analyzed phrases and patterns.

[1285] Step 3:

[1286] Emotion recognition

[1287] At the same time, the device also inputs the voice data into the emotion engine, which analyzes the tone, speed, and pattern of the voice to recognize the user's emotional state. The input is digitized voice data, and the output is the result of the user's emotional state assessment.

[1288] Step 4:

[1289] Fraud likelihood assessment

[1290] The terminal integrates information from the generative model means and the emotion engine. The generative model means reevaluates the emotion data provided by the emotion engine and determines the likelihood of fraud. The input is the analyzed phrases, patterns, and emotion data, and the output is the result of the fraud likelihood assessment.

[1291] Step 5:

[1292] Generate alerts

[1293] If a fraud possibility is determined, the device immediately generates an alert using the alert generation means. This alert includes detailed fraud information and location information. The input is the fraud possibility assessment result, and the output is the generated alert.

[1294] Step 6:

[1295] Alert Notification

[1296] The generated alert is sent to the server via the notification means. The server processes the received alert and sends a notification to registered family members or police. Notifications are sent via a dedicated app, SMS, email, etc. The input is the generated alert, and the output is the notification to family members or police.

[1297] Step 7:

[1298] Facilitating user and family support

[1299] The family member can check the received notification and immediately contact the user or arrange for on-site visit if necessary. The input is the received alert notification, and the output is the appropriate response action.

[1300] As a result of this, the system will be able to combine voice data and emotional data to detect cash card fraud with high accuracy and send notifications to users, their families, and the police in real time.

[1301] The specific processing unit 290 transmits the result of the specific processing to the headset type terminal 314. In the headset type terminal 314, the control unit 46A causes the speaker 240 and the display 343 to output the result of the specific processing. The microphone 238 acquires audio indicating a user input regarding the result of the specific processing. The control unit 46A transmits audio data indicating the user input acquired by the microphone 238 to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the audio data.

[1302] The data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of the data generation model 58 is ChatGPT (Internet Search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search <url: https: gemini.google.com ?hl="ja">) and other generation AIs. The data generation model 58 is obtained by performing deep learning on a neural network. A prompt including an instruction is input to the data generation model 58, and inference data such as voice data indicating voice, text data indicating text, and image data indicating an image is also input. The data generation model 58 performs inference on the input inference data in accordance with the instruction indicated by the prompt, and outputs the inference result in a data format such as voice data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[1303] In the above embodiment, an example was given in which the specific processing is performed by the data processing device 12, but the technology of the present disclosure is not limited to this, and the specific processing may be performed by the headset type terminal 314.

[1304] [Fourth embodiment]

[1305] FIG. 7 shows an example of the configuration of a data processing system 410 according to the fourth embodiment.

[1306] 7, a data processing system 410 includes a data processing device 12 and a robot 414. An example of the data processing device 12 is a server.

[1307] The data processing device 12 includes a computer 22, a database 24, and a communication I / F 26. The computer 22 is an example of a "computer" according to the technology of the present disclosure. The computer 22 includes a processor 28, a RAM 30, and a storage 32. The processor 28, the RAM 30, and the storage 32 are connected to a bus 34. The database 24 and the communication I / F 26 are also connected to the bus 34. The communication I / F 26 is connected to a network 54. Examples of the network 54 include a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[1308] The robot 414 includes a computer 36, a microphone 238, a speaker 240, a camera 42, a communication I / F 44, and a control target 443. The computer 36 includes a processor 46, a RAM 48, and a storage 50. The processor 46, the RAM 48, and the storage 50 are connected to a bus 52. The microphone 238, the speaker 240, the camera 42, and the control target 443 are also connected to the bus 52.

[1309] The microphone 238 receives instructions and the like from the user 20 by receiving voice uttered by the user 20. The microphone 238 captures the voice uttered by the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio in accordance with instructions from the processor 46.

[1310] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an imaging element such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the surroundings of user 20 (for example, an imaging range defined by an angle of view equivalent to the field of vision of a typical healthy person).

[1311] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 are responsible for the exchange of various information between the processor 46 and the processor 28 via the network 54. The exchange of various information between the processor 46 and the processor 28 using the communication I / Fs 44 and 26 is carried out in a secure state.

[1312] The control object 443 includes a display device, LEDs in the eyes, and motors for driving the arms, hands, and feet. The posture and gestures of the robot 414 are controlled by controlling the motors of the arms, hands, and feet. Some of the emotions of the robot 414 can be expressed by controlling these motors. In addition, the facial expressions of the robot 414 can also be expressed by controlling the light emission state of the LEDs in the eyes of the robot 414.

[1313] Fig. 8 shows an example of the main functions of the data processing device 12 and the robot 414. As shown in Fig. 8, in the data processing device 12, a specific process is performed by the processor 28. A specific process program 56 is stored in the storage 32.

[1314] The specific processing program 56 is an example of a "program" according to the technology of the present disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific processing is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific processing program 56 executed on the RAM 30.

[1315] The storage 32 stores a data generation model 58 and an emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[1316] In the robot 414, the processor 46 performs the reception output process. A reception output program 60 is stored in the storage 50. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output process is realized by the processor 46 operating as the control unit 46A in accordance with the reception output program 60 executed on the RAM 48.

[1317] Next, a description will be given of the specific processing performed by the specific processing unit 290 of the data processing device 12. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal."

[1318] The system of the present invention is an advanced crime prevention system for detecting and preventing cash card fraud in real time. The detailed operation of the system and specific methods of use are described below.

[1319] System configuration

[1320] 1. Audio collection method

[1321] The device has a built-in microphone that constantly collects sounds around the user, and the collected sound data is processed digitally in real time.

[1322] 2. Generative Modeling Methods

[1323] To analyze the collected voice data, the device incorporates a generative AI model that uses pre-trained machine learning algorithms to detect specific phrases and patterns in speech.

[1324] 3. Alert Generation Methods

[1325] If the generative AI model determines that there is a high probability of bank card fraud, the device will immediately generate an alert, which will include specific details of the suspected fraudulent conversation and its location.

[1326] 4. Means of notification

[1327] The alert generated on the device is sent to a server, which then notifies registered family members and the police via a dedicated app, SMS, or email.

[1328] Specific examples

[1329] 1. If you receive a fraudulent phone call

[1330] The elderly user receives a call from an unknown number. The caller claims to be a bank employee and asks for confirmation of their cash card number and PIN.

[1331] The device collects this conversation using a voice capture means and analyzes it using a generative model means. The generative AI model detects dangerous phrases such as "cash card" and "PIN number."

[1332] The generative model means determines that there is a high suspicion of fraud and activates the alert generation means, which immediately generates an alert containing detailed information about the fraud and location information.

[1333] The generated alert is sent to a server by a notification means, and the server sends notifications to registered family members and the police in real time.

[1334] Family members and police receive alerts through a dedicated app and can immediately contact the elderly person or go directly to the scene.

