system

Abstract

We provide the system. [Solution] A means for acquiring an audio input signal, means for converting the aforementioned audio input signal into text data, A means for analyzing the aforementioned text data using a generative model to detect patterns related to fraud, A means of generating a warning signal when it is determined that there is a possibility of fraud, Means for transmitting the aforementioned warning signal to a registered terminal or communication device, A system that includes this.

Description

Technical Field

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

Background Art

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

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] There is a lack of means to prevent fraud, which has become a social problem, especially fraud such as refund fraud targeting the elderly. With only the previous enlightenment and warnings, it is difficult to effectively prevent damage to particularly vulnerable groups. Therefore, a system that can detect signs of fraud in real time and send warnings quickly is needed. Furthermore, the system must be easy to use for the elderly and be able to respond quickly.

Means for Solving the Problems

[0005] This invention provides means for acquiring an audio input signal and converting that signal into text data. It also analyzes this text data using a generative model to detect patterns related to fraud. Furthermore, it solves this problem by constructing means for generating a warning signal when a high probability of fraud is determined and transmitting it to a registered terminal or communication device. Since the audio input signal is compressed and encrypted during transmission, it enables efficient processing while protecting privacy. In addition, the warning signal includes a summary of the conversation, which can support a quick and accurate response.

[0006] A "voice input signal" is a signal obtained by converting the user's voice into digital data, and it is an information source that is further processed within the system.

[0007] "Text data" refers to string data obtained by processing an audio input signal, representing the content of the audio as text.

[0008] A "generative model" is a computational model built on machine learning algorithms that learns from past data to detect specific patterns or features.

[0009] A "warning signal" is a signal generated when a potential fraud is detected, and it contains information to urgently inform others of the situation.

[0010] A "terminal" is a portable device or communication device used by a user that is capable of receiving warning signals from the system.

[0011] A "communication device" is a device used to send and receive digital data, and it exchanges data with a system when connected to a network.

[0012] "Conversation summary information" is data that summarizes the main points of a detected fraudulent conversation and is transmitted along with a warning signal. [Brief explanation of the drawing]

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

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

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

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

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

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

[0019] In the following embodiments, the signed communication interface (I / F) is an interface that includes a communication processor and an antenna, etc. The communication interface manages communication between multiple computers. Examples of communication standards applicable to the communication interface include wireless communication standards such as 5G (5th Generation Mobile Communication System), Wi-Fi (registered trademark), or Bluetooth (registered trademark).

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

[0021] [First Embodiment]

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

[0023] As shown in Figure 1, the data processing system 10 includes a data processing device 12 and a smart device 14. An example of the data processing device 12 is a server.

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

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

[0026] The reception device 38 is equipped with a touch panel 38A and a microphone 38B, etc., and receives user input. The touch panel 38A receives user input by detecting contact with an object (e.g., a pen or finger). The microphone 38B receives user input by detecting the user's voice. The control unit 46A transmits data indicating the user input received by the touch panel 38A and microphone 38B to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the data indicating the user input.

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

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

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

[0030] As shown in Figure 2, in the data processing device 12, a specific processing is performed by the processor 28. A specific processing program 56 is stored in the storage 32. The specific processing program 56 is an example of a "program" related to the technology of this disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific processing is realized by the processor 28 operating as a specific processing unit 290 according to the specific processing program 56 executed on the RAM 30.

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

[0032] In the smart device 14, the processor 46 performs the reception output processing. The storage 50 stores the reception output program 60. The reception output program 60 is used in conjunction with a specific processing program 56 by the data processing system 10. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output processing is realized by the processor 46 operating as a control unit 46A according to the reception output program 60 executed on the RAM 48.

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

[0034] This invention uses an application installed on a user's portable device to detect fraud in real time and issue warnings. Specifically, the device continuously acquires voice input signals and sends that data to a server for analysis.

[0035] First, the device records sounds occurring around the user through a high-sensitivity microphone. This audio input signal is compressed and encrypted within the device to ensure the protection of personal information, and then sent to the server periodically or in real time. The compression technology reduces data volume while maintaining sound quality.

[0036] Next, the server converts the received audio into text data. This utilizes speech recognition technology and achieves high accuracy through preprocessing such as noise reduction and speaker separation. This text data is further analyzed by a generative model to determine whether patterns and keywords related to fraud exist.

[0037] When a potential scam is detected, the server immediately generates a warning signal. This signal is sent to the user's device and to registered communication devices such as family members and the police. The warning signal includes a summary of the conversation and a description of the situation, designed to support a quick and appropriate response.

[0038] For example, a user might be asked for personal information over the phone under the pretext of a refund. In this case, the device records the conversation and sends the information to a server. The server analyzes the data and, if it detects any potentially dangerous phrases, immediately sends a warning to the family or the police. This allows the user and those around them to take swift and appropriate action.

[0039] Thus, the present invention provides a system that protects users, especially the elderly, from the threat of fraud by detecting fraud in real time through voice and issuing a rapid warning.

[0040] The following describes the processing flow.

[0041] Step 1:

[0042] The device constantly monitors ambient noise and records the user's voice. This voice input signal is buffered at specific time intervals.

[0043] Step 2:

[0044] The terminal converts the buffered audio into digital data and compresses it to improve communication efficiency. Then, it encrypts the data to ensure security.

[0045] Step 3:

[0046] The device sends compressed and encrypted audio digital data to the server at regular intervals. Transmission is done via Wi-Fi or a mobile data network.

[0047] Step 4:

[0048] The server decodes the received audio data, performs preprocessing such as noise reduction and speaker separation, and then converts it into text data using speech recognition technology.

[0049] Step 5:

[0050] The server uses a generative model to analyze text data and detect patterns and keywords that may indicate fraudulent activity.

[0051] Step 6:

[0052] If the server determines that a conversation may be fraudulent, it will generate a warning signal. This warning signal will include a summary of the relevant conversation.

[0053] Step 7:

[0054] The server sends a warning signal to the user's device and to pre-registered family members and police communication devices.

[0055] Step 8:

[0056] The device notifies the user using audio alerts and screen displays based on the received warning signals. The user can then take appropriate action.

[0057] (Example 1)

[0058] Next, we will describe Example 1. In the following description, the data processing device 12 will be referred to as the "server," and the smart device 14 will be referred to as the "terminal."

[0059] In modern society, fraud, particularly targeting the elderly, remains a persistent problem. Sophisticated scams conducted over the phone or in person are increasingly resulting in the illegal acquisition of personal information and money, necessitating swift and effective countermeasures. Traditional prevention measures struggle to detect fraud in real time, often leading to awareness only after the fact. Therefore, a system that warns of potential fraud in advance is needed.

[0060] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 1 is realized by the following means.

[0061] In this invention, the server includes a device for acquiring audio, a device for converting the audio into digital data, and a device for compressing and encrypting the digital data and transmitting it over a communication channel. This makes it possible to detect fraudulent activity in real time and to securely transmit relevant information.

[0062] "Sound" refers to sounds that are generated as vibrations in the air and perceived by human hearing.

[0063] "Digital data" refers to data obtained by converting analog information, such as audio, into numerical data that can be processed by electronic devices.

[0064] "Compression" is a process that reduces the size of data, which enables more efficient data storage and transfer.

[0065] "Encryption" refers to the process of transforming data using a specific algorithm to prevent it from being deciphered by a third party.

[0066] "Communication channel" refers to a path or means used for data transmission, including wired or wireless networks.

[0067] "Generative technology" refers to technologies that enable artificial intelligence to analyze data according to specific purposes and generate new value or information.

[0068] A "warning signal" is alert information issued when specific conditions are met, with the purpose of prompting the recipient to take some kind of action.

[0069] "Designated device" refers to a device that has been pre-registered or configured for receiving data.

[0070] "Communication equipment" refers to devices used for communication, including mobile phones and computers.

[0071] This invention is a system that uses a user-carried terminal and a server to detect fraud in real time and issue warnings. Specifically, the user-carried terminal is equipped with a high-sensitivity microphone that constantly captures ambient sounds. This sound is converted into digital data, and then compressed (e.g., using the FLAC codec) and encrypted (AES encryption) technologies are used to reduce the data size and ensure security. The data is then transmitted to the server via a communication channel. The SSL / TLS protocol is used to ensure the security of the data during transmission.

[0072] The server decrypts the received data and removes the compression to restore the original audio data. Then, speech recognition technology (e.g., a speech recognition API) is used to convert the audio data into text data. This text data is then analyzed using generative techniques to detect patterns related to fraud. For the generative techniques, widely used generative AI models are employed.

[0073] If the server determines that a fraudulent activity is likely, it generates a warning signal and sends the warning to the user's terminal or a pre-registered communication device. The warning includes a summary of the relevant conversation and a description of the situation, enabling a quick response.

[0074] As a concrete example, consider a case where a user is asked for personal information over the phone under the pretext of a refund. When such a conversation takes place, the device records the audio and waits for analysis on the server. If the server detects a pattern that includes both "refund" and "bank information," the generated warning will include an explanation such as "possible suspicious refund request" and will be instantly displayed on the user's device.

[0075] An example of a prompt message might be: "Build a system that records potentially fraudulent conversations with users and immediately sends a warning when danger is detected."

[0076] This system makes it possible to detect fraud in real time and prevent victims from becoming victims.

[0077] The flow of the specific processing in Example 1 will be explained using Figure 11.

[0078] Step 1:

[0079] The device acquires ambient sound using a high-sensitivity microphone. This audio data is acquired in analog format and converted into digital data. Specifically, an ADC (Analog-to-Digital Converter) is used to convert the analog signal to a digital signal. The input is the raw sound around the user, and the output is digitized audio data.

[0080] Step 2:

[0081] The terminal compresses and encrypts the acquired digital audio data. The data is compressed using the FLAC codec to reduce its size. Next, security is ensured using AES encryption technology. As a result, the input is digital audio data, and the output is compressed and encrypted audio data.

[0082] Step 3:

[0083] The terminal sends compressed and encrypted audio data to the server. The SSL / TLS protocol is used to ensure secure data transmission. The input is compressed and encrypted audio data, and the output is the received audio data stored on the server.

[0084] Step 4:

[0085] The server decrypts the received audio data, removes the compression, and obtains the original digital audio data. First, it decrypts the data using AES, and then removes the FLAC compression. The input is encrypted audio data, and the output is the restored digital audio data.

[0086] Step 5:

[0087] The server converts the restored digital audio data into text data using speech recognition technology. It uses a specific speech recognition API for the audio-to-text conversion. The input is the restored digital audio data, and the output is the generated text data.

[0088] Step 6:

[0089] The server uses a generative AI model to analyze text data and detect patterns and keywords related to fraud. The input is text data, and the output is an analysis result indicating the possibility of fraud. The generative AI model performs analysis based on a pre-trained database.

[0090] Step 7:

[0091] If the server detects a potential scam, it generates a warning signal and sends it to the user's terminal or registered communication device as needed. The warning signal includes a summary of the relevant conversation. The input is the analysis result regarding the potential scam, and the output is the generated warning signal.

[0092] In this way, the fraud detection system functions through the coordination of each step.

[0093] (Application Example 1)

[0094] Next, we will explain Application Example 1. In the following explanation, the data processing device 12 will be referred to as the "server," and the smart device 14 will be referred to as the "terminal."

[0095] Traditionally, fraud detection systems have struggled to quickly and accurately detect fraud conducted via telephone and other communication methods. Especially with the increasing number of fraud cases targeting the elderly, real-time fraud detection and warnings are in high demand. Current systems have limitations in analysis accuracy and warning speed, failing to adequately protect user safety.

[0096] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 1 is realized by the following means.

[0097] In this invention, the server includes a device for acquiring an audio input signal, a device for converting the audio input signal into a text representation, and a device for analyzing the text representation using a generation model and detecting patterns related to fraud. This enables the rapid generation of a warning signal and notification including safety information to registered information processing devices when the possibility of fraud is determined.

[0098] "Audio input signal" refers to a signal collected using a high-sensitivity microphone from sounds occurring around the user.

[0099] "Text representation" refers to character data obtained by converting an audio input signal using speech recognition technology.

[0100] A "generative model" is an algorithm that uses machine learning to analyze text data and detect patterns and keywords related to fraud.

[0101] A "warning signal" is a signal that is generated when a potential fraud is detected, and it sends a warning to the user and registered information processing device.

[0102] An "information processing device" is a communication-enabled device owned by a user or registered party.

[0103] "Encryption" is a process used to prevent the contents of data from being known to third parties.

[0104] A "communication line" is a network path used to send and receive data between a server and a client.

[0105] An "external processing unit" is a system that functions as a central server, performing voice data analysis and notification management.

[0106] "Analysis results using a generative AI model" refers to information about the possibility of fraud based on an analysis of text data converted from speech.

[0107] To implement this invention, a user-facing information processing device, such as a smartphone, is required. The device is equipped with a high-sensitivity microphone, which is used to collect sounds from the user's surroundings. The audio input signal is encrypted and compressed within the device and transmitted in real time or periodically to an external processing device, a server. For communication, an encrypted communication network (e.g., HTTPS protocol) is used to ensure security.

[0108] The server converts the received audio data into a text representation using the Google® Speech-to-Text API. This text is then analyzed for fraud-related patterns using the generative AI model GPT-3®. The generative AI model's analysis also includes preprocessing such as noise reduction and speaker separation.

[0109] If a potential scam is detected, the server generates a warning signal and sends it to the device or registered information processing device. Alternatively, the Twilio API can be used to send the warning signal as an SMS or app notification. In this case, the warning signal includes summary data of the relevant conversation and analysis results using a generative AI model, prompting the user to take prompt and appropriate action.

[0110] As a concrete example, suppose a user receives a phone call requesting personal information in connection with a refund. In this case, the device records the conversation and sends it to a server. The server detects dangerous phrases such as "refund" and "personal information" and immediately sends a warning to the user and their registered family members, thereby ensuring the safety of the user and those around them.

[0111] An example of a prompt message might be: "Please determine if the following text contains potential fraud: 'You are eligible for a refund, please provide your bank information.' If it does, please also explain the factors that indicate it is fraudulent."

[0112] The flow of a specific process in Application Example 1 will be explained using Figure 12.

[0113] Step 1:

[0114] The device collects audio from the user's surroundings through a high-sensitivity microphone. The input is a real-time audio signal, which is then encrypted and compressed within the device. The output is compressed audio data in a format that can be transmitted quickly.

[0115] Step 2:

[0116] The terminal sends compressed and encrypted audio data to a server, which is an external processing unit, via a secure communication network. The input is compressed audio data, which is securely transferred to the server as output. Specifically, encrypted packets are created and sent using the HTTPS protocol.

[0117] Step 3:

[0118] The server uses the Google Speech-to-Text API to convert received audio data into text. The input is compressed audio data, which is then denoised and speaker-separated before generating plain text data as output. The server then passes this data to a generation AI model.

[0119] Step 4:

[0120] The server uses a generative AI model, specifically GPT-3, to analyze text data. The input is transformed text data, utilizing prompts to detect patterns and keywords related to fraud. The output is the analysis result, including data on the likelihood of fraud and specific risk factors.

[0121] Step 5:

[0122] If the analysis reveals a potential for fraud, the server generates a warning signal. The input is the analysis result, and the output is the warning signal based on this result. Specifically, it generates an alert message containing the problematic phrase or analysis result.

[0123] Step 6:

[0124] The server sends warning signals generated via the Twilio API to users and registered information processing devices. The input is the warning signal, and the output is a notification or SMS sent to the user's device or registered recipient. Specifically, a notification is immediately displayed on the user's smartphone.

[0125] Furthermore, an emotion engine that estimates the user's emotions may be incorporated. That is, the identification processing unit 290 may use the emotion identification model 59 to estimate the user's emotions and perform identification processing using the user's emotions.

[0126] This invention provides a system that analyzes voice input signals obtained from a user to detect the possibility of fraudulent activity while simultaneously recognizing the user's emotional state. In particular, by considering the user's psychological response in addition to the content of the voice, the reliability of fraud detection is enhanced.

[0127] The terminal acquires an audio input signal by recording the user's voice. The acquired audio data is compressed and encrypted on the terminal and sent to the server as needed. The server decrypts the received audio data and begins processing.

