system

Abstract

The system according to the embodiment aims to detect refund fraud conversations in real time and respond quickly. [Solution] A system according to an embodiment includes an acquisition unit, an analysis unit, a notification unit, and a transmission unit. The acquisition unit acquires conversation content. The analysis unit analyzes the conversation content acquired by the acquisition unit and determines whether or not there is refund fraud. The notification unit sends an alert based on the result determined by the analysis unit. The transmission unit transmits the alert sent by the notification unit.

Description

[Technical Field]

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

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

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

[0004] Conventional technology has had the problem of making it difficult to detect refund fraud conversations in real time and respond quickly.

[0005] The system according to the embodiment aims to detect refund fraud conversations in real time and respond quickly. [Means for solving the problem]

[0006] The system according to the embodiment includes an acquisition unit, an analysis unit, a notification unit, and a transmission unit. The acquisition unit acquires the conversation content. The analysis unit analyzes the conversation content acquired by the acquisition unit and determines whether or not there is refund fraud. The notification unit transmits an alert based on the result determined by the analysis unit. The transmission unit transmits the alert transmitted by the notification unit. [Effects of the Invention]

[0007] The system according to the embodiment can detect refund fraud conversations in real time and respond quickly. [Brief explanation of the drawings]

[0008] [Figure 1] 1 is a conceptual diagram showing an example of the configuration of a data processing system according to a first embodiment. [Figure 2] 1 is a conceptual diagram showing an example of main functions of a data processing device and a smart device according to a first embodiment. [Figure 3] FIG. 10 is a conceptual diagram showing an example of the configuration of a data processing system according to a second embodiment. [Figure 4] FIG. 10 is a conceptual diagram showing an example of main functions of a data processing device and smart glasses according to a second embodiment. [Figure 5] FIG. 10 is a conceptual diagram showing an example of the configuration of a data processing system according to a third embodiment. [Figure 6] FIG. 11 is a conceptual diagram showing an example of main functions of a data processing device and a headset-type terminal according to a third embodiment. [Figure 7] FIG. 10 is a conceptual diagram showing an example of the configuration of a data processing system according to a fourth embodiment. [Figure 8] FIG. 10 is a conceptual diagram showing an example of main functions of a data processing device and a robot according to a fourth embodiment. [Figure 9] 1 shows an emotion map onto which multiple emotions are mapped. [Figure 10] 1 shows an emotion map onto which multiple emotions are mapped. DETAILED DESCRIPTION OF THE INVENTION

[0009] An example of an embodiment of a system according to the technology of the present disclosure will be described below with reference to the accompanying drawings.

[0010] First, the terms used in the following description will be explained.

[0011] In the following embodiments, a coded processor (hereinafter simply referred to as a "processor") may be a single arithmetic device or a combination of multiple arithmetic devices. Furthermore, the processor may be a single type of arithmetic device or a combination of multiple types of arithmetic devices. Examples of arithmetic devices include a CPU (Central Processing Unit), a GPU (Graphics Processing Unit), a GPGPU (General-Purpose computing on Graphics Processing Units), an APU (Accelerated Processing Unit), or a TPU (Tensor Processing Unit).

[0012] In the following embodiments, a coded RAM (Random Access Memory) is a memory in which information is temporarily stored and is used as a working memory by a processor.

[0013] In the following embodiments, the coded storage is one or more non-volatile storage devices that store various programs, various parameters, etc. Examples of non-volatile storage devices include flash memory (SSD (Solid State Drive)), magnetic disks (e.g., hard disks), and magnetic tapes.

[0014] In the following embodiments, a communication I / F (Interface) with a symbol is an interface including a communication processor, an antenna, etc. The communication I / F controls communication between multiple computers. Examples of communication standards applied to the communication I / F include wireless communication standards including 5G (5th Generation Mobile Communication System), Wi-Fi (registered trademark), and Bluetooth (registered trademark).

[0015] In the following embodiments, "A and / or B" is synonymous with "at least one of A and B." In other words, "A and / or B" means that it may be only A, only B, or a combination of A and B. Furthermore, in this specification, the same concept as "A and / or B" is also applied when three or more things are expressed connected by "and / or."

[0016] [First embodiment] FIG. 1 shows an example of the configuration of a data processing system 10 according to the first embodiment.

[0017] 1, a data processing system 10 includes a data processing device 12 and a smart device 14. An example of the data processing device 12 is a server.