[1335] 2. Normal everyday conversation

[1336] When a user has a daily conversation with a family member over the phone, the terminal similarly collects voice data and analyzes it using the generative model means.

[1337] If the generative AI model does not detect any particularly dangerous phrases, the audio data is securely stored in the system as a log and no alert is generated.

[1338] The server periodically collects log data and uses the analysis results to strengthen security.

[1339] Program processing

[1340] To make it easier to understand, the specific flow of the system's program processing is shown below:

[1341] 1. Initialize the device

[1342] When the device boots up, it loads the generative model means, activates the audio collection means, establishes an internet connection, and synchronizes with the server.

[1343] 2. Audio collection and analysis

[1344] The device constantly collects audio from the user's surroundings and feeds it into a generative AI model in real time, which then analyzes the audio data to detect phrases that could be fraudulent.

[1345] 3. Alert generation and notification

[1346] If the generative model means highly evaluates the possibility of cash card fraud, the alert generation means immediately generates an alert and transmits it to the server.

[1347] The server receives the alert and sends notifications to registered family members and police.

[1348] This will prevent fraud and ensure the safety of the elderly.

[1349] The processing flow will be explained below.

[1350] Program processing flow

[1351] Step 1: Initialization

[1352] Device:

[1353] When the terminal is powered on, it loads the generative model means and enables the audio collection means.

[1354] The device establishes an Internet connection and begins communicating with the server.

[1355] server:

[1356] The server accepts the connection request from the terminal and checks the authentication information.

[1357] User:

[1358] Users install a dedicated app and enter the necessary information (personal information, emergency contact information).

[1359] Step 2: Continuously collect audio data

[1360] Device:

[1361] The device constantly collects surrounding sounds through a microphone, and this collected audio data is processed digitally in real time.

[1362] Step 3: Analyzing the audio data

[1363] Device:

[1364] The device feeds the collected voice data into a generative AI model for real-time analysis.

[1365] The generative modeling means detects specific phrases and patterns in the audio data that are indicative of fraud.

[1366] Step 4: Generate an alert if fraud is suspected

[1367] Device:

[1368] If the generative AI model determines that fraud is highly suspected, the device will generate an alert.

[1369] The alert will include detailed information about the scam and its location.

[1370] Step 5: Sending an alert

[1371] Device:

[1372] The generated alert is sent to the server via a notification means.

[1373] server:

[1374] The server receives the alerts and records them in a database.

[1375] The server will then send notifications to registered family members and the police.

[1376] Step 6: Receive and respond to notifications

[1377] User (family / police):

[1378] Family members and police will receive alerts via a dedicated app, SMS or email.

[1379] Family members and police will check the alert and take necessary measures.

[1380] Step 7: Handling normal everyday conversations

[1381] Device:

[1382] The device also collects speech data from everyday conversations and analyzes it using generative modeling techniques.

[1383] If no dangerous phrases are detected, the data is not processed as an alert but is saved as a log.

[1384] server:

[1385] The server periodically collects the stored log data and uses it, along with the analysis results, to help maintain a safe environment.

[1386] Step 8: Update the generative model

[1387] server:

[1388] The server periodically updates the generative AI model to accommodate new fraud patterns.

[1389] Device:

[1390] The device downloads and applies the updated generative AI model from the server.

[1391] This allows the system to detect cash card fraud in real time and ensure the safety of the elderly.

[1392] Example 1

[1393] Next, a description will be given of Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal."

[1394] In modern society, cash card fraud targeting the elderly is on the rise. This not only causes significant financial losses for these individuals, but also psychological damage. However, current security systems have difficulty detecting fraudulent activity in real time and taking immediate appropriate action. Therefore, a more advanced and rapid system for detecting and notifying fraud is needed.

[1395] The specific processing by the specific processing unit 290 of the data processing device 12 in the first embodiment is realized by the following means.

[1396] In this invention, the server includes an audio recording means for constantly collecting voice data, a generative model means for analyzing the collected voice data and evaluating the possibility of fraud in real time, a warning generation means for issuing an alert when the generative model means determines that there is a high possibility of fraud, and an information transmission means for notifying family members and the police of the issued alert. This makes it possible to quickly and accurately detect fraudulent acts targeting the elderly and to immediately notify family members and the police.

[1397] The "sound recording means" is a device for constantly collecting sound data around the user.

[1398] The "generative model means" is a device or software that analyzes collected voice data and evaluates the likelihood of fraud in real time in conjunction with a database of past fraud cases.

[1399] The "alert generation means" is a device for issuing an alert when the generative model means determines that there is a high possibility of fraud.

[1400] "Information transmission means" refers to a communication device or system for notifying family members or police of an alert that has been issued.

[1401] The "fraud case database" is a storage device that accumulates data on fraud cases that have occurred in the past, and that the generative model means refers to and analyzes this data.

[1402] An "alert" is warning information generated when the generative model means determines that there is a high possibility of fraud, and includes detailed information about the fraud and location information.

[1403] "Location information" is data indicating the geographic location of the point of origin, and is obtained using technology such as GPS.

[1404] The present invention is a system for detecting fraudulent activities targeting elderly people in real time and responding promptly. The system consists of an acoustic recording means, a generative model means, a warning generation means, and an information transmission means.

[1405] System Components

[1406] 1. Acoustic Recording Means

[1407] The device has a built-in microphone that constantly collects sounds around the user. This collected sound data is stored in a digital format. For example, the built-in high-sensitivity microphone records sound in WAV format.

[1408] 2. Generative Modeling Methods

[1409] The device is equipped with a generative AI model to analyze the collected voice data. This generative AI model converts the voice data into text using natural language processing (NLP) techniques and compares it with a database of past fraud cases. Specific examples include machine learning frameworks such as TensorFlow and PyTorch.

[1410] 3. Warning generation means

[1411] If the generative modeling method assesses the likelihood of fraud, the device immediately generates an alert that includes the text of the suspected fraudulent conversation, the time of occurrence, and location information (e.g., GPS data).

[1412] 4. Means of communication

[1413] The alert generated on the device is sent to a server, which then notifies registered family members and the police via a dedicated app, SMS, email, or other means.

[1414] Specific examples

[1415] Example 1: If you receive a fraudulent phone call

[1416] 1. Situation

[1417] The elderly user receives a call from an unknown number. The caller claims to be a bank employee and asks for confirmation of their cash card number and PIN.

[1418] 2. Operation

[1419] The device collects this conversation using a voice capture means and analyzes it using a generative model means. The generative AI model detects dangerous phrases such as "cash card" and "PIN number."

[1420] The generative model means determines that there is a high suspicion of fraud and activates the alert generation means, which generates and sends out an alert.

[1421] The alert is sent to a server via a communication device, and the server then notifies registered family members or police. Family members or police receive the alert through a dedicated app and can immediately contact the elderly person or call the police.