[0128] The server first uses speech recognition technology to convert the audio data into text data. This text data is then analyzed by a generative model designed to detect signs of fraud. The model compares it to existing fraud patterns to find matches.

[0129] Furthermore, the server uses an emotion engine to analyze the user's voice to determine their emotions. This emotion engine determines whether the user is experiencing psychological states such as anxiety or confusion, based on acoustic characteristics such as tone, tempo, and intonation.

[0130] If fraudulent activity is detected or the user is found to be in an unusual emotional state, the server generates a warning signal. This warning signal is sent to the user's device and to registered family members and police communications devices. The warning signal includes details about the detected fraud pattern and a summary of the user's emotional state.

[0131] To give a concrete example, suppose a user is talking about a "risky investment opportunity" over the phone. The device records this conversation and sends it to a server. The server converts it to text, and if the emotion engine detects signs of anxiety in the user's voice, it immediately generates a warning signal. This warning signal is then sent to the user's family, allowing them to take prompt action.

[0132] Thus, by combining voice data and emotional information, the present invention realizes a comprehensive system that not only protects users from fraud but also provides psychological support.

[0133] The following describes the processing flow.

[0134] Step 1:

[0135] The device constantly monitors the audio surrounding the user and starts recording. The audio input signal is acquired through a high-sensitivity microphone.

[0136] Step 2:

[0137] The device compresses recorded audio input signals in real time and encrypts them to ensure security. This achieves both communication efficiency and information protection.

[0138] Step 3:

[0139] The device transmits compressed and encrypted audio data to the server at regular intervals. This transmission uses either Wi-Fi or mobile data, depending on network conditions.

[0140] Step 4:

[0141] The server decodes the received audio data and converts it into text data through a speech recognition system. At this stage, noise reduction is applied to obtain clear text information.

[0142] Step 5:

[0143] The server feeds the text data converted from the speech into a generative model, which analyzes patterns and keywords that suggest potential fraud.

[0144] Step 6:

[0145] The server simultaneously uses an emotion engine to evaluate the user's emotional state based on the tone and tempo of their voice. This can detect unusual mental states such as anxiety or tension.

[0146] Step 7:

[0147] The server generates a warning signal if it detects signs of fraud or unusual emotional states. This warning signal includes information about the analyzed fraud patterns and the user's emotional state.

[0148] Step 8:

[0149] The server sends a warning signal to the user's device and the communication devices of registered family members or the police. This allows for real-time sharing of the situation.

[0150] Step 9:

[0151] The device notifies the user using audio and visual feedback based on the received warning signal. This allows the user to become more vigilant against fraud and to coordinate with family members.

[0152] (Example 2)

[0153] Next, we will describe Example 2. In the following description, the data processing device 12 will be referred to as the "server" and the smart device 14 as the "terminal".

[0154] Conventional acoustic signal analysis systems aimed at detecting fraudulent activity, but they did not adequately consider the user's psychological state. Therefore, there is a need for a method that can quickly grasp emotions and psychological changes to enable more appropriate warnings and interventions.

[0155] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 2 is realized by the following means.

[0156] In this invention, the server includes means for converting acoustic signals into document data, means for analyzing the document data using a generative model to detect patterns related to fraud, and means for analyzing acoustic characteristics to determine the user's psychological state. This makes it possible to detect both signs of fraudulent activity and the user's psychological state, enabling faster and more appropriate warnings and responses.

[0157] An "acoustic signal" is a representation of sound as an electrical signal, which is acquired by devices such as microphones.

[0158] "Document data" refers to data in text format that has been converted from acoustic signals using speech recognition technology.

[0159] A "generative model" is a statistical model used to analyze data and detect specific patterns using machine learning and artificial intelligence techniques.

[0160] "Acoustic characteristics" refer to physical properties such as tone, tempo, and intonation that are extracted from audio data.

[0161] "Psychological state" refers to information that indicates the user's emotions and mental condition, analyzed based on the tone, tempo, and intonation of their voice.

[0162] A "warning signal" is a signal that is generated and transmitted to draw attention when the possibility of fraud or an unusual psychological state is detected.

[0163] "Communication equipment" refers to devices used to send and receive data and signals, and includes, for example, smartphones and tablets.

[0164] An "information transmission device" is a device that has the function of transmitting digital data or signals to other devices.

[0165] This invention is a system that uses acoustic signals to detect fraudulent activity and analyze the user's psychological state. The system converts acoustic signals into document data and uses that data to detect fraud patterns and analyze emotions.

[0166] The terminal acquires audio through the user's microphone. The acquired audio signal is compressed and encrypted locally and sent to the server over the communication network. A common encryption library is used for the encryption process.

[0167] The server decodes the received acoustic signal and first converts it into document data using advanced speech recognition software. Speech recognition technologies such as the Google Speech-to-Text API can be applied at this stage. Next, this document data is analyzed by a generative AI model. This generative AI model is trained on an extensive dataset of fraud cases and detects whether the text data contains signs of fraud.

[0168] Furthermore, the server is equipped with an emotion engine that determines psychological state based on acoustic characteristics. This emotion engine analyzes the tone, tempo, and intonation of the voice to evaluate whether the user is in a specific psychological state, such as anxiety or confusion.

[0169] For example, if a user is talking on the phone about a "business opportunity to make a quick profit," the device records the conversation and sends it to a server. The server converts the audio to text and uses an emotion engine to detect signs of anxiety. If this matches signs of fraud, the server immediately generates a warning signal and sends it to the user's relatives or other relevant parties.

[0170] Examples of prompts to input into a generative AI model:

[0171] Design a system that determines a user's psychological state from their acoustic signals and detects the possibility of fraudulent activity. The input should be acoustic signal data, and the output should be the degree of matching with fraud patterns and the results of the psychological state analysis.

[0172] Therefore, the system can comprehensively address fraudulent activities and provide psychological support.

[0173] The flow of the specific processing in Example 2 will be explained using Figure 13.

[0174] Step 1:

[0175] The device acquires the user's audio signal through a microphone. The input is the user's raw voice, and the output is an audio signal in digital format. In this step, the audio is captured in real time and data conversion processing is performed. The acquired audio signal is compressed and encrypted on the spot to ensure security.

[0176] Step 2:

[0177] The terminal transmits an encrypted acoustic signal to the server using a communication network. The input is the encrypted acoustic signal, and the output is the data transmitted to the server. In this step, the data travels to the server via a network such as the internet.

[0178] Step 3:

[0179] The server decrypts the received acoustic signal to prepare the data. The input is an encrypted acoustic signal, and the output is the available acoustic signal. Decryption is performed using a common encryption technique.

[0180] Step 4:

[0181] The server uses advanced speech recognition technology to convert acoustic signals into document data. The input is the decoded acoustic signal, and the output is text data. This process is performed using a speech recognition engine such as the Google Speech-to-Text API.

[0182] Step 5:

[0183] The server analyzes document data using a generative AI model. The input is text data generated by speech recognition, and the output is an analysis of the likelihood of fraud. The model scans the text and looks for parts that match known fraud patterns.

[0184] Step 6:

[0185] The server operates an emotion engine that determines psychological state based on acoustic characteristics. The input is an acoustic signal, and the output is data indicating the user's emotional state. This analysis is performed based on factors such as the tone, tempo, and intonation of the voice.

[0186] Step 7:

[0187] The server generates an alert signal if it detects signs of fraud or unusual psychological states. The input is fraud detection results and psychological state data, and the output is the alert signal. This signal includes details about the fraud pattern and psychological state.

[0188] Step 8:

[0189] The server sends the generated warning signal to a pre-registered communication device. The input is the warning signal, and the output is the information sent to the communication device. The warning signal reaches the user's relatives and relevant organizations quickly, enabling a rapid response.

[0190] (Application Example 2)

[0191] Next, we will explain application example 2. In the following explanation, the data processing device 12 will be referred to as a "server" and the smart device 14 as a "terminal".

[0192] In modern society, fraud via voice calls is on the rise, making countermeasures an urgent necessity. Furthermore, victims often experience psychological distress and confusion during fraudulent activities, highlighting the need for support. However, current technology is insufficient to simultaneously detect voice-based fraud and analyze emotional states. Therefore, the challenge lies in rapidly detecting potential fraudulent activity and providing information that takes into account the user's emotional state.

[0193] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means.

[0194] In this invention, the server includes a medium for acquiring audio signals, a medium for converting audio signals into corresponding text information, a medium for analyzing the text information using a generation model to identify patterns related to fraud, and a medium for analyzing the user's emotional state based on the audio signals and generating psychological support information if the possibility of fraud or a specific emotional state is recognized. This enables early detection of fraudulent activity and the provision of appropriate support information based on the user's emotional state.

[0195] A "medium for acquiring audio signals" is a device or system that receives audio signals input from a user and records them for processing.

[0196] A "medium for converting into text information" is a device or system that analyzes acquired audio signals and expresses their content as text data.

[0197] "Using generative models" refers to a method of detecting specific patterns or features based on speech or text data, utilizing machine learning and AI technologies.

[0198] A "medium for identifying patterns related to fraud" is a device or mechanism for identifying potential fraud from audio or text based on the characteristics of known fraudulent activities.

[0199] A "medium for analyzing a user's emotional state" is a device or mechanism that analyzes information such as tone, rhythm, and intonation of an audio signal to identify the user's emotions.

[0200] A "medium for generating psychological support information" is a device or mechanism that generates information to provide reassurance or warnings based on the user's emotional state and the possibility of fraud.

[0201] "Warning information" refers to information used to alert those involved when potential fraud or specific emotional states are detected.

[0202] "A medium for transmitting information to a registered device or communication device" refers to a device or mechanism for transmitting generated warning information or support information to a communication device used by a user or related party.

[0203] This invention is a system that converts audio signals into text information using a speech recognition system, analyzes that text information using a generation AI model to detect the possibility of fraud, and further determines the user's emotional state from the audio signal. To enable this, the server and terminal must coordinate their operations.

[0204] The server collects audio data through hardware designed to acquire audio signals (e.g., audio input devices such as smartphones). The acquired audio signals are converted into text information using speech recognition technology on the server. This speech recognition uses common speech recognition libraries or cloud-based APIs (e.g., Google Speech-to-Text API).

[0205] Next, the server uses a generative AI model (e.g., a machine learning model) to analyze the converted text information. This analysis identifies patterns related to fraud. By utilizing a trained model to compare with known fraud patterns, advanced fraud detection is achieved.

[0206] Meanwhile, the server analyzes the user's emotional state from the audio signal. This uses an emotion analysis engine based on acoustic characteristics such as tone, tempo, and intonation. This allows the server to determine whether the user is experiencing emotional states such as anxiety or confusion.

[0207] If potential fraud or a specific emotional state is detected, the server generates a warning message and transmits it to the terminal or communication device. This allows users and registered parties to be aware of the risk of fraud or if the user is in need of psychological support.

[0208] For example, if a user receives a suspicious financial call, the sentiment analysis engine might detect anxiety in the user's voice. In this case, the system would immediately generate a warning and notify the user's family and registered contacts, enabling early intervention.

[0209] As an example of a prompt message to a generative AI model, it could be something like, "Detect if this call is a scam. Text: 'There is a new investment opportunity...'" This allows the model to analyze the likelihood of fraud based on text information.

[0210] The flow of a specific process in Application Example 2 will be explained using Figure 14.

[0211] Step 1:

[0212] The terminal uses a microphone to acquire the user's voice signal. It uses the raw voice signal as input, converts it to a digital format, and sends it to the server. This allows for the preparation of a suitable data format for preprocessing of speech recognition.

[0213] Step 2:

[0214] The server converts the acquired audio signal into text information using speech recognition software (e.g., Google Speech-to-Text API). The input is a digital audio signal, and the output is the corresponding string data. Text information is obtained by analyzing the acoustic characteristics of the audio signal and converting it into text.

[0215] Step 3:

[0216] The server analyzes the character information converted using a generative AI model. It receives string data as input and detects features that indicate potential fraud. The output is a fraud score or information about signs of fraud. A risk assessment is performed based on comparison with known fraud patterns.

[0217] Step 4:

[0218] Simultaneously, the server processes the audio signal itself into an acoustic analysis engine to analyze the user's emotional state. The input is a digital audio signal, and the output is data indicating the user's emotional state. By extracting emotional characteristics from the tone, tempo, and intonation of the voice, the server determines the user's psychological situation.

[0219] Step 5:

[0220] The server generates and sends alert information to a terminal or registered communication device if it detects potential fraud or a specific emotional state. It uses a fraud score and emotional state as input and outputs an alert message. This provides immediate alerts to users and registered recipients.

[0221] Step 6:

[0222] Users take appropriate action based on the warning information they receive. Input is a warning message, and output can include specific actions or notifications to family members. Based on the warning content, support for fraud avoidance and psychological well-being is provided quickly.

[0223] The specific processing unit 290 transmits the result of the specific processing to the smart device 14. In the smart device 14, the control unit 46A causes the output device 40 to output the result of the specific processing. The microphone 38B acquires audio indicating user input for the result of the specific processing. The control unit 46A transmits the audio data indicating user input acquired by the microphone 38B to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the audio data.

[0224] Data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of data generation model 58 is ChatGPT (registered trademark) (Internet search).<URL: https: / / openai.com / blog / chatgpt> ), Gemini (registered trademark) (Internet search) <url: https: gemini.google.com ?hl="ja">Examples of generative AI include those described above. The data generation model 58 is obtained by performing deep learning on a neural network. The data generation model 58 is input with prompts containing instructions, and with inference data such as audio data representing speech, text data representing text, and image data representing images. The data generation model 58 infers from the input inference data according to the instructions shown by the prompts, and outputs the inference results in data formats such as audio data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[0225] In the above embodiment, an example was given in which specific processing is performed by the data processing device 12, but the technology of this disclosure is not limited thereto, and the specific processing may also be performed by the smart device 14.

[0226] [Second Embodiment]

[0227] Figure 3 shows an example of the configuration of the data processing system 210 according to the second embodiment.

[0228] As shown in Figure 3, the data processing system 210 includes a data processing device 12 and smart glasses 214. An example of the data processing device 12 is a server.

[0229] The data processing device 12 comprises a computer 22, a database 24, and a communication interface 26. The computer 22 is an example of a "computer" related to the technology of this disclosure. The computer 22 comprises a processor 28, RAM 30, and storage 32. The processor 28, RAM 30, and storage 32 are connected to a bus 34. The database 24 and the communication interface 26 are also connected to the bus 34. The communication interface 26 is connected to a network 54. An example of the network 54 is a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0230] The smart glasses 214 include a computer 36, a microphone 238, a speaker 240, a camera 42, and a communication interface 44. The computer 36 includes a processor 46, RAM 48, and storage 50. The processor 46, RAM 48, and storage 50 are connected to a bus 52. The microphone 238, speaker 240, and camera 42 are also connected to the bus 52.

[0231] The microphone 238 receives voice signals from the user 20 and receives instructions from the user 20. The microphone 238 captures the voice signals from the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio according to the instructions from the processor 46.

[0232] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an image sensor such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the area around the user 20 (for example, an imaging range defined by a field of view equivalent to the width of a typical healthy person's field of vision).

[0233] Communication interface 44 is connected to network 54. Communication interfaces 44 and 26 are responsible for the exchange of various information between processor 46 and processor 28 via network 54. The exchange of various information between processor 46 and processor 28 using communication interfaces 44 and 26 is performed in a secure manner.

[0234] Figure 4 shows an example of the main functions of the data processing device 12 and the smart glasses 214. As shown in Figure 4, the data processing device 12 performs specific processing using the processor 28. The storage 32 stores the specific processing program 56.

[0235] The specific processing program 56 is an example of a "program" relating to the technology of this disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific processing is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific processing program 56 executed on the RAM 30.

[0236] The storage 32 stores the data generation model 58 and the emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[0237] In the smart glasses 214, the processor 46 performs the reception output processing. The storage 50 stores the reception output program 60. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output processing is realized by the processor 46 operating as a control unit 46A according to the reception output program 60 executed on the RAM 48.

[0238] Next, the identification processing performed by the identification processing unit 290 of the data processing device 12 will be described. In the following description, the data processing device 12 will be referred to as the "server" and the smart glasses 214 will be referred to as the "terminal".