[1422] Example 2: Normal everyday conversation

[1423] 1. Situation

[1424] The user has everyday conversations with his family.

[1425] 2. Operation

[1426] The device also collects voice data and analyzes it using a generative model. If the generative AI model does not detect any particularly dangerous phrases, the voice data is safely stored in the system as a log and no alert is generated.

[1427] The server periodically collects log data and uses it, along with the analysis results, to improve the accuracy of the system.

[1428] Prompt Sentence Examples

[1429] "If families want to protect their elderly relatives from fraud, please tell us about the development of a system that can detect and notify them of fraudulent phone calls."

[1430] "Give us an example of how we could design a system that leverages generative AI models to prevent bank card fraud."

[1431] The system configuration and operating principles described above make it possible to quickly detect and immediately deal with fraudulent acts targeting the elderly, making this system particularly effective in ensuring the safety of the elderly.

[1432] The flow of the identification process in the first embodiment will be described with reference to FIG.

[1433] System program processing flow

[1434] Step 1:

[1435] Initializing the device

[1436] Input: The device is powered on.

[1437] Processing: The device first loads the generative AI model into memory, using libraries such as TensorFlow or PyTorch.

[1438] Specific operation: The trained model file of the generative AI model is loaded and the model is initialized. Next, the microphone, which is the audio collection means, is enabled and set up so that audio data can be collected in digital format. Next, the system connects to the Internet using Wi-Fi or wired communication and synchronizes with the server.

[1439] Output: An initialized AI model, ready to collect audio.

[1440] Step 2:

[1441] Audio data collection

[1442] Input: Audio around the user.

[1443] Processing: The device continuously collects surrounding audio and stores it in digital format (e.g., WAV format).

[1444] Specific operation: The audio captured by the microphone undergoes pre-processing such as noise removal using digital signal processing (DSP), and is then input into the generative AI model.

[1445] Output: Preprocessed digital audio data.

[1446] Step 3:

[1447] Analysis of audio data

[1448] Input: Digital audio data collected in step 2.

[1449] Processing: The device feeds collected voice data into a generative AI model to detect phrases and patterns.

[1450] How it works: The generative AI model converts digital voice data into text and then uses natural language processing (NLP) techniques to detect potentially fraudulent phrases from a list.

[1451] Output: A fraud likelihood assessment score.

[1452] Step 4:

[1453] Generate alerts

[1454] Input: Analysis results from step 3 (likely fraudulent reputation score).

[1455] Processing: If the generative AI model determines that there is a high possibility of fraud, the alert generation means operates and generates an alert.

[1456] What it does: The alert will include the text of the allegedly fraudulent conversation, the time it occurred, and location information (GPS data). This information will be compiled into a single message.

[1457] Output: The generated alert message.

[1458] Step 5:

[1459] Alert Notification

[1460] Input: The alert message generated in step 4.

[1461] Processing: The terminal sends the generated alert message to the server.

[1462] How it works: Communication is via HTTP / HTTPS protocol and is encrypted for data security. The server analyzes the received alert and notifies registered family members and police.

[1463] Output: Notification message sent to family and police.

[1464] The above is an explanation of the specific operations, inputs, and outputs at each processing step. This system will enable us to quickly and effectively detect fraud targeting the elderly and respond immediately.

[1465] (Application example 1)

[1466] Next, a description will be given of Application Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal."

[1467] The risk of elderly people and general users becoming victims of cash card fraud and other fraudulent activities is increasing. In particular, fraudulent activities over the phone are becoming more sophisticated, and many users are becoming victims without realizing it. Therefore, there is a need for an effective system that can detect fraudulent activities in real time and immediately notify relevant parties and crime prevention agencies.

[1468] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 1 is realized by the following means.

[1469] In this invention, the server includes an audio collection means for constantly collecting audio data, an AI model generation means for analyzing the collected audio data and evaluating the possibility of fraudulent activity in real time, an alert generation means for issuing an alert when the AI ​​model generation means determines that there is a high possibility of fraudulent activity, and a notification means for notifying relevant parties and crime prevention agencies of the issued alert, thereby enabling fraudulent activity to be detected in real time and dealt with promptly.

[1470] "Audio collection means" refers to devices or technologies for constantly collecting surrounding audio data.

[1471] "Generative AI model means" refers to technology that uses a generative AI model to analyze collected voice data and assess the possibility of fraudulent activity in real time.

[1472] "Alert generation means" refers to a device or technology for issuing an alert when the generation AI model means determines that there is a high possibility of fraudulent activity.

[1473] "Notification means" refers to devices or technologies for notifying relevant parties and crime prevention agencies of the issued alert.

[1474] "Fraudulent acts" are acts such as cash card fraud that attempt to deceive users and illegally obtain money or personal information.

[1475] The "database of past misconduct cases" is a database that stores information about misconduct that has occurred in the past.

[1476] This invention aims to realize a security system for detecting and preventing fraudulent activities. The system combines a voice collection means, a generative AI model means, an alert generation means, and a notification means. Here, a specific example of this system will be described.

[1477] System Components

[1478] 1. Audio collection method

[1479] The device is equipped with a built-in microphone that constantly collects sounds around the user. This microphone has high sensitivity and noise-canceling capabilities, allowing it to collect clear audio data.

[1480] 2. Generative AI Model Means

[1481] The collected voice data is sent to an on-device generative AI modeling facility, which uses a pre-trained natural language processing model (e.g., GPT-4) to analyze the voice data and detect specific phrases and patterns associated with fraudulent activity. The model works in conjunction with a database of past fraud cases to improve detection accuracy.

[1482] 3. Alert Generation Methods

[1483] If the generation AI model means determines that there is a high possibility of fraudulent activity, the alert generation means immediately issues an alert, which includes the user's location information and details of the dangerous phrase, allowing for a prompt response.

[1484] 4. Means of notification

[1485] The alert is sent from the device to a server, which then notifies relevant parties and crime prevention agencies via SMS, email, a dedicated app, etc.

[1486] Example

[1487] When an elderly person receives a fraudulent phone call

[1488] When an elderly person hears phrases such as "cash card," "PIN number," or "bank employee" over the phone, the device's microphone collects the audio and sends it to a generative AI model. The model analyzes these phrases and, if it determines there is a high possibility of fraud, the alert generation means immediately sends out an alert. This alert also includes location information, allowing relevant parties and crime prevention agencies to respond quickly.

[1489] Use in everyday conversation

[1490] Similarly, when a user has a daily conversation with a family member on the phone, the device collects the audio and sends it to the generative AI model. If the model does not detect any particularly dangerous phrases, the audio data is securely stored as a log within the system. This allows the system to prevent false alarms and provide accurate information when needed.

[1491] Prompt Sentence Examples

[1492] Analyze audio data of conversations containing phrases such as "cash card," "PIN number," "bank employee," and "provided" to assess the likelihood of fraudulent activity.

[1493] In this way, the present invention provides a system that reduces the risk of users becoming involved in fraud or other fraudulent activity and allows for a quick and accurate response.