[0239] This invention uses an application installed on a user's portable device to detect fraud in real time and issue warnings. Specifically, the device continuously acquires voice input signals and sends that data to a server for analysis.

[0240] First, the device records sounds occurring around the user through a high-sensitivity microphone. This audio input signal is compressed and encrypted within the device to ensure the protection of personal information, and then sent to the server periodically or in real time. The compression technology reduces data volume while maintaining sound quality.

[0241] Next, the server converts the received audio into text data. This utilizes speech recognition technology and achieves high accuracy through preprocessing such as noise reduction and speaker separation. This text data is further analyzed by a generative model to determine whether patterns and keywords related to fraud exist.

[0242] When a potential scam is detected, the server immediately generates a warning signal. This signal is sent to the user's device and to registered communication devices such as family members and the police. The warning signal includes a summary of the conversation and a description of the situation, designed to support a quick and appropriate response.

[0243] For example, a user might be asked for personal information over the phone under the pretext of a refund. In this case, the device records the conversation and sends the information to a server. The server analyzes the data and, if it detects any potentially dangerous phrases, immediately sends a warning to the family or the police. This allows the user and those around them to take swift and appropriate action.

[0244] Thus, the present invention provides a system that protects users, especially the elderly, from the threat of fraud by detecting fraud in real time through voice and issuing a rapid warning.

[0245] The following describes the processing flow.

[0246] Step 1:

[0247] The device constantly monitors ambient noise and records the user's voice. This voice input signal is buffered at specific time intervals.

[0248] Step 2:

[0249] The terminal converts the buffered audio into digital data and compresses it to improve communication efficiency. Then, it encrypts the data to ensure security.

[0250] Step 3:

[0251] The device sends compressed and encrypted audio digital data to the server at regular intervals. Transmission is done via Wi-Fi or a mobile data network.

[0252] Step 4:

[0253] The server decodes the received audio data, performs preprocessing such as noise reduction and speaker separation, and then converts it into text data using speech recognition technology.

[0254] Step 5:

[0255] The server uses a generative model to analyze text data and detect patterns and keywords that may indicate fraudulent activity.

[0256] Step 6:

[0257] If the server determines that a conversation may be fraudulent, it will generate a warning signal. This warning signal will include a summary of the relevant conversation.

[0258] Step 7:

[0259] The server sends a warning signal to the user's device and to pre-registered family members and police communication devices.

[0260] Step 8:

[0261] The device notifies the user using audio alerts and screen displays based on the received warning signals. The user can then take appropriate action.

[0262] (Example 1)

[0263] Next, we will describe Example 1. In the following description, the data processing device 12 will be referred to as the "server," and the smart glasses 214 will be referred to as the "terminal."

[0264] In modern society, fraud, particularly targeting the elderly, remains a persistent problem. Sophisticated scams conducted over the phone or in person are increasingly resulting in the illegal acquisition of personal information and money, necessitating swift and effective countermeasures. Traditional prevention measures struggle to detect fraud in real time, often leading to awareness only after the fact. Therefore, a system that warns of potential fraud in advance is needed.

[0265] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 1 is realized by the following means.

[0266] In this invention, the server includes a device for acquiring audio, a device for converting the audio into digital data, and a device for compressing and encrypting the digital data and transmitting it over a communication channel. This makes it possible to detect fraudulent activity in real time and to securely transmit relevant information.

[0267] "Sound" refers to sounds that are generated as vibrations in the air and perceived by human hearing.

[0268] "Digital data" refers to data obtained by converting analog information, such as audio, into numerical data that can be processed by electronic devices.

[0269] "Compression" is a process that reduces the size of data, which enables more efficient data storage and transfer.

[0270] "Encryption" refers to the process of transforming data using a specific algorithm to prevent it from being deciphered by a third party.

[0271] "Communication channel" refers to a path or means used for data transmission, including wired or wireless networks.

[0272] "Generative technology" refers to technologies that enable artificial intelligence to analyze data according to specific purposes and generate new value or information.

[0273] A "warning signal" is alert information issued when specific conditions are met, with the purpose of prompting the recipient to take some kind of action.

[0274] "Designated device" refers to a device that has been pre-registered or configured for receiving data.

[0275] "Communication equipment" refers to devices used for communication, including mobile phones and computers.

[0276] This invention is a system that uses a user-carried terminal and a server to detect fraud in real time and issue warnings. Specifically, the user-carried terminal is equipped with a high-sensitivity microphone that constantly captures ambient sounds. This sound is converted into digital data, and then compressed (e.g., using the FLAC codec) and encrypted (AES encryption) technologies are used to reduce the data size and ensure security. The data is then transmitted to the server via a communication channel. The SSL / TLS protocol is used to ensure the security of the data during transmission.

[0277] The server decrypts the received data and removes the compression to restore the original audio data. Then, speech recognition technology (e.g., a speech recognition API) is used to convert the audio data into text data. This text data is then analyzed using generative techniques to detect patterns related to fraud. For the generative techniques, widely used generative AI models are employed.

[0278] If the server determines that a fraudulent activity is likely, it generates a warning signal and sends the warning to the user's terminal or a pre-registered communication device. The warning includes a summary of the relevant conversation and a description of the situation, enabling a quick response.

[0279] As a specific example, consider the case where a user is asked for personal information over the phone on the grounds of a refund. When such a conversation takes place, the terminal records the voice and waits for analysis by the server. If the server detects a pattern that includes both "refund" and "bank information" simultaneously, the generated warning will include an explanation of "possibility of a suspicious refund request" and will be instantly displayed on the user terminal.

[0280] As an example of a prompt text, the content "Please build a system that records conversations where there is a possibility of fraud by the user and immediately sends a warning when danger is detected." can be considered.

[0281] With this system, it becomes possible to detect fraud in real time and prevent damage.

[0282] The flow of the specific process in Example 1 will be described using FIG. 11.

[0283] Step 1:

[0284] The terminal acquires the surrounding voice with a high-sensitivity microphone. This voice data is acquired in analog form and converted into digital data. Specifically, an ADC (Analog-to-Digital Converter) is used to convert the analog signal into a digital signal. The input is the raw voice around the user, and the output is the digitized voice data.

[0285] Step 2:

[0286] The terminal compresses and encrypts the acquired digital voice data. The data is compressed with the FLAC codec to reduce the data size.Next, the AES encryption technology is used to ensure security. As a result, the input is digital voice data, and the output is compressed and encrypted voice data.

[0287] Step   3:

[0288] The terminal sends compressed and encrypted audio data to the server. The SSL / TLS protocol is used to ensure secure data transmission. The input is compressed and encrypted audio data, and the output is the received audio data stored on the server.

[0289] Step 4:

[0290] The server decrypts the received audio data, removes the compression, and obtains the original digital audio data. First, it decrypts the data using AES, and then removes the FLAC compression. The input is encrypted audio data, and the output is the restored digital audio data.

[0291] Step 5:

[0292] The server converts the restored digital audio data into text data using speech recognition technology. It uses a specific speech recognition API for the audio-to-text conversion. The input is the restored digital audio data, and the output is the generated text data.

[0293] Step 6:

[0294] The server uses a generative AI model to analyze text data and detect patterns and keywords related to fraud. The input is text data, and the output is an analysis result indicating the possibility of fraud. The generative AI model performs analysis based on a pre-trained database.

[0295] Step 7:

[0296] If the server detects a potential scam, it generates a warning signal and sends it to the user's terminal or registered communication device as needed. The warning signal includes a summary of the relevant conversation. The input is the analysis result regarding the potential scam, and the output is the generated warning signal.

[0297] In this way, the fraud detection system functions through the coordination of each step.

[0298] (Application Example 1)

[0299] Next, we will explain Application Example 1. In the following explanation, the data processing device 12 will be referred to as the "server," and the smart glasses 214 will be referred to as the "terminal."

[0300] Traditionally, fraud detection systems have struggled to quickly and accurately detect fraud conducted via telephone and other communication methods. Especially with the increasing number of fraud cases targeting the elderly, real-time fraud detection and warnings are in high demand. Current systems have limitations in analysis accuracy and warning speed, failing to adequately protect user safety.

[0301] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 1 is realized by the following means.

[0302] In this invention, the server includes a device for acquiring an audio input signal, a device for converting the audio input signal into a text representation, and a device for analyzing the text representation using a generation model and detecting patterns related to fraud. This enables the rapid generation of a warning signal and notification including safety information to registered information processing devices when the possibility of fraud is determined.

[0303] "Audio input signal" refers to a signal collected using a high-sensitivity microphone from sounds occurring around the user.

[0304] "Text representation" refers to character data obtained by converting an audio input signal using speech recognition technology.

[0305] A "generative model" is an algorithm that uses machine learning to analyze text data and detect patterns and keywords related to fraud.

[0306] A "warning signal" is a signal that is generated when a potential fraud is detected, and it sends a warning to the user and registered information processing device.

[0307] An "information processing device" is a communicable device owned by a user or a registered related person.

[0308] "Encryption" is a process for preventing the content of data from being known to a third party.

[0309] A "communication line" is a network path for transmitting and receiving data between a server and a client.

[0310] An "external processing device" is a system that functions as a central server and performs analysis of voice data and notification management.

[0311] "Analysis results using a generative AI model" are information about the possibility of fraud based on analysis using text data converted from voice.

[0312] To implement this invention, an information processing device that a user normally uses, such as a smartphone, is required. A high-sensitivity microphone is attached to the terminal, and this microphone is used to collect the voices around the user. The voice input signal is encrypted and compressed within the terminal and transmitted to a server, which is an external processing device, in real time or periodically. For communication, an encrypted communication network (e.g., HTTPS protocol) is used to ensure security.

[0313] The server converts the received voice data into text representation using the Google Speech-to-Text API. This text is analyzed using the GPT-3, which is a generative AI model, to detect the presence or absence of patterns related to fraud. The analysis using the generative AI model also includes preprocessing such as noise removal and speaker separation.

[0314] When the possibility of fraud is detected, the server generates a warning signal and transmits it to the terminal or the registered information processing device. There is also a means to transmit the warning signal as an SMS or an app notification using the Twilio API. In this case, the warning signal includes summary data of the corresponding conversation and analysis results using the generative AI model, prompting the user to take prompt and appropriate action.

[0315] As a concrete example, suppose a user receives a phone call requesting personal information in connection with a refund. In this case, the device records the conversation and sends it to a server. The server detects dangerous phrases such as "refund" and "personal information" and immediately sends a warning to the user and their registered family members, thereby ensuring the safety of the user and those around them.

[0316] An example of a prompt message might be: "Please determine if the following text contains potential fraud: 'You are eligible for a refund, please provide your bank information.' If it does, please also explain the factors that indicate it is fraudulent."

[0317] The flow of a specific process in Application Example 1 will be explained using Figure 12.

[0318] Step 1:

[0319] The device collects audio from the user's surroundings through a high-sensitivity microphone. The input is a real-time audio signal, which is then encrypted and compressed within the device. The output is compressed audio data in a format that can be transmitted quickly.

[0320] Step 2:

[0321] The terminal sends compressed and encrypted audio data to a server, which is an external processing unit, via a secure communication network. The input is compressed audio data, which is securely transferred to the server as output. Specifically, encrypted packets are created and sent using the HTTPS protocol.

[0322] Step 3:

[0323] The server uses the Google Speech-to-Text API to convert received audio data into text. The input is compressed audio data, which is then denoised and speaker-separated before generating plain text data as output. The server then passes this data to a generation AI model.

[0324] Step 4:

[0325] The server uses a generative AI model, specifically GPT-3, to analyze text data. The input is transformed text data, utilizing prompts to detect patterns and keywords related to fraud. The output is the analysis result, including data on the likelihood of fraud and specific risk factors.

[0326] Step 5:

[0327] If the analysis reveals a potential for fraud, the server generates a warning signal. The input is the analysis result, and the output is the warning signal based on this result. Specifically, it generates an alert message containing the problematic phrase or analysis result.

[0328] Step 6:

[0329] The server sends warning signals generated via the Twilio API to users and registered information processing devices. The input is the warning signal, and the output is a notification or SMS sent to the user's device or registered recipient. Specifically, a notification is immediately displayed on the user's smartphone.

[0330] Furthermore, an emotion engine that estimates the user's emotions may be incorporated. That is, the identification processing unit 290 may use the emotion identification model 59 to estimate the user's emotions and perform identification processing using the user's emotions.

[0331] This invention provides a system that analyzes voice input signals obtained from a user to detect the possibility of fraudulent activity while simultaneously recognizing the user's emotional state. In particular, by considering the user's psychological response in addition to the content of the voice, the reliability of fraud detection is enhanced.

[0332] The terminal acquires an audio input signal by recording the user's voice. The acquired audio data is compressed and encrypted on the terminal and sent to the server as needed. The server decrypts the received audio data and begins processing.

[0333] The server first uses speech recognition technology to convert the audio data into text data. This text data is then analyzed by a generative model designed to detect signs of fraud. The model compares it to existing fraud patterns to find matches.

[0334] Furthermore, the server uses an emotion engine to analyze the user's voice to determine their emotions. This emotion engine determines whether the user is experiencing psychological states such as anxiety or confusion, based on acoustic characteristics such as tone, tempo, and intonation.

[0335] If fraudulent activity is detected or the user is found to be in an unusual emotional state, the server generates a warning signal. This warning signal is sent to the user's device and to registered family members and police communications devices. The warning signal includes details about the detected fraud pattern and a summary of the user's emotional state.

[0336] To give a concrete example, suppose a user is talking about a "risky investment opportunity" over the phone. The device records this conversation and sends it to a server. The server converts it to text, and if the emotion engine detects signs of anxiety in the user's voice, it immediately generates a warning signal. This warning signal is then sent to the user's family, allowing them to take prompt action.

[0337] Thus, by combining voice data and emotional information, the present invention realizes a comprehensive system that not only protects users from fraud but also provides psychological support.

[0338] The following describes the processing flow.

[0339] Step 1:

[0340] The device constantly monitors the audio surrounding the user and starts recording. The audio input signal is acquired through a high-sensitivity microphone.

[0341] Step 2:

[0342] The device compresses recorded audio input signals in real time and encrypts them to ensure security. This achieves both communication efficiency and information protection.

[0343] Step 3:

[0344] The device transmits compressed and encrypted audio data to the server at regular intervals. This transmission uses either Wi-Fi or mobile data, depending on network conditions.

[0345] Step 4:

[0346] The server decodes the received audio data and converts it into text data through a speech recognition system. At this stage, noise reduction is applied to obtain clear text information.

[0347] Step 5:

[0348] The server feeds the text data converted from the speech into a generative model, which analyzes patterns and keywords that suggest potential fraud.

[0349] Step 6:

[0350] The server simultaneously uses an emotion engine to evaluate the user's emotional state based on the tone and tempo of their voice. This can detect unusual mental states such as anxiety or tension.

[0351] Step 7:

[0352] The server generates a warning signal if it detects signs of fraud or unusual emotional states. This warning signal includes information about the analyzed fraud patterns and the user's emotional state.

[0353] Step 8:

[0354] The server sends a warning signal to the user's device and the communication devices of registered family members or the police. This allows for real-time sharing of the situation.

[0355] Step 9:

[0356] The device notifies the user using audio and visual feedback based on the received warning signal. This allows the user to become more vigilant against fraud and to coordinate with family members.

[0357] (Example 2)

[0358] Next, we will describe Example 2. In the following description, the data processing device 12 will be referred to as the "server" and the smart glasses 214 will be referred to as the "terminal".

[0359] Conventional acoustic signal analysis systems aimed at detecting fraudulent activity, but they did not adequately consider the user's psychological state. Therefore, there is a need for a method that can quickly grasp emotions and psychological changes to enable more appropriate warnings and interventions.

[0360] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 2 is realized by the following means.

[0361] In this invention, the server includes means for converting acoustic signals into document data, means for analyzing the document data using a generative model to detect patterns related to fraud, and means for analyzing acoustic characteristics to determine the user's psychological state. This makes it possible to detect both signs of fraudulent activity and the user's psychological state, enabling faster and more appropriate warnings and responses.

[0362] An "acoustic signal" is a representation of sound as an electrical signal, which is acquired by devices such as microphones.