[1494] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[1495] Step 1:

[1496] Initialization:

[1497] When the device boots up, it first loads the generative AI model, activates the audio capture method, prepares a microphone with high sensitivity and noise cancellation, establishes an internet connection, and synchronizes with the server.

[1498] Input: Start terminal

[1499] Output: Loading the generative AI model, enabling the audio capture method, and establishing an internet connection

[1500] Step 2:

[1501] Audio Collection:

[1502] The microphone constantly collects sounds around the user, and the audio data is converted into a digital format in real time and stored on the device.

[1503] Input: Ambient audio

[1504] Output: Digital audio data

[1505] Step 3:

[1506] Audio Analysis:

[1507] The collected voice data is sent to a generative AI model, which analyzes the speech based on prompts to detect specific phrases and patterns associated with fraudulent activity.

[1508] Input: Digital audio data

[1509] Output: The result of assessing the likelihood of fraud.

[1510] Step 4:

[1511] Alert generation:

[1512] If the generative AI model determines that fraud is likely, the alert generator will immediately send an alert, which will include the user's location and details of the dangerous phrase.

[1513] Input: Results of evaluation of potential fraud

[1514] Output: Alert information (location information, dangerous phrases)

[1515] Step 5:

[1516] notification:

[1517] The alert is sent from the device to a server, which then notifies relevant parties and crime prevention agencies via SMS, email, a dedicated app, etc.

[1518] Input: Alert information (location information, dangerous phrases)

[1519] Output: Notification to relevant parties and security agencies

[1520] Furthermore, an emotion engine that estimates the user's emotion may be further combined. That is, the identification processing unit 290 may estimate the user's emotion using the emotion identification model 59, and perform identification processing using the user's emotion.

[1521] The system of the present invention is an advanced crime prevention system for detecting and preventing cash card fraud in real time, and combines a voice collection means, a generative model means, an alert generation means, a notification means, and an emotion engine that recognizes the user's emotions. The detailed operation of the system and specific usage methods are described below.

[1522] System configuration

[1523] 1. Audio collection method

[1524] The device has a built-in microphone that constantly collects sounds around the user, and this sound data is processed digitally in real time.

[1525] 2. Generative Modeling Methods

[1526] To analyze the collected voice data, the device incorporates a generative AI model that uses machine learning algorithms to detect specific phrases and patterns in conversation.

[1527] 3. Emotion Engine

[1528] The device is equipped with an emotion engine that recognizes the user's emotions from collected voice data, and determines the user's emotional state from the tone, speed, and patterns of the voice.

[1529] 4. Alert Generation Methods

[1530] If the generative model and emotion engine determine that fraud is likely, the device generates an alert, which includes details about the fraud and its location.

[1531] 5. Means of notification

[1532] The generated alert is sent to the server via the notification method. The server receives the alert and notifies registered family members and the police. Notifications are sent via a dedicated app, SMS, or email.

[1533] Specific examples

[1534] 1. If you receive a fraudulent phone call

[1535] The elderly user receives a call from an unknown number. The caller claims to be a bank employee and asks for confirmation of their cash card number and PIN.

[1536] The device collects this conversation using a voice capture means and analyzes it using a generative model means. The generative AI model detects dangerous phrases such as "cash card" and "PIN number."

[1537] At the same time, the emotion engine recognizes changes in emotion from the tone and rate of the user's voice, and if a suspected change in emotion is detected, it feeds this data to the generative model means.

[1538] The generative modeling means takes the sentiment data into account to reassess the likelihood of fraud, and if suspicion of fraud increases, it activates the alert generation means, which immediately generates an alert.

[1539] The generated alert is sent to a server by a notification means, and the server sends notifications to registered family members and the police in real time.

[1540] Family members and police receive alerts through a dedicated app and can immediately contact the elderly person or go directly to the scene.

[1541] 2. Normal everyday conversation

[1542] When a user has a daily conversation with a family member over the phone, the terminal similarly collects voice data and analyzes it using the generative model means.

[1543] If no dangerous phrases are detected, and the results of the emotion engine's analysis of the user's emotions are normal, the data will not be processed as an alert, but will be saved as a log.

[1544] The server periodically collects the stored log data and uses it, along with the analysis results, to strengthen the security of the system.

[1545] Program processing

[1546] To make it easier to understand, the specific flow of the system's program processing is shown below:

[1547] 1. Initialize the device

[1548] When the device starts up, it loads the generative model means and emotion engine, activates the voice collection means, establishes an Internet connection, and starts communication with the server.

[1549] 2. Audio collection and analysis

[1550] The device constantly collects audio from the user's surroundings and feeds it into a generative AI model and emotion engine in real time. The generative model analyzes the audio data to detect specific phrases and patterns that may be indicative of fraud.

[1551] 3. Emotion recognition and evaluation

[1552] The emotion engine recognizes the user's emotions from the collected voice data and supplies changes in emotions to the generative model means.

[1553] 4. Alert Generation and Notification

[1554] If the generative model means assesses the likelihood of bank card fraud highly or if the emotion engine detects a suspicious change in emotion, the terminal generates an alert.

[1555] The alert will include detailed information about the fraud and location information and will be sent to the server via the notification method.

[1556] The server receives the alert and sends notifications to registered family members and police.

[1557] This allows the system to combine voice analysis and emotion recognition to detect cash card fraud in real time and ensure the safety of the elderly more precisely.

[1558] The processing flow will be explained below.

[1559] Processing flow of a system that combines emotion engines

[1560] Step 1: Initialization

[1561] Device:

[1562] When the device is powered on, it loads the generative model means and emotion engine and enables the voice collection means.

[1563] The device establishes an Internet connection and begins communicating with the server.

[1564] server:

[1565] The server accepts the connection request from the terminal and checks the authentication information.

[1566] User:

[1567] Users install a dedicated app and enter the necessary information (personal information, emergency contact information).

[1568] Step 2: Continuously collect audio data

[1569] Device:

[1570] The device constantly collects surrounding sounds through a microphone and processes them digitally in real time.

[1571] Step 3: Analyzing the audio data

[1572] Device:

[1573] The device feeds the collected voice data into a generative AI model for real-time analysis.

[1574] The generative modeling means detects specific phrases and patterns in the audio data that are indicative of fraud.

[1575] Step 4: Recognizing Emotional Data

[1576] Device:

[1577] At the same time, the device analyzes the voice data using an emotion engine to recognize the user's emotional state, which determines emotions based on the tone, speed, and patterns of the voice.

[1578] Step 5: Overall Scam Assessment

[1579] Device:

[1580] Based on the analysis results from the generative model and the recognition results from the emotion engine, the possibility of fraud is comprehensively evaluated.

[1581] The user's emotion data recognized by the emotion engine is supplied to the generative model means and reflected in the fraud probability assessment.

[1582] Step 6: Alert Generation

[1583] Device:

[1584] If the generative model means assesses the likelihood of bank card fraud highly or if the emotion engine detects a suspicious change in emotion, the terminal generates an alert.

[1585] The alert will include detailed information about the scam and its location.

[1586] Step 7: Sending an alert

[1587] Device:

[1588] The generated alert is sent to the server via a notification means.

[1589] server:

[1590] The server receives the alerts and records them in a database.

[1591] The server will then send notifications to registered family members and the police.

[1592] Step 8: Receive and respond to notifications

[1593] User (family / police):

[1594] Family members and police will receive alerts via a dedicated app, SMS or email.

[1595] Family members and police will check the alert and take necessary measures.

[1596] Step 9: Handling normal everyday conversations

[1597] Device:

[1598] The device also collects voice data from everyday conversations and analyzes it using a generative modeling method and emotion engine.

[1599] If no dangerous phrases are detected and the user's sentiment according to the sentiment engine is stable, the data is not processed as an alert but is stored as a log.

[1600] server:

[1601] The server periodically collects the stored log data and uses it, along with the analysis results, to help maintain a safe environment.

[1602] Step 10: Update the generative model

[1603] server:

[1604] The server periodically updates the generative AI model to accommodate new fraud patterns.

[1605] Device:

[1606] The device downloads and applies the updated generative AI model from the server.

[1607] Through these steps, this system combines emotion analysis and voice analysis to detect cash card fraud in real time and ensure the safety of the elderly with greater accuracy.

[1608] Example 2

[1609] Next, a description will be given of Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal."

[1610] Cash card fraud is a crime that particularly targets the elderly, and the damage it causes is serious. In addition, fraud methods are becoming more sophisticated, so traditional simple crime prevention systems are no longer sufficient to deal with it. Current technology does not have a system in place to detect signs of fraud in real time and prevent damage before it occurs.

[1611] The specific processing by the specific processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means.

[1612] In this invention, the server includes: a voice collection means for constantly collecting voice data; a generative model means for analyzing the collected voice data and evaluating the possibility of cash card fraud in real time; an emotion engine for analyzing the voice data to recognize emotions; an alert generation means for reevaluating the possibility of cash card fraud based on data from the generative model means and the emotion engine and issuing an alert if it is determined that fraud is highly likely; and a notification means for notifying family members and the police of the issued alert. This makes it possible to detect cash card fraud in real time by combining analysis of voice data and emotion recognition, and to prevent damage before it occurs by taking prompt action.

[1613] The "voice collection means" is hardware or software for constantly collecting voices around the user.

[1614] The "generative model means" is a system that includes a machine learning algorithm for analyzing collected voice data and assessing the likelihood of cash card fraud in real time.

[1615] The "Emotion Engine" is a software module that analyzes voice data to recognize the user's emotions. It determines the emotional state from the tone, rate, and patterns of the voice.

[1616] The "alert generation means" is a system that has the function of reevaluating the possibility of cash card fraud based on data from the generative model means and the emotion engine, and issuing an alert if it is determined that there is a high possibility of fraud.

[1617] "Notification methods" refer to the technology used to notify family members and the police of the alert. Notifications can be sent via a dedicated app, SMS, email, or other means.

[1618] "Location information" refers to geographical data of the point of origin, and this information is included in the alert.

[1619] "Fraud details" refers to specific information about cash card fraud that is included in the alert, such as the time the fraud occurred and the details of the fraud.

[1620] The "database of past cash card fraud cases" is a database that accumulates cases of cash card fraud that have occurred in the past.

[1621] The "user emotion database" is a database that stores past data on user emotions. It is used to operate the emotion engine.

[1622] The system of the present invention is an advanced security system for detecting and preventing cash card fraud in real time, and combines a voice collection means, a generative model means, an alert generation means, a notification means, and an emotion engine that recognizes user emotions. Specific embodiments for implementing the present invention are described in detail below.

[1623] System configuration

[1624] 1. Audio collection method

[1625] The device uses a built-in microphone to continuously collect sounds around the user, and this audio data is processed digitally in real time. For example, this is the case with microphones found on devices such as PCs, smartphones, and tablets.

[1626] 2. Generative Modeling Methods

[1627] The device incorporates a generative AI model (e.g., GPT-4) to analyze the collected voice data. This generative AI model uses machine learning algorithms to detect specific phrases and patterns in conversations, including risky phrases like "cash card" and "PIN number."

[1628] 3. Emotion Engine

[1629] The device is equipped with an emotion engine that recognizes the user's emotions from collected voice data. The emotion engine determines the user's emotional state from the tone, speed, and patterns of the voice, detecting changes such as anxiety or tension.

[1630] 4. Alert Generation Methods

[1631] If the device determines that fraud is likely based on information from the generative model means and emotion engine, it generates an alert. This alert includes detailed information about the fraud and location information. The alert generation means operates when a relevant dangerous phrase or suspicious emotion change is detected.

[1632] 5. Means of notification

[1633] The generated alert is sent to the server via a notification method. The server receives the alert and notifies registered family members and the police. Notifications are sent via a dedicated app, SMS, email, etc. Location information and detailed information about the fraud are included, urging immediate action.

[1634] Specific examples

[1635] 1. If you receive a fraudulent phone call

[1636] The elderly user receives a call from an unknown number. The caller claims to be a bank employee and asks for their cash card and PIN number.

[1637] The device collects this conversation using a voice capture method and analyzes it with a generative AI model, which detects dangerous phrases such as "cash card" and "PIN number."

[1638] At the same time, the emotion engine recognizes changes in emotions from the tone, rate, and patterns of the user's voice, detecting anxiety and tension.

[1639] The generative model determines if there is a high risk of fraud and generates an alert, which includes details of the fraud and location information.

[1640] The alert is sent to a server via the notification method, and the server then sends a real-time notification to family members or the police via a dedicated app, SMS, email, etc., allowing for immediate action.

[1641] Prompt Sentence Examples

[1642] By using prompt sentences like the one below for the generative AI model, you can verify and adjust the system's operation.

[1643] "Please give an example of how you would respond if you were asked to verify your bank card over the phone."

[1644] "How does the emotion engine recognize changes in emotion from the tone and rate of a user's voice?"

[1645] "What information is included in the generated alert?"

[1646] This system combines analysis of user voice data with emotion recognition to detect cash card fraud in real time and promote quick and appropriate responses.

[1647] The flow of the identification process in the second embodiment will be described with reference to FIG.

[1648] Step 1:

[1649] When the device starts up, it loads the generative modeling means and emotion engine, activates the voice collection means, establishes an internet connection, and starts communication with the server, allowing the system to be fully functional.

[1650] Input: Power on

[1651] Output: System ready state

[1652] Step 2:

[1653] The device constantly collects the user's surrounding sounds through a built-in microphone. This audio data is processed digitally in real time and input to the generative modeling means and emotion engine.