[0363] "Document data" refers to data in text format that has been converted from acoustic signals using speech recognition technology.

[0364] A "generative model" is a statistical model used to analyze data and detect specific patterns using machine learning and artificial intelligence techniques.

[0365] "Acoustic characteristics" refer to physical properties such as tone, tempo, and intonation that are extracted from audio data.

[0366] "Psychological state" refers to information that indicates the user's emotions and mental condition, analyzed based on the tone, tempo, and intonation of their voice.

[0367] A "warning signal" is a signal that is generated and transmitted to draw attention when the possibility of fraud or an unusual psychological state is detected.

[0368] "Communication equipment" refers to devices used to send and receive data and signals, and includes, for example, smartphones and tablets.

[0369] An "information transmission device" is a device that has the function of transmitting digital data or signals to other devices.

[0370] This invention is a system that uses acoustic signals to detect fraudulent activity and analyze the user's psychological state. The system converts acoustic signals into document data and uses that data to detect fraud patterns and analyze emotions.

[0371] The terminal acquires audio through the user's microphone. The acquired audio signal is compressed and encrypted locally and sent to the server over the communication network. A common encryption library is used for the encryption process.

[0372] The server decodes the received acoustic signal and first converts it into document data using advanced speech recognition software. Speech recognition technologies such as the Google Speech-to-Text API can be applied at this stage. Next, this document data is analyzed by a generative AI model. This generative AI model is trained on an extensive dataset of fraud cases and detects whether the text data contains signs of fraud.

[0373] Furthermore, the server is equipped with an emotion engine that determines psychological state based on acoustic characteristics. This emotion engine analyzes the tone, tempo, and intonation of the voice to evaluate whether the user is in a specific psychological state, such as anxiety or confusion.

[0374] For example, if a user is talking on the phone about a "business opportunity to make a quick profit," the device records the conversation and sends it to a server. The server converts the audio to text and uses an emotion engine to detect signs of anxiety. If this matches signs of fraud, the server immediately generates a warning signal and sends it to the user's relatives or other relevant parties.

[0375] Examples of prompts to input into a generative AI model:

[0376] Design a system that determines a user's psychological state from their acoustic signals and detects the possibility of fraudulent activity. The input should be acoustic signal data, and the output should be the degree of matching with fraud patterns and the results of the psychological state analysis.

[0377] Therefore, the system can comprehensively address fraudulent activities and provide psychological support.

[0378] The flow of the specific processing in Example 2 will be explained using Figure 13.

[0379] Step 1:

[0380] The device acquires the user's audio signal through a microphone. The input is the user's raw voice, and the output is an audio signal in digital format. In this step, the audio is captured in real time and data conversion processing is performed. The acquired audio signal is compressed and encrypted on the spot to ensure security.

[0381] Step 2:

[0382] The terminal transmits an encrypted acoustic signal to the server using a communication network. The input is the encrypted acoustic signal, and the output is the data transmitted to the server. In this step, the data travels to the server via a network such as the internet.

[0383] Step 3:

[0384] The server decrypts the received acoustic signal to prepare the data. The input is an encrypted acoustic signal, and the output is the available acoustic signal. Decryption is performed using a common encryption technique.

[0385] Step 4:

[0386] The server uses advanced speech recognition technology to convert acoustic signals into document data. The input is the decoded acoustic signal, and the output is text data. This process is performed using a speech recognition engine such as the Google Speech-to-Text API.

[0387] Step 5:

[0388] The server analyzes document data using a generative AI model. The input is text data generated by speech recognition, and the output is an analysis of the likelihood of fraud. The model scans the text and looks for parts that match known fraud patterns.

[0389] Step 6:

[0390] The server operates an emotion engine that determines psychological state based on acoustic characteristics. The input is an acoustic signal, and the output is data indicating the user's emotional state. This analysis is performed based on factors such as the tone, tempo, and intonation of the voice.

[0391] Step 7:

[0392] The server generates an alert signal if it detects signs of fraud or unusual psychological states. The input is fraud detection results and psychological state data, and the output is the alert signal. This signal includes details about the fraud pattern and psychological state.

[0393] Step 8:

[0394] The server sends the generated warning signal to a pre-registered communication device. The input is the warning signal, and the output is the information sent to the communication device. The warning signal reaches the user's relatives and relevant organizations quickly, enabling a rapid response.

[0395] (Application Example 2)

[0396] Next, we will explain application example 2. In the following explanation, the data processing device 12 will be referred to as the "server," and the smart glasses 214 will be referred to as the "terminal."

[0397] In modern society, fraud via voice calls is on the rise, making countermeasures an urgent necessity. Furthermore, victims often experience psychological distress and confusion during fraudulent activities, highlighting the need for support. However, current technology is insufficient to simultaneously detect voice-based fraud and analyze emotional states. Therefore, the challenge lies in rapidly detecting potential fraudulent activity and providing information that takes into account the user's emotional state.

[0398] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means.

[0399] In this invention, the server includes a medium for acquiring audio signals, a medium for converting audio signals into corresponding text information, a medium for analyzing the text information using a generation model to identify patterns related to fraud, and a medium for analyzing the user's emotional state based on the audio signals and generating psychological support information if the possibility of fraud or a specific emotional state is recognized. This enables early detection of fraudulent activity and the provision of appropriate support information based on the user's emotional state.

[0400] A "medium for acquiring audio signals" is a device or system that receives audio signals input from a user and records them for processing.

[0401] A "medium for converting into text information" is a device or system that analyzes acquired audio signals and expresses their content as text data.

[0402] "Using generative models" refers to a method of detecting specific patterns or features based on speech or text data, utilizing machine learning and AI technologies.

[0403] A "medium for identifying patterns related to fraud" is a device or mechanism for identifying potential fraud from audio or text based on the characteristics of known fraudulent activities.

[0404] A "medium for analyzing a user's emotional state" is a device or mechanism that analyzes information such as tone, rhythm, and intonation of an audio signal to identify the user's emotions.

[0405] A "medium for generating psychological support information" is a device or mechanism that generates information to provide reassurance or warnings based on the user's emotional state and the possibility of fraud.

[0406] "Warning information" refers to information used to alert those involved when potential fraud or specific emotional states are detected.

[0407] "A medium for transmitting information to a registered device or communication device" refers to a device or mechanism for transmitting generated warning information or support information to a communication device used by a user or related party.

[0408] This invention is a system that converts audio signals into text information using a speech recognition system, analyzes that text information using a generation AI model to detect the possibility of fraud, and further determines the user's emotional state from the audio signal. To enable this, the server and terminal must coordinate their operations.

[0409] The server collects audio data through hardware designed to acquire audio signals (e.g., audio input devices such as smartphones). The acquired audio signals are converted into text information using speech recognition technology on the server. This speech recognition uses common speech recognition libraries or cloud-based APIs (e.g., Google Speech-to-Text API).

[0410] Next, the server uses a generative AI model (e.g., a machine learning model) to analyze the converted text information. This analysis identifies patterns related to fraud. By utilizing a trained model to compare with known fraud patterns, advanced fraud detection is achieved.

[0411] Meanwhile, the server analyzes the user's emotional state from the audio signal. This uses an emotion analysis engine based on acoustic characteristics such as tone, tempo, and intonation. This allows the server to determine whether the user is experiencing emotional states such as anxiety or confusion.

[0412] If potential fraud or a specific emotional state is detected, the server generates a warning message and transmits it to the terminal or communication device. This allows users and registered parties to be aware of the risk of fraud or if the user is in need of psychological support.

[0413] For example, if a user receives a suspicious financial call, the sentiment analysis engine might detect anxiety in the user's voice. In this case, the system would immediately generate a warning and notify the user's family and registered contacts, enabling early intervention.

[0414] As an example of a prompt message to a generative AI model, it could be something like, "Detect if this call is a scam. Text: 'There is a new investment opportunity...'" This allows the model to analyze the likelihood of fraud based on text information.

[0415] The flow of a specific process in Application Example 2 will be explained using Figure 14.

[0416] Step 1:

[0417] The terminal uses a microphone to acquire the user's voice signal. It uses the raw voice signal as input, converts it to a digital format, and sends it to the server. This allows for the preparation of a suitable data format for preprocessing of speech recognition.

[0418] Step 2:

[0419] The server converts the acquired audio signal into text information using speech recognition software (e.g., Google Speech-to-Text API). The input is a digital audio signal, and the output is the corresponding string data. Text information is obtained by analyzing the acoustic characteristics of the audio signal and converting it into text.

[0420] Step 3:

[0421] The server analyzes the character information converted using a generative AI model. It receives string data as input and detects features that indicate potential fraud. The output is a fraud score or information about signs of fraud. A risk assessment is performed based on comparison with known fraud patterns.

[0422] Step 4:

[0423] Simultaneously, the server processes the audio signal itself into an acoustic analysis engine to analyze the user's emotional state. The input is a digital audio signal, and the output is data indicating the user's emotional state. By extracting emotional characteristics from the tone, tempo, and intonation of the voice, the server determines the user's psychological situation.

[0424] Step 5:

[0425] The server generates and sends alert information to a terminal or registered communication device if it detects potential fraud or a specific emotional state. It uses a fraud score and emotional state as input and outputs an alert message. This provides immediate alerts to users and registered recipients.

[0426] Step 6:

[0427] Users take appropriate action based on the warning information they receive. Input is a warning message, and output can include specific actions or notifications to family members. Based on the warning content, support for fraud avoidance and psychological well-being is provided quickly.

[0428] The specific processing unit 290 transmits the result of the specific processing to the smart glasses 214. In the smart glasses 214, the control unit 46A causes the speaker 240 to output the result of the specific processing. The microphone 238 acquires audio indicating user input for the result of the specific processing. The control unit 46A transmits the audio data indicating user input acquired by the microphone 238 to the data processing unit 12. In the data processing unit 12, the specific processing unit 290 acquires the audio data.

[0429] Data generation model 58 is a type of so-called generative AI (Artificial Intelligence). One example of data generation model 58 is ChatGPT (Internet search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search) <url: https: gemini.google.com ?hl="ja">Examples of generative AI include those described above. The data generation model 58 is obtained by performing deep learning on a neural network. The data generation model 58 is input with prompts containing instructions, and with inference data such as audio data representing speech, text data representing text, and image data representing images. The data generation model 58 infers from the input inference data according to the instructions shown by the prompts, and outputs the inference results in data formats such as audio data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[0430] In the above embodiment, an example was given in which specific processing is performed by the data processing device 12, but the technology of this disclosure is not limited thereto, and the specific processing may also be performed by the smart glasses 214.

[0431] [Third Embodiment]

[0432] Figure 5 shows an example of the configuration of the data processing system 310 according to the third embodiment.

[0433] As shown in Figure 5, the data processing system 310 includes a data processing device 12 and a headset terminal 314. An example of the data processing device 12 is a server.

[0434] The data processing device 12 comprises a computer 22, a database 24, and a communication interface 26. The computer 22 is an example of a "computer" related to the technology of this disclosure. The computer 22 comprises a processor 28, RAM 30, and storage 32. The processor 28, RAM 30, and storage 32 are connected to a bus 34. The database 24 and the communication interface 26 are also connected to the bus 34. The communication interface 26 is connected to a network 54. An example of the network 54 is a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0435] The headset terminal 314 includes a computer 36, a microphone 238, a speaker 240, a camera 42, a communication interface 44, and a display 343. The computer 36 includes a processor 46, RAM 48, and storage 50. The processor 46, RAM 48, and storage 50 are connected to a bus 52. The microphone 238, speaker 240, camera 42, and display 343 are also connected to the bus 52.

[0436] The microphone 238 receives voice signals from the user 20 and receives instructions from the user 20. The microphone 238 captures the voice signals from the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio according to the instructions from the processor 46.

[0437] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an image sensor such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the area around the user 20 (for example, an imaging range defined by a field of view equivalent to the width of a typical healthy person's field of vision).

[0438] Communication interface 44 is connected to network 54. Communication interfaces 44 and 26 are responsible for the exchange of various information between processor 46 and processor 28 via network 54. The exchange of various information between processor 46 and processor 28 using communication interfaces 44 and 26 is performed in a secure manner.

[0439] Figure 6 shows an example of the main functions of the data processing device 12 and the headset terminal 314. As shown in Figure 6, the data processing device 12 performs specific processing using the processor 28. The storage 32 stores the specific processing program 56.

[0440] The specific processing program 56 is an example of a "program" relating to the technology of this disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific processing is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific processing program 56 executed on the RAM 30.

[0441] The storage 32 stores the data generation model 58 and the emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[0442] In the headset terminal 314, the processor 46 performs the reception output processing. The storage 50 stores the reception output program 60. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output processing is realized by the processor 46 operating as a control unit 46A according to the reception output program 60 executed on the RAM 48.

[0443] Next, the specific processing performed by the specific processing unit 290 of the data processing device 12 will be described. In the following description, the data processing device 12 will be referred to as the "server" and the headset terminal 314 will be referred to as the "terminal".

[0444] This invention uses an application installed on a user's portable device to detect fraud in real time and issue warnings. Specifically, the device continuously acquires voice input signals and sends that data to a server for analysis.

[0445] First, the device records sounds occurring around the user through a high-sensitivity microphone. This audio input signal is compressed and encrypted within the device to ensure the protection of personal information, and then sent to the server periodically or in real time. The compression technology reduces data volume while maintaining sound quality.

[0446] Next, the server converts the received audio into text data. This utilizes speech recognition technology and achieves high accuracy through preprocessing such as noise reduction and speaker separation. This text data is further analyzed by a generative model to determine whether patterns and keywords related to fraud exist.

[0447] When a potential scam is detected, the server immediately generates a warning signal. This signal is sent to the user's device and to registered communication devices such as family members and the police. The warning signal includes a summary of the conversation and a description of the situation, designed to support a quick and appropriate response.

[0448] For example, a user might be asked for personal information over the phone under the pretext of a refund. In this case, the device records the conversation and sends the information to a server. The server analyzes the data and, if it detects any potentially dangerous phrases, immediately sends a warning to the family or the police. This allows the user and those around them to take swift and appropriate action.

[0449] Thus, the present invention provides a system that protects users, especially the elderly, from the threat of fraud by detecting fraud in real time through voice and issuing a rapid warning.

[0450] The following describes the processing flow.

[0451] Step 1:

[0452] The device constantly monitors ambient noise and records the user's voice. This voice input signal is buffered at specific time intervals.

[0453] Step 2:

[0454] The terminal converts the buffered audio into digital data and compresses it to improve communication efficiency. Then, it encrypts the data to ensure security.

[0455] Step 3:

[0456] The device sends compressed and encrypted audio digital data to the server at regular intervals. Transmission is done via Wi-Fi or a mobile data network.

[0457] Step 4:

[0458] The server decodes the received audio data, performs preprocessing such as noise reduction and speaker separation, and then converts it into text data using speech recognition technology.

[0459] Step 5:

[0460] The server uses a generative model to analyze text data and detect patterns and keywords that may indicate fraudulent activity.

[0461] Step 6:

[0462] If the server determines that a conversation may be fraudulent, it will generate a warning signal. This warning signal will include a summary of the relevant conversation.

[0463] Step 7:

[0464] The server sends a warning signal to the user's device and to pre-registered family members and police communication devices.

[0465] Step 8:

[0466] The device notifies the user using audio alerts and screen displays based on the received warning signals. The user can then take appropriate action.

[0467] (Example 1)

[0468] Next, we will describe Example 1. In the following description, the data processing device 12 will be referred to as the "server," and the headset-type terminal 314 will be referred to as the "terminal."

[0469] In modern society, fraud, particularly targeting the elderly, remains a persistent problem. Sophisticated scams conducted over the phone or in person are increasingly resulting in the illegal acquisition of personal information and money, necessitating swift and effective countermeasures. Traditional prevention measures struggle to detect fraud in real time, often leading to awareness only after the fact. Therefore, a system that warns of potential fraud in advance is needed.

[0470] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 1 is realized by the following means.

[0471] In this invention, the server includes a device for acquiring audio, a device for converting the audio into digital data, and a device for compressing and encrypting the digital data and transmitting it over a communication channel. This makes it possible to detect fraudulent activity in real time and to securely transmit relevant information.