[1654] Input: Audio around the user

[1655] Output: Digital audio data

[1656] Step 3:

[1657] The device's generative modeling means analyzes the collected voice data to detect specific phrases and patterns. A generative AI model (e.g., GPT-4) detects risky phrases such as "cash card" and "PIN number." This analysis assesses the likelihood of fraud.

[1658] Input: Digital audio data

[1659] Output: Detection results for specific phrases and fraud likelihood rating

[1660] Step 4:

[1661] The emotion engine analyzes the user's emotions from the voice data, determining the user's emotional state from the tone, rate, and patterns of the voice, and detecting emotional changes such as anxiety or tension.

[1662] Input: Digital audio data

[1663] Output: Evaluation result of the user's emotional state

[1664] Step 5:

[1665] The generative modeling means considers the emotion data from the emotion engine and reassess the risk of fraud. If it determines that fraud is likely, the alert generation means operates to generate an alert that includes detailed information and location information.

[1666] Input: Phrase detection results and emotional state evaluation results

[1667] Output: Alert generation decision and alert data

[1668] Step 6:

[1669] The generated alert is sent to the server via the notification method. The server receives the alert and notifies registered family members and the police. Notifications are sent via a dedicated app, SMS, email, etc.

[1670] Input: Alert data

[1671] Output: Notify family and police

[1672] Step 7:

[1673] The server evaluates responses from family members and police who receive the notification and, in some cases, provides further details or escalates the alert.

[1674] Input: Responses from family and police

[1675] Output: Additional details and escalation instructions

[1676] Through these processing steps, the system combines voice data analysis and emotion recognition to detect cash card fraud in real time and promptly respond.

[1677] (Application example 2)

[1678] Next, a description will be given of Application Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal."

[1679] In recent years, there has been an increase in cash card frauds targeting the elderly, and effective measures to prevent such damage are needed. Current security systems have difficulty detecting signs of fraud in real time, and are particularly ineffective against telephone fraud. Furthermore, because it is difficult for elderly people to recognize fraud themselves, more advanced fraud detection systems must be developed. Furthermore, fraud detection systems must be able to recognize users' emotions and take their changes into account for more accurate detection. This requires immediate and appropriate alerts in situations where there is a high possibility of fraud, and prompt notification to family members and the police.

[1680] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 2 is realized by the following means.

[1681] In this invention, the server includes voice collection means for constantly collecting voice data, generative model means for analyzing the collected voice data and evaluating the possibility of cash card fraud in real time, alert generation means for issuing an alert when the generative model means determines that there is a high possibility of cash card fraud, notification means for notifying family members and the police of the issued alert, an emotion engine for recognizing the user's emotions from the voice data, and means for reevaluating the possibility of fraud when the emotion engine determines that a change in emotion is suspicious. This enables highly accurate fraud detection that takes into account changes in the user's emotions and rapid issuance of alerts and notifications.

[1682] The "voice collection means" is a device or system for constantly collecting voices around the user.

[1683] A "generative model means" is a device or system that incorporates a machine learning algorithm for analyzing collected voice data and assessing the likelihood of cash card fraud in real time.

[1684] The "alert generation means" is a device or system that issues an alert when the generative model means determines that there is a high possibility of cash card fraud.

[1685] "Notification means" refers to a communication device or system for notifying family members or the police of an alert that has been issued.

[1686] An "emotion engine" is a device or algorithm that recognizes a user's emotions from voice data and determines changes in emotions.

[1687] A "means for reassessing" is a device or system that reassess the possibility of fraud when the emotion engine determines that a change in emotion is suspicious.

[1688] The system for implementing this invention is broadly composed of the following elements: a voice collection means, a generative model means, an emotion engine, an alert generation means, and a notification means. These elements are realized by combining multiple pieces of software and hardware.

[1689] 1. Audio collection method

[1690] The voice collection means uses a microphone built into a device such as a smartphone. It constantly collects voices around the user and converts them into data in real time. This voice data is then input into the generative model means and emotion engine in the next step.

[1691] 2. Generative Modeling Methods

[1692] The generative model means includes a machine learning algorithm for analyzing the collected voice data. Specifically, the generative AI model is used to detect specific phrases and patterns in the voice data that indicate possible bank card fraud. The model works in conjunction with a pre-trained database (past bank card fraud cases).

[1693] 3. Emotion Engine

[1694] The emotion engine includes algorithms for recognizing a user's emotional state from speech data. The emotion engine analyzes the tone, rate, and patterns of speech to determine changes in the user's emotions. This information is then used to assess the likelihood of fraud by means of a generative model.

[1695] 4. Alert Generation Methods

[1696] The alert generator operates based on information from the generative model and the emotion engine. If a fraud is deemed likely or a suspicious change in emotion is detected, an alert is generated. The alert includes detailed information about the fraud and data about the user's current location.

[1697] 5. Means of notification

[1698] The notification method sends the generated alert to a server, which processes the received alert and sends notifications to registered family members and the police via a dedicated application, SMS, email, etc.

[1699] Hardware and software used

[1700] Hardware: Smartphone (built-in microphone)

[1701] Software: Audio collection library (e.g., some_audio_library), generative AI model library (e.g., some_ml_library), emotion recognition library (e.g., emotion_recognition_library)

[1702] Cloud server: Alert notifications and log data storage

[1703] Adding specific examples

[1704] For example, consider a scenario in which an elderly person receives a scam call from an unknown number. If the generative AI model and emotion recognition engine detect a suspicious phrase or emotional change, the following prompt will be generated:

[1705] Scam Guardian Alert: Call from unknown number may be a bank card scam. Details: Scam details, Location: Current location

[1706] This prompt will alert you and your family immediately so you can take any necessary action.

[1707] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[1708] Step 1:

[1709] Start audio collection

[1710] When the device is started, it activates the built-in microphone and starts the audio collection method. It continuously collects audio data around the user and converts it into a digital format in real time. The input is the surrounding audio, and the output is the digitized audio data.

[1711] Step 2:

[1712] Analysis of audio data

[1713] The device inputs the collected voice data into a generative AI model means. The generative AI model uses a machine learning algorithm to analyze the voice data and detect specific phrases or patterns that may indicate possible bank card fraud. The input is the digitized voice data, and the output is the results of the analyzed phrases and patterns.

[1714] Step 3:

[1715] Emotion recognition

[1716] At the same time, the device also inputs the voice data into the emotion engine, which analyzes the tone, speed, and pattern of the voice to recognize the user's emotional state. The input is digitized voice data, and the output is the result of the user's emotional state assessment.

[1717] Step 4:

[1718] Fraud likelihood assessment

[1719] The terminal integrates information from the generative model means and the emotion engine. The generative model means reevaluates the emotion data provided by the emotion engine and determines the likelihood of fraud. The input is the analyzed phrases, patterns, and emotion data, and the output is the result of the fraud likelihood assessment.