[0472] "Sound" refers to sounds that are generated as vibrations in the air and perceived by human hearing.

[0473] "Digital data" refers to data obtained by converting analog information, such as audio, into numerical data that can be processed by electronic devices.

[0474] "Compression" is a process that reduces the size of data, which enables more efficient data storage and transfer.

[0475] "Encryption" refers to the process of transforming data using a specific algorithm to prevent it from being deciphered by a third party.

[0476] "Communication channel" refers to a path or means used for data transmission, including wired or wireless networks.

[0477] "Generative technology" refers to technologies that enable artificial intelligence to analyze data according to specific purposes and generate new value or information.

[0478] A "warning signal" is alert information issued when specific conditions are met, with the purpose of prompting the recipient to take some kind of action.

[0479] "Designated device" refers to a device that has been pre-registered or configured for receiving data.

[0480] "Communication equipment" refers to devices used for communication, including mobile phones and computers.

[0481] This invention is a system that uses a user-carried terminal and a server to detect fraud in real time and issue warnings. Specifically, the user-carried terminal is equipped with a high-sensitivity microphone that constantly captures ambient sounds. This sound is converted into digital data, and then compressed (e.g., using the FLAC codec) and encrypted (AES encryption) technologies are used to reduce the data size and ensure security. The data is then transmitted to the server via a communication channel. The SSL / TLS protocol is used to ensure the security of the data during transmission.

[0482] The server decrypts the received data and removes the compression to restore the original audio data. Then, speech recognition technology (e.g., a speech recognition API) is used to convert the audio data into text data. This text data is then analyzed using generative techniques to detect patterns related to fraud. For the generative techniques, widely used generative AI models are employed.

[0483] If the server determines that a fraudulent activity is likely, it generates a warning signal and sends the warning to the user's terminal or a pre-registered communication device. The warning includes a summary of the relevant conversation and a description of the situation, enabling a quick response.

[0484] As a concrete example, consider a case where a user is asked for personal information over the phone under the pretext of a refund. When such a conversation takes place, the device records the audio and waits for analysis on the server. If the server detects a pattern that includes both "refund" and "bank information," the generated warning will include an explanation such as "possible suspicious refund request" and will be instantly displayed on the user's device.

[0485] An example of a prompt message might be: "Build a system that records potentially fraudulent conversations with users and immediately sends a warning when danger is detected."

[0486] This system makes it possible to detect fraud in real time and prevent victims from becoming victims.

[0487] The flow of the specific processing in Example 1 will be explained using Figure 11.

[0488] Step 1:

[0489] The device acquires ambient sound using a high-sensitivity microphone. This audio data is acquired in analog format and converted into digital data. Specifically, an ADC (Analog-to-Digital Converter) is used to convert the analog signal to a digital signal. The input is the raw sound around the user, and the output is digitized audio data.

[0490] Step 2:

[0491] The terminal compresses and encrypts the acquired digital audio data. The data is compressed using the FLAC codec to reduce its size. Next, security is ensured using AES encryption technology. As a result, the input is digital audio data, and the output is compressed and encrypted audio data.

[0492] Step 3:

[0493] The terminal sends compressed and encrypted audio data to the server. The SSL / TLS protocol is used to ensure secure data transmission. The input is compressed and encrypted audio data, and the output is the received audio data stored on the server.

[0494] Step 4:

[0495] The server decrypts the received audio data, removes the compression, and obtains the original digital audio data. First, it decrypts the data using AES, and then removes the FLAC compression. The input is encrypted audio data, and the output is the restored digital audio data.

[0496] Step 5:

[0497] The server converts the restored digital audio data into text data using speech recognition technology. It uses a specific speech recognition API for the audio-to-text conversion. The input is the restored digital audio data, and the output is the generated text data.

[0498] Step 6:

[0499] The server uses a generative AI model to analyze text data and detect patterns and keywords related to fraud. The input is text data, and the output is an analysis result indicating the possibility of fraud. The generative AI model performs analysis based on a pre-trained database.

[0500] Step 7:

[0501] If the server detects a potential scam, it generates a warning signal and sends it to the user's terminal or registered communication device as needed. The warning signal includes a summary of the relevant conversation. The input is the analysis result regarding the potential scam, and the output is the generated warning signal.

[0502] In this way, the fraud detection system functions through the coordination of each step.

[0503] (Application Example 1)

[0504] Next, we will explain Application Example 1. In the following explanation, the data processing device 12 will be referred to as the "server," and the headset-type terminal 314 will be referred to as the "terminal."

[0505] Traditionally, fraud detection systems have struggled to quickly and accurately detect fraud conducted via telephone and other communication methods. Especially with the increasing number of fraud cases targeting the elderly, real-time fraud detection and warnings are in high demand. Current systems have limitations in analysis accuracy and warning speed, failing to adequately protect user safety.

[0506] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 1 is realized by the following means.

[0507] In this invention, the server includes a device for acquiring an audio input signal, a device for converting the audio input signal into a text representation, and a device for analyzing the text representation using a generation model and detecting patterns related to fraud. This enables the rapid generation of a warning signal and notification including safety information to registered information processing devices when the possibility of fraud is determined.

[0508] "Audio input signal" refers to a signal collected using a high-sensitivity microphone from sounds occurring around the user.

[0509] "Text representation" refers to character data obtained by converting an audio input signal using speech recognition technology.

[0510] A "generative model" is an algorithm that uses machine learning to analyze text data and detect patterns and keywords related to fraud.

[0511] A "warning signal" is a signal that is generated when a potential fraud is detected, and it sends a warning to the user and registered information processing device.

[0512] An "information processing device" is a communication-enabled device owned by a user or registered party.

[0513] "Encryption" is a process used to prevent the contents of data from being known to third parties.

[0514] A "communication line" is a network path used to send and receive data between a server and a client.

[0515] An "external processing unit" is a system that functions as a central server, performing voice data analysis and notification management.

[0516] "Analysis results using a generative AI model" refers to information about the possibility of fraud based on an analysis of text data converted from speech.

[0517] To implement this invention, a user-facing information processing device, such as a smartphone, is required. The device is equipped with a high-sensitivity microphone, which is used to collect sounds from the user's surroundings. The audio input signal is encrypted and compressed within the device and transmitted in real time or periodically to an external processing device, a server. For communication, an encrypted communication network (e.g., HTTPS protocol) is used to ensure security.

[0518] The server converts the received audio data into a text representation using the Google Speech-to-Text API. This text is then analyzed for fraud-related patterns using the generative AI model GPT-3. The generative AI model's analysis also includes preprocessing such as noise reduction and speaker separation.

[0519] If a potential scam is detected, the server generates a warning signal and sends it to the device or registered information processing device. Alternatively, the Twilio API can be used to send the warning signal as an SMS or app notification. In this case, the warning signal includes summary data of the relevant conversation and analysis results using a generative AI model, prompting the user to take prompt and appropriate action.

[0520] As a concrete example, suppose a user receives a phone call requesting personal information in connection with a refund. In this case, the device records the conversation and sends it to a server. The server detects dangerous phrases such as "refund" and "personal information" and immediately sends a warning to the user and their registered family members, thereby ensuring the safety of the user and those around them.

[0521] An example of a prompt message might be: "Please determine if the following text contains potential fraud: 'You are eligible for a refund, please provide your bank information.' If it does, please also explain the factors that indicate it is fraudulent."

[0522] The flow of a specific process in Application Example 1 will be explained using Figure 12.

[0523] Step 1:

[0524] The device collects audio from the user's surroundings through a high-sensitivity microphone. The input is a real-time audio signal, which is then encrypted and compressed within the device. The output is compressed audio data in a format that can be transmitted quickly.

[0525] Step 2:

[0526] The terminal sends compressed and encrypted audio data to a server, which is an external processing unit, via a secure communication network. The input is compressed audio data, which is securely transferred to the server as output. Specifically, encrypted packets are created and sent using the HTTPS protocol.

[0527] Step 3:

[0528] The server uses the Google Speech-to-Text API to convert received audio data into text. The input is compressed audio data, which is then denoised and speaker-separated before generating plain text data as output. The server then passes this data to a generation AI model.

[0529] Step 4:

[0530] The server uses a generative AI model, specifically GPT-3, to analyze text data. The input is transformed text data, utilizing prompts to detect patterns and keywords related to fraud. The output is the analysis result, including data on the likelihood of fraud and specific risk factors.

[0531] Step 5:

[0532] If the analysis reveals a potential for fraud, the server generates a warning signal. The input is the analysis result, and the output is the warning signal based on this result. Specifically, it generates an alert message containing the problematic phrase or analysis result.

[0533] Step 6:

[0534] The server sends warning signals generated via the Twilio API to users and registered information processing devices. The input is the warning signal, and the output is a notification or SMS sent to the user's device or registered recipient. Specifically, a notification is immediately displayed on the user's smartphone.

[0535] Furthermore, an emotion engine that estimates the user's emotions may be incorporated. That is, the identification processing unit 290 may use the emotion identification model 59 to estimate the user's emotions and perform identification processing using the user's emotions.

[0536] This invention provides a system that analyzes voice input signals obtained from a user to detect the possibility of fraudulent activity while simultaneously recognizing the user's emotional state. In particular, by considering the user's psychological response in addition to the content of the voice, the reliability of fraud detection is enhanced.

[0537] The terminal acquires an audio input signal by recording the user's voice. The acquired audio data is compressed and encrypted on the terminal and sent to the server as needed. The server decrypts the received audio data and begins processing.

[0538] The server first uses speech recognition technology to convert the audio data into text data. This text data is then analyzed by a generative model designed to detect signs of fraud. The model compares it to existing fraud patterns to find matches.

[0539] Furthermore, the server uses an emotion engine to analyze the user's voice to determine their emotions. This emotion engine determines whether the user is experiencing psychological states such as anxiety or confusion, based on acoustic characteristics such as tone, tempo, and intonation.

[0540] If fraudulent activity is detected or the user is found to be in an unusual emotional state, the server generates a warning signal. This warning signal is sent to the user's device and to registered family members and police communications devices. The warning signal includes details about the detected fraud pattern and a summary of the user's emotional state.

[0541] To give a concrete example, suppose a user is talking about a "risky investment opportunity" over the phone. The device records this conversation and sends it to a server. The server converts it to text, and if the emotion engine detects signs of anxiety in the user's voice, it immediately generates a warning signal. This warning signal is then sent to the user's family, allowing them to take prompt action.

[0542] Thus, by combining voice data and emotional information, the present invention realizes a comprehensive system that not only protects users from fraud but also provides psychological support.

[0543] The following describes the processing flow.

[0544] Step 1:

[0545] The device constantly monitors the audio surrounding the user and starts recording. The audio input signal is acquired through a high-sensitivity microphone.

[0546] Step 2:

[0547] The device compresses recorded audio input signals in real time and encrypts them to ensure security. This achieves both communication efficiency and information protection.

[0548] Step 3:

[0549] The device transmits compressed and encrypted audio data to the server at regular intervals. This transmission uses either Wi-Fi or mobile data, depending on network conditions.

[0550] Step 4:

[0551] The server decodes the received audio data and converts it into text data through a speech recognition system. At this stage, noise reduction is applied to obtain clear text information.

[0552] Step 5:

[0553] The server feeds the text data converted from the speech into a generative model, which analyzes patterns and keywords that suggest potential fraud.

[0554] Step 6:

[0555] The server simultaneously uses an emotion engine to evaluate the user's emotional state based on the tone and tempo of their voice. This can detect unusual mental states such as anxiety or tension.

[0556] Step 7:

[0557] The server generates a warning signal if it detects signs of fraud or unusual emotional states. This warning signal includes information about the analyzed fraud patterns and the user's emotional state.

[0558] Step 8:

[0559] The server sends a warning signal to the user's device and the communication devices of registered family members or the police. This allows for real-time sharing of the situation.

[0560] Step 9:

[0561] The device notifies the user using audio and visual feedback based on the received warning signal. This allows the user to become more vigilant against fraud and to coordinate with family members.

[0562] (Example 2)

[0563] Next, we will describe Example 2. In the following description, the data processing device 12 will be referred to as the "server," and the headset-type terminal 314 will be referred to as the "terminal."

[0564] Conventional acoustic signal analysis systems aimed at detecting fraudulent activity, but they did not adequately consider the user's psychological state. Therefore, there is a need for a method that can quickly grasp emotions and psychological changes to enable more appropriate warnings and interventions.

[0565] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 2 is realized by the following means.

[0566] In this invention, the server includes means for converting acoustic signals into document data, means for analyzing the document data using a generative model to detect patterns related to fraud, and means for analyzing acoustic characteristics to determine the user's psychological state. This makes it possible to detect both signs of fraudulent activity and the user's psychological state, enabling faster and more appropriate warnings and responses.

[0567] An "acoustic signal" is a representation of sound as an electrical signal, which is acquired by devices such as microphones.

[0568] "Document data" refers to data in text format that has been converted from acoustic signals using speech recognition technology.

[0569] A "generative model" is a statistical model used to analyze data and detect specific patterns using machine learning and artificial intelligence techniques.

[0570] "Acoustic characteristics" refer to physical properties such as tone, tempo, and intonation that are extracted from audio data.

[0571] "Psychological state" refers to information that indicates the user's emotions and mental condition, analyzed based on the tone, tempo, and intonation of their voice.

[0572] A "warning signal" is a signal that is generated and transmitted to draw attention when the possibility of fraud or an unusual psychological state is detected.

[0573] "Communication equipment" refers to devices used to send and receive data and signals, and includes, for example, smartphones and tablets.

[0574] An "information transmission device" is a device that has the function of transmitting digital data or signals to other devices.

[0575] This invention is a system that uses acoustic signals to detect fraudulent activity and analyze the user's psychological state. The system converts acoustic signals into document data and uses that data to detect fraud patterns and analyze emotions.

[0576] The terminal acquires audio through the user's microphone. The acquired audio signal is compressed and encrypted locally and sent to the server over the communication network. A common encryption library is used for the encryption process.

[0577] The server decodes the received acoustic signal and first converts it into document data using advanced speech recognition software. Speech recognition technologies such as the Google Speech-to-Text API can be applied at this stage. Next, this document data is analyzed by a generative AI model. This generative AI model is trained on an extensive dataset of fraud cases and detects whether the text data contains signs of fraud.

[0578] Furthermore, the server is equipped with an emotion engine that determines psychological state based on acoustic characteristics. This emotion engine analyzes the tone, tempo, and intonation of the voice to evaluate whether the user is in a specific psychological state, such as anxiety or confusion.

[0579] For example, if a user is talking on the phone about a "business opportunity to make a quick profit," the device records the conversation and sends it to a server. The server converts the audio to text and uses an emotion engine to detect signs of anxiety. If this matches signs of fraud, the server immediately generates a warning signal and sends it to the user's relatives or other relevant parties.

[0580] Examples of prompts to input into a generative AI model:

[0581] Design a system that determines a user's psychological state from their acoustic signals and detects the possibility of fraudulent activity. The input should be acoustic signal data, and the output should be the degree of matching with fraud patterns and the results of the psychological state analysis.

[0582] Therefore, the system can comprehensively address fraudulent activities and provide psychological support.

[0583] The flow of the specific processing in Example 2 will be explained using Figure 13.

[0584] Step 1:

[0585] The device acquires the user's audio signal through a microphone. The input is the user's raw voice, and the output is an audio signal in digital format. In this step, the audio is captured in real time and data conversion processing is performed. The acquired audio signal is compressed and encrypted on the spot to ensure security.

[0586] Step 2:

[0587] The terminal transmits an encrypted acoustic signal to the server using a communication network. The input is the encrypted acoustic signal, and the output is the data transmitted to the server. In this step, the data travels to the server via a network such as the internet.

[0588] Step 3:

[0589] The server decrypts the received acoustic signal to prepare the data. The input is an encrypted acoustic signal, and the output is the available acoustic signal. Decryption is performed using a common encryption technique.

[0590] Step 4:

[0591] The server uses advanced speech recognition technology to convert acoustic signals into document data. The input is the decoded acoustic signal, and the output is text data. This process is performed using a speech recognition engine such as the Google Speech-to-Text API.

[0592] Step 5:

[0593] The server analyzes document data using a generative AI model. The input is text data generated by speech recognition, and the output is an analysis of the likelihood of fraud. The model scans the text and looks for parts that match known fraud patterns.

[0594] Step 6:

[0595] The server operates an emotion engine that determines psychological state based on acoustic characteristics. The input is an acoustic signal, and the output is data indicating the user's emotional state. This analysis is performed based on factors such as the tone, tempo, and intonation of the voice.

[0596] Step 7:

[0597] The server generates an alert signal if it detects signs of fraud or unusual psychological states. The input is fraud detection results and psychological state data, and the output is the alert signal. This signal includes details about the fraud pattern and psychological state.

[0598] Step 8:

[0599] The server sends the generated warning signal to a pre-registered communication device. The input is the warning signal, and the output is the information sent to the communication device. The warning signal reaches the user's relatives and relevant organizations quickly, enabling a rapid response.

[0600] (Application Example 2)

[0601] Next, we will explain application example 2. In the following explanation, the data processing device 12 will be referred to as the "server," and the headset-type terminal 314 will be referred to as the "terminal."

[0602] In modern society, fraud via voice calls is on the rise, making countermeasures an urgent necessity. Furthermore, victims often experience psychological distress and confusion during fraudulent activities, highlighting the need for support. However, current technology is insufficient to simultaneously detect voice-based fraud and analyze emotional states. Therefore, the challenge lies in rapidly detecting potential fraudulent activity and providing information that takes into account the user's emotional state.

[0603] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means.

[0604] In this invention, the server includes a medium for acquiring audio signals, a medium for converting audio signals into corresponding text information, a medium for analyzing the text information using a generation model to identify patterns related to fraud, and a medium for analyzing the user's emotional state based on the audio signals and generating psychological support information if the possibility of fraud or a specific emotional state is recognized. This enables early detection of fraudulent activity and the provision of appropriate support information based on the user's emotional state.

[0605] A "medium for acquiring audio signals" is a device or system that receives audio signals input from a user and records them for processing.

[0606] A "medium for converting into text information" is a device or system that analyzes acquired audio signals and expresses their content as text data.

[0607] "Using generative models" refers to a method of detecting specific patterns or features based on speech or text data, utilizing machine learning and AI technologies.

[0608] A "medium for identifying patterns related to fraud" is a device or mechanism for identifying potential fraud from audio or text based on the characteristics of known fraudulent activities.

[0609] A "medium for analyzing a user's emotional state" is a device or mechanism that analyzes information such as tone, rhythm, and intonation of an audio signal to identify the user's emotions.

[0610] A "medium for generating psychological support information" is a device or mechanism that generates information to provide reassurance or warnings based on the user's emotional state and the possibility of fraud.

[0611] "Warning information" refers to information used to alert those involved when potential fraud or specific emotional states are detected.

[0612] "A medium for transmitting information to a registered device or communication device" refers to a device or mechanism for transmitting generated warning information or support information to a communication device used by a user or related party.

[0613] This invention is a system that converts audio signals into text information using a speech recognition system, analyzes that text information using a generation AI model to detect the possibility of fraud, and further determines the user's emotional state from the audio signal. To enable this, the server and terminal must coordinate their operations.

[0614] The server collects audio data through hardware designed to acquire audio signals (e.g., audio input devices such as smartphones). The acquired audio signals are converted into text information using speech recognition technology on the server. This speech recognition uses common speech recognition libraries or cloud-based APIs (e.g., Google Speech-to-Text API).

[0615] Next, the server uses a generative AI model (e.g., a machine learning model) to analyze the converted text information. This analysis identifies patterns related to fraud. By utilizing a trained model to compare with known fraud patterns, advanced fraud detection is achieved.

[0616] Meanwhile, the server analyzes the user's emotional state from the audio signal. This uses an emotion analysis engine based on acoustic characteristics such as tone, tempo, and intonation. This allows the server to determine whether the user is experiencing emotional states such as anxiety or confusion.

[0617] If potential fraud or a specific emotional state is detected, the server generates a warning message and transmits it to the terminal or communication device. This allows users and registered parties to be aware of the risk of fraud or if the user is in need of psychological support.

[0618] For example, if a user receives a suspicious financial call, the sentiment analysis engine might detect anxiety in the user's voice. In this case, the system would immediately generate a warning and notify the user's family and registered contacts, enabling early intervention.

[0619] As an example of a prompt message to a generative AI model, it could be something like, "Detect if this call is a scam. Text: 'There is a new investment opportunity...'" This allows the model to analyze the likelihood of fraud based on text information.

[0620] The flow of a specific process in Application Example 2 will be explained using Figure 14.

[0621] Step 1:

[0622] The terminal uses a microphone to acquire the user's voice signal. It uses the raw voice signal as input, converts it to a digital format, and sends it to the server. This allows for the preparation of a suitable data format for preprocessing of speech recognition.

[0623] Step 2:

[0624] The server converts the acquired audio signal into text information using speech recognition software (e.g., Google Speech-to-Text API). The input is a digital audio signal, and the output is the corresponding string data. Text information is obtained by analyzing the acoustic characteristics of the audio signal and converting it into text.

[0625] Step 3:

[0626] The server analyzes the character information converted using a generative AI model. It receives string data as input and detects features that indicate potential fraud. The output is a fraud score or information about signs of fraud. A risk assessment is performed based on comparison with known fraud patterns.

[0627] Step 4:

[0628] Simultaneously, the server processes the audio signal itself into an acoustic analysis engine to analyze the user's emotional state. The input is a digital audio signal, and the output is data indicating the user's emotional state. By extracting emotional characteristics from the tone, tempo, and intonation of the voice, the server determines the user's psychological situation.

[0629] Step 5:

[0630] The server generates and sends alert information to a terminal or registered communication device if it detects potential fraud or a specific emotional state. It uses a fraud score and emotional state as input and outputs an alert message. This provides immediate alerts to users and registered recipients.

[0631] Step 6:

[0632] Users take appropriate action based on the warning information they receive. Input is a warning message, and output can include specific actions or notifications to family members. Based on the warning content, support for fraud avoidance and psychological well-being is provided quickly.

[0633] The specific processing unit 290 transmits the result of the specific processing to the headset terminal 314. In the headset terminal 314, the control unit 46A causes the speaker 240 and display 343 to output the result of the specific processing. The microphone 238 acquires audio indicating user input for the result of the specific processing. The control unit 46A transmits the audio data indicating user input acquired by the microphone 238 to the data processing unit 12. In the data processing unit 12, the specific processing unit 290 acquires the audio data.

[0634] Data generation model 58 is a type of so-called generative AI (Artificial Intelligence). One example of data generation model 58 is ChatGPT (Internet search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search) <url: https: gemini.google.com ?hl="ja">Examples of generative AI include those described above. The data generation model 58 is obtained by performing deep learning on a neural network. The data generation model 58 is input with prompts containing instructions, and with inference data such as audio data representing speech, text data representing text, and image data representing images. The data generation model 58 infers from the input inference data according to the instructions shown by the prompts, and outputs the inference results in data formats such as audio data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[0635] In the above embodiment, an example was given in which specific processing is performed by the data processing device 12, but the technology of this disclosure is not limited thereto, and specific processing may also be performed by the headset terminal 314.

[0636] [Fourth Embodiment]

[0637] Figure 7 shows an example of the configuration of the data processing system 410 according to the fourth embodiment.

[0638] As shown in Figure 7, the data processing system 410 includes a data processing device 12 and a robot 414. An example of the data processing device 12 is a server.

[0639] The data processing device 12 comprises a computer 22, a database 24, and a communication interface 26. The computer 22 is an example of a "computer" related to the technology of this disclosure. The computer 22 comprises a processor 28, RAM 30, and storage 32. The processor 28, RAM 30, and storage 32 are connected to a bus 34. The database 24 and the communication interface 26 are also connected to the bus 34. The communication interface 26 is connected to a network 54. An example of the network 54 is a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0640] The robot 414 includes a computer 36, a microphone 238, a speaker 240, a camera 42, a communication interface 44, and a controlled object 443. The computer 36 includes a processor 46, RAM 48, and storage 50. The processor 46, RAM 48, and storage 50 are connected to a bus 52. The microphone 238, speaker 240, camera 42, and controlled object 443 are also connected to the bus 52.

[0641] The microphone 238 receives voice signals from the user 20 and receives instructions from the user 20. The microphone 238 captures the voice signals from the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio according to the instructions from the processor 46.

[0642] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an image sensor such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the area around the user 20 (for example, an imaging range defined by a field of view equivalent to the width of a typical healthy person's field of vision).

[0643] Communication interface 44 is connected to network 54. Communication interfaces 44 and 26 are responsible for the exchange of various information between processor 46 and processor 28 via network 54. The exchange of various information between processor 46 and processor 28 using communication interfaces 44 and 26 is performed in a secure manner.

[0644] The controlled object 443 includes a display device, LEDs in the eyes, and motors that drive the arms, hands, and feet. The posture and gestures of the robot 414 are controlled by controlling the motors of the arms, hands, and feet. Some of the robot 414's emotions can be expressed by controlling these motors. Furthermore, the robot 414's facial expressions can also be expressed by controlling the illumination state of the LEDs in its eyes.

[0645] Figure 8 shows an example of the main functions of the data processing device 12 and the robot 414. As shown in Figure 8, the data processing device 12 performs specific processing using the processor 28. The storage 32 stores the specific processing program 56.

[0646] The specific processing program 56 is an example of a "program" relating to the technology of this disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific processing is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific processing program 56 executed on the RAM 30.

[0647] The storage 32 stores the data generation model 58 and the emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[0648] In robot 414, the processor 46 performs the reception output processing. The storage 50 stores the reception output program 60. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output processing is realized by the processor 46 operating as a control unit 46A according to the reception output program 60 executed on the RAM 48.

[0649] Next, the specific processing performed by the specific processing unit 290 of the data processing device 12 will be described. In the following description, the data processing device 12 will be referred to as the "server" and the robot 414 as the "terminal".

[0650] This invention uses an application installed on a user's portable device to detect fraud in real time and issue warnings. Specifically, the device continuously acquires voice input signals and sends that data to a server for analysis.

[0651] First, the device records sounds occurring around the user through a high-sensitivity microphone. This audio input signal is compressed and encrypted within the device to ensure the protection of personal information, and then sent to the server periodically or in real time. The compression technology reduces data volume while maintaining sound quality.

[0652] Next, the server converts the received audio into text data. This utilizes speech recognition technology and achieves high accuracy through preprocessing such as noise reduction and speaker separation. This text data is further analyzed by a generative model to determine whether patterns and keywords related to fraud exist.

[0653] When a potential scam is detected, the server immediately generates a warning signal. This signal is sent to the user's device and to registered communication devices such as family members and the police. The warning signal includes a summary of the conversation and a description of the situation, designed to support a quick and appropriate response.

[0654] For example, a user might be asked for personal information over the phone under the pretext of a refund. In this case, the device records the conversation and sends the information to a server. The server analyzes the data and, if it detects any potentially dangerous phrases, immediately sends a warning to the family or the police. This allows the user and those around them to take swift and appropriate action.

[0655] Thus, the present invention provides a system that protects users, especially the elderly, from the threat of fraud by detecting fraud in real time through voice and issuing a rapid warning.

[0656] The following describes the processing flow.

[0657] Step 1:

[0658] The device constantly monitors ambient noise and records the user's voice. This voice input signal is buffered at specific time intervals.

[0659] Step 2:

[0660] The terminal converts the buffered audio into digital data and compresses it to improve communication efficiency. Then, it encrypts the data to ensure security.

[0661] Step 3:

[0662] The device sends compressed and encrypted audio digital data to the server at regular intervals. Transmission is done via Wi-Fi or a mobile data network.

[0663] Step 4:

[0664] The server decodes the received audio data, performs preprocessing such as noise reduction and speaker separation, and then converts it into text data using speech recognition technology.

[0665] Step 5:

[0666] The server uses a generative model to analyze text data and detect patterns and keywords that may indicate fraudulent activity.

[0667] Step 6:

[0668] If the server determines that a conversation may be fraudulent, it will generate a warning signal. This warning signal will include a summary of the relevant conversation.

[0669] Step 7:

[0670] The server sends a warning signal to the user's device and to pre-registered family members and police communication devices.

[0671] Step 8:

[0672] The device notifies the user using audio alerts and screen displays based on the received warning signals. The user can then take appropriate action.

[0673] (Example 1)

[0674] Next, we will describe Example 1. In the following description, the data processing device 12 will be referred to as the "server" and the robot 414 as the "terminal".

[0675] In modern society, fraud, particularly targeting the elderly, remains a persistent problem. Sophisticated scams conducted over the phone or in person are increasingly resulting in the illegal acquisition of personal information and money, necessitating swift and effective countermeasures. Traditional prevention measures struggle to detect fraud in real time, often leading to awareness only after the fact. Therefore, a system that warns of potential fraud in advance is needed.

[0676] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 1 is realized by the following means.

[0677] In this invention, the server includes a device for acquiring audio, a device for converting the audio into digital data, and a device for compressing and encrypting the digital data and transmitting it over a communication channel. This makes it possible to detect fraudulent activity in real time and to securely transmit relevant information.

[0678] "Sound" refers to sounds that are generated as vibrations in the air and perceived by human hearing.

[0679] "Digital data" refers to data obtained by converting analog information, such as audio, into numerical data that can be processed by electronic devices.

[0680] "Compression" is a process that reduces the size of data, which enables more efficient data storage and transfer.

[0681] "Encryption" refers to the process of transforming data using a specific algorithm to prevent it from being deciphered by a third party.

[0682] "Communication channel" refers to a path or means used for data transmission, including wired or wireless networks.

[0683] "Generative technology" refers to technologies that enable artificial intelligence to analyze data according to specific purposes and generate new value or information.

[0684] A "warning signal" is alert information issued when specific conditions are met, with the purpose of prompting the recipient to take some kind of action.

[0685] "Designated device" refers to a device that has been pre-registered or configured for receiving data.

[0686] "Communication equipment" refers to devices used for communication, including mobile phones and computers.

[0687] This invention is a system that uses a user-carried terminal and a server to detect fraud in real time and issue warnings. Specifically, the user-carried terminal is equipped with a high-sensitivity microphone that constantly captures ambient sounds. This sound is converted into digital data, and then compressed (e.g., using the FLAC codec) and encrypted (AES encryption) technologies are used to reduce the data size and ensure security. The data is then transmitted to the server via a communication channel. The SSL / TLS protocol is used to ensure the security of the data during transmission.

[0688] The server decrypts the received data and removes the compression to restore the original audio data. Then, speech recognition technology (e.g., a speech recognition API) is used to convert the audio data into text data. This text data is then analyzed using generative techniques to detect patterns related to fraud. For the generative techniques, widely used generative AI models are employed.

[0689] If the server determines that a fraudulent activity is likely, it generates a warning signal and sends the warning to the user's terminal or a pre-registered communication device. The warning includes a summary of the relevant conversation and a description of the situation, enabling a quick response.

[0690] As a concrete example, consider a case where a user is asked for personal information over the phone under the pretext of a refund. When such a conversation takes place, the device records the audio and waits for analysis on the server. If the server detects a pattern that includes both "refund" and "bank information," the generated warning will include an explanation such as "possible suspicious refund request" and will be instantly displayed on the user's device.

[0691] An example of a prompt message might be: "Build a system that records potentially fraudulent conversations with users and immediately sends a warning when danger is detected."

[0692] This system makes it possible to detect fraud in real time and prevent victims from becoming victims.

[0693] The flow of the specific processing in Example 1 will be explained using Figure 11.

[0694] Step 1:

[0695] The device acquires ambient sound using a high-sensitivity microphone. This audio data is acquired in analog format and converted into digital data. Specifically, an ADC (Analog-to-Digital Converter) is used to convert the analog signal to a digital signal. The input is the raw sound around the user, and the output is digitized audio data.

[0696] Step 2:

[0697] The terminal compresses and encrypts the acquired digital audio data. The data is compressed using the FLAC codec to reduce its size. Next, security is ensured using AES encryption technology. As a result, the input is digital audio data, and the output is compressed and encrypted audio data.