[1720] Step 5:

[1721] Generate alerts

[1722] If a fraud possibility is determined, the device immediately generates an alert using the alert generation means. This alert includes detailed fraud information and location information. The input is the fraud possibility assessment result, and the output is the generated alert.

[1723] Step 6:

[1724] Alert Notification

[1725] The generated alert is sent to the server via the notification means. The server processes the received alert and sends a notification to registered family members or police. Notifications are sent via a dedicated app, SMS, email, etc. The input is the generated alert, and the output is the notification to family members or police.

[1726] Step 7:

[1727] Facilitating user and family support

[1728] The family member can check the received notification and immediately contact the user or arrange for on-site visit if necessary. The input is the received alert notification, and the output is the appropriate response action.

[1729] As a result of this, the system will be able to combine voice data and emotional data to detect cash card fraud with high accuracy and send notifications to users, their families, and the police in real time.

[1730] The specific processing unit 290 transmits the result of the specific processing to the robot 414. In the robot 414, the control unit 46A causes the speaker 240 and the control target 443 to output the result of the specific processing. The microphone 238 acquires voice indicating a user input regarding the result of the specific processing. The control unit 46A transmits voice data indicating the user input acquired by the microphone 238 to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the voice data.

[1731] The data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of the data generation model 58 is ChatGPT (Internet Search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search <url: https: gemini.google.com ?hl="ja">) and other generation AIs. The data generation model 58 is obtained by performing deep learning on a neural network. A prompt including an instruction is input to the data generation model 58, and inference data such as voice data indicating voice, text data indicating text, and image data indicating an image is also input. The data generation model 58 performs inference on the input inference data in accordance with the instruction indicated by the prompt, and outputs the inference result in a data format such as voice data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[1732] In the above embodiment, an example was given in which the specific processing is performed by the data processing device 12, but the technology of the present disclosure is not limited to this, and the specific processing may be performed by the robot 414.

[1733] The emotion identification model 59 as an emotion engine may determine the user's emotion according to a specific mapping. Specifically, the emotion identification model 59 may determine the user's emotion according to an emotion map (see FIG. 9), which is a specific mapping. Similarly, the emotion identification model 59 may determine the robot's emotion, and the identification processing unit 290 may perform identification processing using the robot's emotion.

[1734] FIG. 9 is a diagram illustrating an emotion map 400 on which multiple emotions are mapped. In the emotion map 400, emotions are arranged in concentric circles radiating from the center. Emotions closer to the center of the concentric circles are more primitive. Emotions representing states and actions arising from a state of mind are arranged on the outer edges of the concentric circles. The concept of emotion includes both affect and mental states. Emotions generally generated from reactions occurring in the brain are arranged on the left side of the concentric circles. Emotions generally induced by situational judgment are arranged on the right side of the concentric circles. Emotions generally generated from reactions occurring in the brain and induced by situational judgment are arranged on the upper and lower sides of the concentric circles. Furthermore, the emotion of "pleasure" is arranged on the upper side of the concentric circles, and the emotion of "discomfort" is arranged on the lower side. In this way, in the emotion map 400, multiple emotions are mapped based on the structure by which emotions are generated, and emotions that tend to occur simultaneously are mapped close to each other.

[1735] These emotions are distributed in the 3 o'clock direction on emotion map 400, and typically fluctuate between relief and anxiety. In the right half of emotion map 400, situational awareness dominates over internal sensations, resulting in a sense of calm.

[1736] The inside of emotion map 400 represents what is going on in the mind, and the outside of emotion map 400 represents behavior, so the further you go outside emotion map 400, the more visible the emotions become (the more they are expressed in behavior).

[1737] Human emotions are based on various balances, such as posture and blood sugar levels. When these balances deviate from the ideal, a state of discomfort is indicated, and when they approach the ideal, a state of pleasure is indicated. Emotions can also be created for robots, automobiles, and motorcycles, based on various balances, such as posture and remaining battery life. When these balances deviate from the ideal, a state of discomfort is indicated, and when they approach the ideal, a state of pleasure is indicated. An emotion map can be generated, for example, based on Dr. Mitsuyoshi's emotion map (Research on Voice Emotion Recognition and Emotional Brain Physiological Signal Analysis Systems, Tokushima University, Doctoral Dissertation: https: / / ci.nii.ac.jp / naid / 500000375379). The left half of the emotion map lists emotions belonging to the "reaction" domain, where sensation is dominant. The right half of the emotion map lists emotions belonging to the "situation" domain, where situational awareness is dominant.

[1738] The emotion map defines two emotions that promote learning. One is a negative emotion on the situation side, around the middle of "repentance" or "reflection." In other words, this occurs when the robot experiences negative emotions such as "I never want to feel this way again" or "I don't want to be scolded again." The other is a positive emotion on the response side, around "desire." In other words, this occurs when the robot experiences positive feelings such as "I want more" or "I want to know more."

[1739] The emotion identification model 59 inputs user input into a pre-trained neural network, obtains emotion values ​​indicating each emotion shown in the emotion map 400, and determines the user's emotion. This neural network is pre-trained based on multiple pieces of training data that are combinations of user input and emotion values ​​indicating each emotion shown in the emotion map 400. Furthermore, this neural network is trained so that emotions that are located close to each other have similar values, as in the emotion map 900 shown in FIG. 10. FIG. 10 shows an example in which multiple emotions, "relieved," "calm," and "reassuring," have similar emotion values.

[1740] The system according to the present disclosure has been described above mainly with respect to the functions of the data processing device 12, but the system according to the present disclosure is not necessarily implemented on a server. The system according to the present disclosure may be implemented as a general information processing system. The present disclosure may be implemented, for example, as a software program running on a personal computer or an application running on a smartphone, etc. The method according to the present disclosure may be provided to users in the form of SaaS (Software as a Service).

[1741] In the above embodiment, an example was given in which the specific processing is performed by one computer 22, but the technology of the present disclosure is not limited to this, and the specific processing may be distributed and performed by a plurality of computers including the computer 22. For example, the data generation model 58 may be provided in an external device of the data processing device 12, and data may be generated in the external device in accordance with input data.

[1742] In the above embodiment, an example in which the specific processing program 56 is stored in the storage 32 has been described, but the technology of the present disclosure is not limited to this. For example, the specific processing program 56 may be stored in a portable, computer-readable, non-transitory storage medium such as a USB (Universal Serial Bus) memory. The specific processing program 56 stored in the non-transitory storage medium is installed in the computer 22 of the data processing device 12. The processor 28 executes the specific processing in accordance with the specific processing program 56.

[1743] 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.

[1744] It is not necessary to store all of the specific processing program 56 in a storage device such as a server connected to the data processing device 12 via the network 54, or to store all of the specific processing program 56 in the storage 32; only a portion of the specific processing program 56 may be stored.