[0698] Step 3:

[0699] The terminal sends compressed and encrypted audio data to the server. The SSL / TLS protocol is used to ensure secure data transmission. The input is compressed and encrypted audio data, and the output is the received audio data stored on the server.

[0700] Step 4:

[0701] The server decrypts the received audio data, removes the compression, and obtains the original digital audio data. First, it decrypts the data using AES, and then removes the FLAC compression. The input is encrypted audio data, and the output is the restored digital audio data.

[0702] Step 5:

[0703] The server converts the restored digital audio data into text data using speech recognition technology. It uses a specific speech recognition API for the audio-to-text conversion. The input is the restored digital audio data, and the output is the generated text data.

[0704] Step 6:

[0705] The server uses a generative AI model to analyze text data and detect patterns and keywords related to fraud. The input is text data, and the output is an analysis result indicating the possibility of fraud. The generative AI model performs analysis based on a pre-trained database.

[0706] Step 7:

[0707] If the server detects a potential scam, it generates a warning signal and sends it to the user's terminal or registered communication device as needed. The warning signal includes a summary of the relevant conversation. The input is the analysis result regarding the potential scam, and the output is the generated warning signal.

[0708] In this way, the fraud detection system functions through the coordination of each step.

[0709] (Application Example 1)

[0710] Next, we will explain Application Example 1. In the following explanation, the data processing device 12 will be referred to as the "server" and the robot 414 as the "terminal".

[0711] Traditionally, fraud detection systems have struggled to quickly and accurately detect fraud conducted via telephone and other communication methods. Especially with the increasing number of fraud cases targeting the elderly, real-time fraud detection and warnings are in high demand. Current systems have limitations in analysis accuracy and warning speed, failing to adequately protect user safety.

[0712] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 1 is realized by the following means.

[0713] In this invention, the server includes a device for acquiring an audio input signal, a device for converting the audio input signal into a text representation, and a device for analyzing the text representation using a generation model and detecting patterns related to fraud. This enables the rapid generation of a warning signal and notification including safety information to registered information processing devices when the possibility of fraud is determined.

[0714] "Audio input signal" refers to a signal collected using a high-sensitivity microphone from sounds occurring around the user.

[0715] "Text representation" refers to character data obtained by converting an audio input signal using speech recognition technology.

[0716] A "generative model" is an algorithm that uses machine learning to analyze text data and detect patterns and keywords related to fraud.

[0717] A "warning signal" is a signal that is generated when a potential fraud is detected, and it sends a warning to the user and registered information processing device.

[0718] An "information processing device" is a communication-enabled device owned by a user or registered party.

[0719] "Encryption" is a process used to prevent the contents of data from being known to third parties.

[0720] A "communication line" is a network path used to send and receive data between a server and a client.

[0721] An "external processing unit" is a system that functions as a central server, performing voice data analysis and notification management.

[0722] "Analysis results using a generative AI model" refers to information about the possibility of fraud based on an analysis of text data converted from speech.

[0723] To implement this invention, a user-facing information processing device, such as a smartphone, is required. The device is equipped with a high-sensitivity microphone, which is used to collect sounds from the user's surroundings. The audio input signal is encrypted and compressed within the device and transmitted in real time or periodically to an external processing device, a server. For communication, an encrypted communication network (e.g., HTTPS protocol) is used to ensure security.

[0724] The server converts the received audio data into a text representation using the Google Speech-to-Text API. This text is then analyzed for fraud-related patterns using the generative AI model GPT-3. The generative AI model's analysis also includes preprocessing such as noise reduction and speaker separation.

[0725] If a potential scam is detected, the server generates a warning signal and sends it to the device or registered information processing device. Alternatively, the Twilio API can be used to send the warning signal as an SMS or app notification. In this case, the warning signal includes summary data of the relevant conversation and analysis results using a generative AI model, prompting the user to take prompt and appropriate action.

[0726] As a concrete example, suppose a user receives a phone call requesting personal information in connection with a refund. In this case, the device records the conversation and sends it to a server. The server detects dangerous phrases such as "refund" and "personal information" and immediately sends a warning to the user and their registered family members, thereby ensuring the safety of the user and those around them.

[0727] An example of a prompt message might be: "Please determine if the following text contains potential fraud: 'You are eligible for a refund, please provide your bank information.' If it does, please also explain the factors that indicate it is fraudulent."

[0728] The flow of a specific process in Application Example 1 will be explained using Figure 12.

[0729] Step 1:

[0730] The device collects audio from the user's surroundings through a high-sensitivity microphone. The input is a real-time audio signal, which is then encrypted and compressed within the device. The output is compressed audio data in a format that can be transmitted quickly.

[0731] Step 2:

[0732] The terminal sends compressed and encrypted audio data to a server, which is an external processing unit, via a secure communication network. The input is compressed audio data, which is securely transferred to the server as output. Specifically, encrypted packets are created and sent using the HTTPS protocol.

[0733] Step 3:

[0734] The server uses the Google Speech-to-Text API to convert received audio data into text. The input is compressed audio data, which is then denoised and speaker-separated before generating plain text data as output. The server then passes this data to a generation AI model.

[0735] Step 4:

[0736] The server uses a generative AI model, specifically GPT-3, to analyze text data. The input is transformed text data, utilizing prompts to detect patterns and keywords related to fraud. The output is the analysis result, including data on the likelihood of fraud and specific risk factors.

[0737] Step 5:

[0738] If the analysis reveals a potential for fraud, the server generates a warning signal. The input is the analysis result, and the output is the warning signal based on this result. Specifically, it generates an alert message containing the problematic phrase or analysis result.

[0739] Step 6:

[0740] The server sends warning signals generated via the Twilio API to users and registered information processing devices. The input is the warning signal, and the output is a notification or SMS sent to the user's device or registered recipient. Specifically, a notification is immediately displayed on the user's smartphone.

[0741] Furthermore, an emotion engine that estimates the user's emotions may be incorporated. That is, the identification processing unit 290 may use the emotion identification model 59 to estimate the user's emotions and perform identification processing using the user's emotions.

[0742] This invention provides a system that analyzes voice input signals obtained from a user to detect the possibility of fraudulent activity while simultaneously recognizing the user's emotional state. In particular, by considering the user's psychological response in addition to the content of the voice, the reliability of fraud detection is enhanced.

[0743] The terminal acquires an audio input signal by recording the user's voice. The acquired audio data is compressed and encrypted on the terminal and sent to the server as needed. The server decrypts the received audio data and begins processing.

[0744] The server first uses speech recognition technology to convert the audio data into text data. This text data is then analyzed by a generative model designed to detect signs of fraud. The model compares it to existing fraud patterns to find matches.

[0745] Furthermore, the server uses an emotion engine to analyze the user's voice to determine their emotions. This emotion engine determines whether the user is experiencing psychological states such as anxiety or confusion, based on acoustic characteristics such as tone, tempo, and intonation.

[0746] If fraudulent activity is detected or the user is found to be in an unusual emotional state, the server generates a warning signal. This warning signal is sent to the user's device and to registered family members and police communications devices. The warning signal includes details about the detected fraud pattern and a summary of the user's emotional state.

[0747] To give a concrete example, suppose a user is talking about a "risky investment opportunity" over the phone. The device records this conversation and sends it to a server. The server converts it to text, and if the emotion engine detects signs of anxiety in the user's voice, it immediately generates a warning signal. This warning signal is then sent to the user's family, allowing them to take prompt action.

[0748] Thus, by combining voice data and emotional information, the present invention realizes a comprehensive system that not only protects users from fraud but also provides psychological support.

[0749] The following describes the processing flow.

[0750] Step 1:

[0751] The device constantly monitors the audio surrounding the user and starts recording. The audio input signal is acquired through a high-sensitivity microphone.

[0752] Step 2:

[0753] The device compresses recorded audio input signals in real time and encrypts them to ensure security. This achieves both communication efficiency and information protection.

[0754] Step 3:

[0755] The device transmits compressed and encrypted audio data to the server at regular intervals. This transmission uses either Wi-Fi or mobile data, depending on network conditions.

[0756] Step 4:

[0757] The server decodes the received audio data and converts it into text data through a speech recognition system. At this stage, noise reduction is applied to obtain clear text information.

[0758] Step 5:

[0759] The server feeds the text data converted from the speech into a generative model, which analyzes patterns and keywords that suggest potential fraud.

[0760] Step 6:

[0761] The server simultaneously uses an emotion engine to evaluate the user's emotional state based on the tone and tempo of their voice. This can detect unusual mental states such as anxiety or tension.

[0762] Step 7:

[0763] The server generates a warning signal if it detects signs of fraud or unusual emotional states. This warning signal includes information about the analyzed fraud patterns and the user's emotional state.

[0764] Step 8:

[0765] The server sends a warning signal to the user's device and the communication devices of registered family members or the police. This allows for real-time sharing of the situation.

[0766] Step 9:

[0767] The device notifies the user using audio and visual feedback based on the received warning signal. This allows the user to become more vigilant against fraud and to coordinate with family members.

[0768] (Example 2)

[0769] Next, we will describe Example 2. In the following description, the data processing device 12 will be referred to as the "server" and the robot 414 as the "terminal".

[0770] Conventional acoustic signal analysis systems aimed at detecting fraudulent activity, but they did not adequately consider the user's psychological state. Therefore, there is a need for a method that can quickly grasp emotions and psychological changes to enable more appropriate warnings and interventions.

[0771] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 2 is realized by the following means.

[0772] In this invention, the server includes means for converting acoustic signals into document data, means for analyzing the document data using a generative model to detect patterns related to fraud, and means for analyzing acoustic characteristics to determine the user's psychological state. This makes it possible to detect both signs of fraudulent activity and the user's psychological state, enabling faster and more appropriate warnings and responses.

[0773] An "acoustic signal" is a representation of sound as an electrical signal, which is acquired by devices such as microphones.

[0774] "Document data" refers to data in text format that has been converted from acoustic signals using speech recognition technology.

[0775] A "generative model" is a statistical model used to analyze data and detect specific patterns using machine learning and artificial intelligence techniques.

[0776] "Acoustic characteristics" refer to physical properties such as tone, tempo, and intonation that are extracted from audio data.

[0777] "Psychological state" refers to information that indicates the user's emotions and mental condition, analyzed based on the tone, tempo, and intonation of their voice.

[0778] A "warning signal" is a signal that is generated and transmitted to draw attention when the possibility of fraud or an unusual psychological state is detected.

[0779] "Communication equipment" refers to devices used to send and receive data and signals, and includes, for example, smartphones and tablets.

[0780] An "information transmission device" is a device that has the function of transmitting digital data or signals to other devices.

[0781] This invention is a system that uses acoustic signals to detect fraudulent activity and analyze the user's psychological state. The system converts acoustic signals into document data and uses that data to detect fraud patterns and analyze emotions.

[0782] The terminal acquires audio through the user's microphone. The acquired audio signal is compressed and encrypted locally and sent to the server over the communication network. A common encryption library is used for the encryption process.

[0783] The server decodes the received acoustic signal and first converts it into document data using advanced speech recognition software. Speech recognition technologies such as the Google Speech-to-Text API can be applied at this stage. Next, this document data is analyzed by a generative AI model. This generative AI model is trained on an extensive dataset of fraud cases and detects whether the text data contains signs of fraud.

[0784] Furthermore, the server is equipped with an emotion engine that determines psychological state based on acoustic characteristics. This emotion engine analyzes the tone, tempo, and intonation of the voice to evaluate whether the user is in a specific psychological state, such as anxiety or confusion.

[0785] For example, if a user is talking on the phone about a "business opportunity to make a quick profit," the device records the conversation and sends it to a server. The server converts the audio to text and uses an emotion engine to detect signs of anxiety. If this matches signs of fraud, the server immediately generates a warning signal and sends it to the user's relatives or other relevant parties.

[0786] Examples of prompts to input into a generative AI model:

[0787] Design a system that determines a user's psychological state from their acoustic signals and detects the possibility of fraudulent activity. The input should be acoustic signal data, and the output should be the degree of matching with fraud patterns and the results of the psychological state analysis.

[0788] Therefore, the system can comprehensively address fraudulent activities and provide psychological support.

[0789] The flow of the specific processing in Example 2 will be explained using Figure 13.

[0790] Step 1:

[0791] The device acquires the user's audio signal through a microphone. The input is the user's raw voice, and the output is an audio signal in digital format. In this step, the audio is captured in real time and data conversion processing is performed. The acquired audio signal is compressed and encrypted on the spot to ensure security.

[0792] Step 2:

[0793] The terminal transmits an encrypted acoustic signal to the server using a communication network. The input is the encrypted acoustic signal, and the output is the data transmitted to the server. In this step, the data travels to the server via a network such as the internet.

[0794] Step 3:

[0795] The server decrypts the received acoustic signal to prepare the data. The input is an encrypted acoustic signal, and the output is the available acoustic signal. Decryption is performed using a common encryption technique.

[0796] Step 4:

[0797] The server uses advanced speech recognition technology to convert acoustic signals into document data. The input is the decoded acoustic signal, and the output is text data. This process is performed using a speech recognition engine such as the Google Speech-to-Text API.

[0798] Step 5:

[0799] The server analyzes document data using a generative AI model. The input is text data generated by speech recognition, and the output is an analysis of the likelihood of fraud. The model scans the text and looks for parts that match known fraud patterns.

[0800] Step 6:

[0801] The server operates an emotion engine that determines psychological state based on acoustic characteristics. The input is an acoustic signal, and the output is data indicating the user's emotional state. This analysis is performed based on factors such as the tone, tempo, and intonation of the voice.

[0802] Step 7:

[0803] The server generates an alert signal if it detects signs of fraud or unusual psychological states. The input is fraud detection results and psychological state data, and the output is the alert signal. This signal includes details about the fraud pattern and psychological state.

[0804] Step 8:

[0805] The server sends the generated warning signal to a pre-registered communication device. The input is the warning signal, and the output is the information sent to the communication device. The warning signal reaches the user's relatives and relevant organizations quickly, enabling a rapid response.

[0806] (Application Example 2)

[0807] Next, we will explain application example 2. In the following explanation, the data processing device 12 will be referred to as the "server" and the robot 414 as the "terminal".

[0808] In modern society, fraud via voice calls is on the rise, making countermeasures an urgent necessity. Furthermore, victims often experience psychological distress and confusion during fraudulent activities, highlighting the need for support. However, current technology is insufficient to simultaneously detect voice-based fraud and analyze emotional states. Therefore, the challenge lies in rapidly detecting potential fraudulent activity and providing information that takes into account the user's emotional state.

[0809] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means.

[0810] In this invention, the server includes a medium for acquiring audio signals, a medium for converting audio signals into corresponding text information, a medium for analyzing the text information using a generation model to identify patterns related to fraud, and a medium for analyzing the user's emotional state based on the audio signals and generating psychological support information if the possibility of fraud or a specific emotional state is recognized. This enables early detection of fraudulent activity and the provision of appropriate support information based on the user's emotional state.

[0811] A "medium for acquiring audio signals" is a device or system that receives audio signals input from a user and records them for processing.

[0812] A "medium for converting into text information" is a device or system that analyzes acquired audio signals and expresses their content as text data.

[0813] "Using generative models" refers to a method of detecting specific patterns or features based on speech or text data, utilizing machine learning and AI technologies.

[0814] A "medium for identifying patterns related to fraud" is a device or mechanism for identifying potential fraud from audio or text based on the characteristics of known fraudulent activities.

[0815] A "medium for analyzing a user's emotional state" is a device or mechanism that analyzes information such as tone, rhythm, and intonation of an audio signal to identify the user's emotions.

[0816] A "medium for generating psychological support information" is a device or mechanism that generates information to provide reassurance or warnings based on the user's emotional state and the possibility of fraud.

[0817] "Warning information" refers to information used to alert those involved when potential fraud or specific emotional states are detected.

[0818] "A medium for transmitting information to a registered device or communication device" refers to a device or mechanism for transmitting generated warning information or support information to a communication device used by a user or related party.

[0819] This invention is a system that converts audio signals into text information using a speech recognition system, analyzes that text information using a generation AI model to detect the possibility of fraud, and further determines the user's emotional state from the audio signal. To enable this, the server and terminal must coordinate their operations.

[0820] The server collects audio data through hardware designed to acquire audio signals (e.g., audio input devices such as smartphones). The acquired audio signals are converted into text information using speech recognition technology on the server. This speech recognition uses common speech recognition libraries or cloud-based APIs (e.g., Google Speech-to-Text API).

[0821] Next, the server uses a generative AI model (e.g., a machine learning model) to analyze the converted text information. This analysis identifies patterns related to fraud. By utilizing a trained model to compare with known fraud patterns, advanced fraud detection is achieved.

[0822] Meanwhile, the server analyzes the user's emotional state from the audio signal. This uses an emotion analysis engine based on acoustic characteristics such as tone, tempo, and intonation. This allows the server to determine whether the user is experiencing emotional states such as anxiety or confusion.

[0823] If potential fraud or a specific emotional state is detected, the server generates a warning message and transmits it to the terminal or communication device. This allows users and registered parties to be aware of the risk of fraud or if the user is in need of psychological support.

[0824] For example, if a user receives a suspicious financial call, the sentiment analysis engine might detect anxiety in the user's voice. In this case, the system would immediately generate a warning and notify the user's family and registered contacts, enabling early intervention.

[0825] As an example of a prompt message to a generative AI model, it could be something like, "Detect if this call is a scam. Text: 'There is a new investment opportunity...'" This allows the model to analyze the likelihood of fraud based on text information.

[0826] The flow of a specific process in Application Example 2 will be explained using Figure 14.

[0827] Step 1:

[0828] The terminal uses a microphone to acquire the user's voice signal. It uses the raw voice signal as input, converts it to a digital format, and sends it to the server. This allows for the preparation of a suitable data format for preprocessing of speech recognition.

[0829] Step 2:

[0830] The server converts the acquired audio signal into text information using speech recognition software (e.g., Google Speech-to-Text API). The input is a digital audio signal, and the output is the corresponding string data. Text information is obtained by analyzing the acoustic characteristics of the audio signal and converting it into text.

[0831] Step 3:

[0832] The server analyzes the character information converted using a generative AI model. It receives string data as input and detects features that indicate potential fraud. The output is a fraud score or information about signs of fraud. A risk assessment is performed based on comparison with known fraud patterns.

[0833] Step 4:

[0834] Simultaneously, the server processes the audio signal itself into an acoustic analysis engine to analyze the user's emotional state. The input is a digital audio signal, and the output is data indicating the user's emotional state. By extracting emotional characteristics from the tone, tempo, and intonation of the voice, the server determines the user's psychological situation.

[0835] Step 5:

[0836] The server generates and sends alert information to a terminal or registered communication device if it detects potential fraud or a specific emotional state. It uses a fraud score and emotional state as input and outputs an alert message. This provides immediate alerts to users and registered recipients.

[0837] Step 6:

[0838] Users take appropriate action based on the warning information they receive. Input is a warning message, and output can include specific actions or notifications to family members. Based on the warning content, support for fraud avoidance and psychological well-being is provided quickly.

[0839] The specific processing unit 290 transmits the result of the specific processing to the robot 414. In the robot 414, the control unit 46A causes the speaker 240 and the controlled object 443 to output the result of the specific processing. The microphone 238 acquires audio indicating user input for the result of the specific processing. The control unit 46A transmits the audio data indicating user input acquired by the microphone 238 to the data processing unit 12. In the data processing unit 12, the specific processing unit 290 acquires the audio data.

[0840] Data generation model 58 is a type of so-called generative AI (Artificial Intelligence). One example of data generation model 58 is ChatGPT (Internet search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search) <url: https: gemini.google.com ?hl="ja">Examples of generative AI include those described above. The data generation model 58 is obtained by performing deep learning on a neural network. The data generation model 58 is input with prompts containing instructions, and with inference data such as audio data representing speech, text data representing text, and image data representing images. The data generation model 58 infers from the input inference data according to the instructions shown by the prompts, and outputs the inference results in data formats such as audio data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[0841] 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 this disclosure is not limited thereto, and the specific processing may also be performed by the robot 414.

[0842] Furthermore, the emotion identification model 59, acting as an emotion engine, may determine the user's emotion according to a specific mapping. Specifically, the emotion identification model 59 may determine the user's emotion according to a specific mapping, which is an emotion map (see Figure 9). Similarly, the emotion identification model 59 may also determine the robot's emotion, and the identification processing unit 290 may perform identification processing using the robot's emotion.

[0843] Figure 9 shows an emotion map 400 in which multiple emotions are mapped. In the emotion map 400, emotions are arranged in concentric circles radiating from the center. The closer to the center of the concentric circles, the more primitive the emotions are located. Further out of the concentric circles, emotions representing states and actions arising from mental states are located. Emotion is a concept that includes feelings and mental states. On the left side of the concentric circles, emotions that are generally generated from reactions occurring in the brain are located. On the right side of the concentric circles, emotions that are generally induced by situational judgment are located. Above and below the concentric circles, emotions that are generally generated from reactions occurring in the brain and induced by situational judgment are located. In addition, the emotion of "pleasure" is located on the upper side of the concentric circles, and the emotion of "displeasure" is located on the lower side. Thus, in the emotion map 400, multiple emotions are mapped based on the structure in which emotions arise, and emotions that are likely to occur simultaneously are mapped close together.

[0844] These emotions are distributed at the 3 o'clock position on the Emotion Map 400, and usually fluctuate between feelings of security and anxiety. In the right half of the Emotion Map 400, situational awareness takes precedence over internal feelings, resulting in a calm impression.

[0845] The inside of the Emotion Map 400 represents inner thoughts, while the outside represents actions. Therefore, the further you go from the outside of the Emotion Map 400, the more visible (expressed in actions) your emotions become.

[0846] Here, human emotions are based on various balances, such as posture and blood sugar levels. When these balances deviate from the ideal, it results in discomfort, and when they approach the ideal, it results in pleasure. Similarly, in robots, cars, motorcycles, etc., emotions can be created based on various balances, such as posture and battery level. When these balances deviate from the ideal, it results in discomfort, and when they approach the ideal, it results in pleasure. The emotion map can be generated, for example, based on Dr. Mitsuyoshi's emotion map (Research on a system for analyzing brain physiological signals of speech emotion recognition and emotion, Tokushima University, doctoral dissertation: https: / / ci.nii.ac.jp / naid / 500000375379). The left half of the emotion map contains emotions belonging to a region called "response," where sensation is dominant. The right half of the emotion map contains emotions belonging to a region called "situation," where situational awareness is dominant.

[0847] The emotion map defines two emotions that promote learning. One is the emotion around the middle of the negative "repentance" and "reflection" on the situation side. In other words, it is when the robot experiences negative emotions such as "I never want to feel this way again" or "I don't want to be scolded again." The other is the emotion around the positive "desire" on the reaction side. In other words, it is when the robot has positive feelings such as "I want more" or "I want to know more."

[0848] The emotion identification model 59 inputs user input into a pre-trained neural network, obtains emotion values ​​representing each emotion shown in the emotion map 400, and determines the user's emotion. This neural network is pre-trained based on multiple training data sets, which are combinations of user input and emotion values ​​representing each emotion shown in the emotion map 400. Furthermore, this neural network is trained so that emotions located close together have similar values, as shown in the emotion map 900 in Figure 10. Figure 10 shows an example where multiple emotions such as "reassured," "calm," and "confident" have similar emotion values.

[0849] The above description primarily focuses on the functions of the data processing device 12 in relation to this disclosure. However, the system related to this disclosure is not necessarily implemented on a server. The system related to this disclosure may be implemented as a general information processing system. This disclosure may be implemented, for example, as a software program that runs on a personal computer or as an application that runs on a smartphone. The method related to this disclosure may be provided to users in SaaS (Software as a Service) format.

[0850] In the above embodiment, an example was given in which a specific process is performed by a single computer 22. However, the technology of this disclosure is not limited thereto, and a distributed processing of the specific process may be performed by multiple computers, including computer 22. For example, a data generation model 58 may be provided in an external device of the data processing device 12, and the external device may generate data according to the input data.

[0851] In the above embodiment, an example was given in which the specific processing program 56 is stored in the storage 32, but the technology of this disclosure is not limited thereto. For example, the specific processing program 56 may be stored in a portable, computer-readable, non-temporary storage medium such as a USB (Universal Serial Bus) memory. The specific processing program 56 stored in the non-temporary storage medium is installed in the computer 22 of the data processing device 12. The processor 28 executes specific processing according to the specific processing program 56.

[0852] 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.

[0853] Furthermore, it is not necessary to store the entirety of the specific processing program 56 in a storage device such as a server connected to the data processing device 12 via the network 54, or to store the entirety of the specific processing program 56 in the storage 32; it is acceptable to store only a portion of the specific processing program 56.

[0854] The following types of processors can be used as hardware resources to perform specific processing. Examples of processors include a CPU, a general-purpose processor that functions as a hardware resource to perform specific processing by executing software, i.e., a program. Other examples of processors include dedicated electrical circuits, such as FPGAs (Field-Programmable Gate Arrays), PLDs (Programmable Logic Devices), or ASICs (Application Specific Integrated Circuits), which have circuit configurations specifically designed to perform specific processing. All of these processors have built-in or connected memory, and all of them perform specific processing by using memory.

[0855] The hardware resource that performs a specific process may consist of one of these various processors, or it may consist of a combination of two or more processors of the same or different types (for example, a combination of multiple FPGAs, or a combination of a CPU and an FPGA). Alternatively, the hardware resource that performs a specific process may consist of a single processor.

[0856] Examples of configurations using a single processor include, firstly, a configuration in which one or more CPUs and software are combined to form a single processor, and this processor functions as a hardware resource that performs a specific process. Secondly, there is a configuration using a processor that realizes the functions of the entire system, including multiple hardware resources that perform a specific process, on a single IC chip, as exemplified by SoCs (System-on-a-chip). In this way, a specific process is realized using one or more of the above types of processors as hardware resources.

[0857] Furthermore, the hardware structure of these various processors can more specifically utilize electrical circuits that combine circuit elements such as semiconductor devices. Also, the specific processing described above is merely an example. Therefore, it goes without saying that unnecessary steps can be deleted, new steps added, or the processing order rearranged, as long as it does not deviate from the main purpose.

[0858] The descriptions and illustrations presented above are detailed explanations of the technical aspects of this disclosure and are merely examples of the technical aspects. For example, the above descriptions of the structure, function, operation, and effect are examples of the structure, function, operation, and effect of the technical aspects of this disclosure. Therefore, it goes without saying that you may delete unnecessary parts, add new elements, or replace elements in the descriptions and illustrations presented above, as long as you do not deviate from the essence of the technical aspects of this disclosure. Furthermore, in order to avoid confusion and facilitate understanding of the technical aspects of this disclosure, explanations of common technical knowledge and the like that do not require special explanation to enable the implementation of the technical aspects of this disclosure have been omitted from the descriptions and illustrations presented above.

[0859] All documents, patent applications, and technical standards described herein are incorporated by reference to the same extent as if each individual document, patent application, and technical standard were specifically and individually noted as being incorporated by reference.

[0860] The following is further disclosed regarding the embodiments described above.

[0861] (Claim 1)

[0862] A means for acquiring an audio input signal,

[0863] means for converting the aforementioned audio input signal into text data,

[0864] A means for analyzing the aforementioned text data using a generative model to detect patterns related to fraud,

[0865] A means of generating a warning signal when it is determined that there is a possibility of fraud,

[0866] Means for transmitting the aforementioned warning signal to a registered terminal or communication device,

[0867] A system that includes this.

[0868] (Claim 2)

[0869] The system according to claim 1, wherein the aforementioned audio input signal is compressed and encrypted and transmitted over a communication line.

[0870] (Claim 3)

[0871] The system according to claim 1, which transmits a summary of the conversation corresponding to the warning signal.

[0872] "Example 1"

[0873] (Claim 1)

[0874] A device for acquiring sound,

[0875] A device for converting the aforementioned audio into digital data,

[0876] A device for compressing and encrypting the aforementioned digital data,

[0877] A device for transmitting the compressed and encrypted data over a communication channel,

[0878] A device that receives and decodes the transmitted data,

[0879] A device that converts decoded audio data into text data,

[0880] A device that uses generation technology to analyze the aforementioned text data and detect patterns related to fraud,

[0881] A device that generates a warning signal when it is determined that there is a possibility of fraud,

[0882] A device that transmits the aforementioned warning signal to a designated device or communication device,

[0883] A system that includes this.

[0884] (Claim 2)

[0885] The system according to claim 1, which compresses the aforementioned digital data with high efficiency and encrypts it to ensure the security of the information.

[0886] (Claim 3)

[0887] The system according to claim 1, which transmits a summary of the conversation and a description of the situation related to the warning signal.

[0888] "Application Example 1"

[0889] (Claim 1)

[0890] A device for acquiring audio input signals,

[0891] A device that converts the aforementioned audio input signal into text representation,

[0892] A device that analyzes the aforementioned text representation using a generative model and detects patterns related to fraud,

[0893] A device that generates a warning signal when fraud is detected,

[0894] A device that transmits the aforementioned warning signal to a registered information processing device or communication means,

[0895] A device that encrypts acquired audio data and transmits it to an external processing unit via a secure communication line,

[0896] An external processing device analyzes voice data with high accuracy to quickly warn of fraud risks and notifies registered information processing devices, including security information.

[0897] A system that includes this.

[0898] (Claim 2)

[0899] The system according to claim 1, comprising a device for compressing and encrypting acquired audio data and transmitting it to an external processing device using secure communication means.

[0900] (Claim 3)

[0901] The system according to claim 1, comprising a device that transmits the warning signal including summary data of the relevant conversation and analysis results using a generative model.

[0902] "Example 2 of combining an emotion engine"

[0903] (Claim 1)

[0904] Means for acquiring acoustic signals,

[0905] means for converting the aforementioned acoustic signal into document data,

[0906] A means for analyzing the aforementioned document data using a generative model to detect patterns related to fraud,

[0907] A method for determining psychological state by analyzing acoustic characteristics,

[0908] A means for generating a warning signal when the possibility of fraud or an unusual psychological state is detected,

[0909] Means for transmitting the aforementioned warning signal to a registered communication device or information transmission device,

[0910] ...

[0911] A system that includes this.

[0912] (Claim 2)

[0913] The system according to claim 1, wherein the aforementioned acoustic signal is compressed and encrypted and transmitted over an information line.

[0914] (Claim 3)

[0915] The system according to claim 1, which transmits a summary of the conversation corresponding to the warning signal and the identified psychological state.

[0916] "Application example 2 when combining with an emotional engine"

[0917] (Claim 1)

[0918] A medium for acquiring audio signals,

[0919] A medium for converting the aforementioned audio signal into corresponding character information,

[0920] A medium that analyzes the aforementioned textual information using a generative model to identify patterns related to fraud,

[0921] A medium that generates warning information when it is determined that there is a possibility of fraud,

[0922] A medium for transmitting the aforementioned warning information to a registered device or communication device,

[0923] A medium that analyzes the user's emotional state based on the aforementioned audio signal and generates psychological support information if the possibility of fraud or a specific emotional state is identified,

[0924] A system that includes this.

[0925] (Claim 2)

[0926] The system according to claim 1, wherein the aforementioned audio signal is compressed and encrypted and transmitted via a communication path.

[0927] (Claim 3)

[0928] The system according to claim 1, which transmits summary information of the conversation corresponding to the warning information and information on the user's emotional state. [Explanation of Symbols]

[0929] 10, 210, 310, 410 Data Processing Systems 12 Data Processing Devices 14 Smart Devices 214 Smart Glasses 314 Headset-type terminal 414 Robots< / url:> < / url:> < / url:> < / url:>

Claims

1. A means for acquiring an audio input signal, means for converting the aforementioned audio input signal into text data, A means for analyzing the aforementioned text data using a generative model to detect patterns related to fraud, A means of generating a warning signal when it is determined that there is a possibility of fraud, Means for transmitting the aforementioned warning signal to a registered terminal or communication device, A system that includes this.

2. The system according to claim 1, wherein the aforementioned audio input signal is compressed and encrypted and transmitted over a communication line.

3. The system according to claim 1, which transmits a summary of the conversation corresponding to the warning signal.

Images