[1745] The hardware resource for executing a specific process can be any of the following processors: An example of a processor is a CPU, which is a general-purpose processor that functions as a hardware resource for executing a specific process by executing software, i.e., a program. Another example of a processor is a dedicated electrical circuit, such as an FPGA (Field-Programmable Gate Array), a PLD (Programmable Logic Device), or an ASIC (Application Specific Integrated Circuit), which is a processor with a circuit configuration designed specifically for executing a specific process. Each processor has built-in or connected memory, and each processor uses the memory to execute the specific process.

[1746] The hardware resource that executes the specific processing may be configured with one of these various processors, or may be configured with a combination of two or more processors of the same or different types (for example, a combination of multiple FPGAs, or a combination of a CPU and an FPGA). Also, the hardware resource that executes the specific processing may be a single processor.

[1747] As an example of a system configured with a single processor, first, one processor is configured by combining one or more CPUs and software, and this processor functions as a hardware resource that executes a specific process. Second, there is a system that uses a processor that realizes the functions of an entire system including multiple hardware resources that execute a specific process on a single IC chip, as typified by SoC (System-on-a-chip). In this way, a specific process is realized using one or more of the above-mentioned various processors as hardware resources.

[1748] Furthermore, the hardware structure of these various processors can be, more specifically, an electric circuit that combines circuit elements such as semiconductor devices. The specific processing described above is merely an example. Therefore, it goes without saying that unnecessary steps may be deleted, new steps may be added, or the processing order may be rearranged, without departing from the spirit of the invention.

[1749] The above-described description and illustrations are a detailed explanation of the parts related to the technology of the present disclosure and are merely an example of the technology of the present disclosure. For example, the above description of the configuration, functions, actions, and effects is an explanation of an example of the configuration, functions, actions, and effects of the parts related to the technology of the present disclosure. Therefore, it goes without saying that unnecessary parts may be deleted, new elements may be added, or replacements may be made to the above-described description and illustrations within the scope of the gist of the technology of the present disclosure. Furthermore, to avoid confusion and facilitate understanding of the parts related to the technology of the present disclosure, the above-described description and illustrations omit explanations of common technical knowledge that do not require particular explanation to enable the implementation of the technology of the present disclosure.

[1750] All publications, patent applications, and technical standards mentioned in this specification are herein incorporated by reference to the same extent as if each individual publication, patent application, or technical standard was specifically and individually indicated to be incorporated by reference.

[1751] The following is further disclosed regarding the above embodiment.

[1752] (Claim 1)

[1753] [A voice collection means for constantly collecting voice data;]

[1754] [A generative model that analyzes collected voice data and evaluates the likelihood of bank card fraud in real time; and]

[1755] [an alert generation means for issuing an alert when the generative model means determines that there is a high possibility of cash card fraud; and]

[1756] [Method of notifying family members and police of the alert that was issued]

[1757] A system including:

[1758] (Claim 2)

[1759] The system according to claim 1, wherein the alert sent by the notification means includes location information of the sending point.

[1760] (Claim 3)

[1761] The system according to claim 1, wherein the generative model means operates in conjunction with a database of past cash card fraud cases.

[1762] "Example 1"

[1763] (Claim 1)

[1764] [an acoustic recording means for constantly collecting audio data; and]

[1765] [A generative model that analyzes collected voice data and assesses the likelihood of fraud in real time; and]

[1766] [an alert generation means for issuing an alert when the generative model means determines that fraud is likely; and]

[1767] [A means of communication to notify family members and police of the alert that has been issued]

[1768] A system including:

[1769] (Claim 2)

[1770] The system according to claim 1, wherein the alert transmitted by the information transmission means includes location information of the transmission point.

[1771] (Claim 3)

[1772] The system of claim 1, wherein the generative model means operates in conjunction with a database of past fraud cases.

[1773] "Application Example 1"

[1774] (Claim 1)

[1775] [A voice collection means for constantly collecting voice data;]

[1776] [A generative AI model means for analyzing collected voice data and assessing potential fraud in real time; and]

[1777] [an alert generation means for issuing an alert when the generative AI model means determines that there is a high possibility of fraudulent activity; and]

[1778] [Means of notifying relevant parties and crime prevention agencies of the issued alert, and]

[1779] A system including:

[1780] (Claim 2)

[1781] The system according to claim 1, wherein the alert sent by the notification means includes location information of the sending point and details of the danger phrase.

[1782] (Claim 3)

[1783] The system of claim 1, wherein the generative AI model means operates in conjunction with a database of past fraud cases.

[1784] "Example 2: Combining Emotion Engines"

[1785] (Claim 1)

[1786] [A voice collection means for constantly collecting voice data;]

[1787] [A generative model that analyzes collected voice data and evaluates the likelihood of bank card fraud in real time; and]

[1788] [An emotion engine that analyzes voice data to recognize emotions;]

[1789] [an alert generation means for reassessing the likelihood of bank card fraud based on data from the generative model means and the sentiment engine, and issuing an alert if it is determined that fraud is likely; and]

[1790] [Method of notifying family members and police of the alert that was issued]

[1791] A system including:

[1792] (Claim 2)

[1793] The system of claim 1, wherein the alert sent by the notification means includes location information of the point of origin and detailed information about the fraud.

[1794] (Claim 3)

[1795] The system according to claim 1, wherein the generative model means operates in cooperation with a database of past cash card fraud cases and a database of user emotions.

[1796] "Application example 2 when combining emotion engines"

[1797] (Claim 1)

[1798] [A voice collection means for constantly collecting voice data;]

[1799] [A generative model that analyzes collected voice data and evaluates the likelihood of bank card fraud in real time; and]

[1800] [an alert generation means for issuing an alert when the generative model means determines that there is a high possibility of cash card fraud; and]

[1801] [Method of notifying family members and police of the alert that was issued]

[1802] [An emotion engine that recognizes user emotions from voice data]

[1803] [Means to reassess the likelihood of fraud if the sentiment engine determines that a change in sentiment is suspicious; and]

[1804] A system including:

[1805] (Claim 2)

[1806] The system according to claim 1, wherein the alert sent by the notification means includes location information of the sending point.

[1807] (Claim 3)

[1808] The system of claim 1, wherein the generative model means operates in conjunction with a database of past bank card fraud cases and takes into account emotion data provided by the emotion engine. [Explanation of symbols]

[1809] 10, 210, 310, 410 Data Processing Systems 12 Data Processing Device 14 Smart Devices 214 Smart Glasses 314 Headset-type terminal 414 Robot< / url:> < / url:> < / url:> < / url:>

Claims

1. a voice collecting means for constantly collecting voice data; a generative modeling means for analyzing the collected voice data and evaluating the possibility of cash card fraud in real time; an alert generation means for issuing an alert when the generative model means determines that there is a high possibility of cash card fraud; A notification method to notify family members and police of the alert that was issued, A system including:

2. 2. The system according to claim 1, wherein the alert sent by the notification means includes location information of the sending point.

3. 2. The system according to claim 1, wherein the generative model means operates in conjunction with a database of past cash card fraud cases.

Images