[0018] The data processing device 12 includes a computer 22, a database 24, and a communication I / F 26. The computer 22 includes a processor 28, a RAM 30, and a storage 32. The processor 28, RAM 30, and storage 32 are connected to a bus 34. The database 24 and the communication I / F 26 are also connected to the bus 34. The communication I / F 26 is connected to a network 54. Examples of the network 54 include a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0019] The smart device 14 includes a computer 36, a reception device 38, an output device 40, a camera 42, and a communication I / F 44. The computer 36 includes a processor 46, a RAM 48, and a storage 50. The processor 46, the RAM 48, and the storage 50 are connected to a bus 52. The reception device 38, the output device 40, and the camera 42 are also connected to the bus 52.

[0020] The reception device 38 includes a touch panel 38A and a microphone 38B, and receives user input. The touch panel 38A detects contact with a pointer (for example, a pen or a finger) to receive user input by the touch of the pointer. The microphone 38B detects the user's voice to receive user input by voice. The control unit 46A transmits data indicating the user input received by the touch panel 38A and the microphone 38B to the data processing device 12. In the data processing device 12, the specific processing unit 290 (see FIG. 2) acquires the data indicating the user input.

[0021] Output device 40 includes a display 40A and a speaker 40B, and presents data to a user by outputting the data in a form of expression that the user can perceive (e.g., audio and / or text). Display 40A displays visible information such as text and images in accordance with instructions from processor 46. Speaker 40B outputs audio in accordance with instructions from processor 46. Camera 42 is a compact digital camera equipped with an optical system including a lens, aperture, and shutter, and an imaging element such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor.

[0022] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 control the exchange of various information between the processor 46 and the processor 28 via the network 54.

[0023] FIG. 2 shows an example of the main functions of the data processing device 12 and the smart device 14.

[0024] 2, in the data processing device 12, a specific process is performed by the processor 28. A specific processing program 56 is stored in the storage 32. The specific processing program 56 is an example of a "program" according to the technology of the present disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific process is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific processing program 56 executed on the RAM 30.

[0025] The storage 32 stores a data generation model 58 and an emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290. The identification processing unit 290 can estimate a user's emotion using the emotion identification model 59 and perform identification processing using the user's emotion. The emotion estimation function (emotion identification function) using the emotion identification model 59 performs various estimations and predictions regarding the user's emotion, including estimation and prediction of the user's emotion, but is not limited to these examples. Furthermore, the estimation and prediction of emotion also includes, for example, emotion analysis.

[0026] In the smart device 14, the specific processing is performed by the processor 46. The storage 50 stores a specific processing program 60. The specific processing program 60 is used together with the specific processing program 56 by the data processing system 10. The processor 46 reads the specific processing program 60 from the storage 50 and executes the read specific processing program 60 on the RAM 48. The specific processing is realized by the processor 46 operating as the control unit 46A in accordance with the specific processing program 60 executed on the RAM 48. Note that the smart device 14 has a data generation model and an emotion identification model similar to the data generation model 58 and the emotion identification model 59, and can also perform processing similar to that of the specific processing unit 290 using these models.

[0027] Note that a device other than the data processing device 12 may have the data generation model 58. For example, a server device (e.g., a generation server) may have the data generation model 58. In this case, the data processing device 12 obtains a processing result (prediction result, etc.) using the data generation model 58 by communicating with the server device having the data generation model 58. Furthermore, the data processing device 12 may be a server device, or may be a terminal device owned by a user (e.g., a mobile phone, a robot, a home appliance, etc.). Next, an example of processing by the data processing system 10 according to the first embodiment will be described.

[0028] (Example 1) A refund fraud detection system according to an embodiment of the present invention acquires conversation content, analyzes the possibility of refund fraud, and sends an alert. The refund fraud detection system acquires conversation content, analyzes the possibility of refund fraud, and sends an alert, thereby preventing fraud before it occurs. For example, the refund fraud detection system installs a dedicated app on a dedicated terminal or smartphone. This app is equipped with a generation AI and can constantly listen to the owner's conversations. The generation AI then analyzes the conversation content and determines whether there is a possibility of refund fraud. For example, if keywords such as "refund," "transfer," and "bank account" are included, the generation AI will evaluate the possibility of fraud. If a possible fraud is detected, an alert is sent to the owner's family or the police. For example, the dedicated app automatically notifies the owner's family or the police, allowing them to take measures before the owner is defrauded. In this way, the refund fraud detection system can prevent refund fraud before it occurs. For example, the system constantly listens to the owner's conversations, detects conversations that may be tax refund fraud, and sends an alert to family members or the police, allowing for swift countermeasures to be taken. Privacy is thoroughly protected, and when the generating AI analyzes the content of conversations, it is designed to prevent personal information from leaking to the outside. For example, the content of conversations is analyzed only within the device and is not sent to an external server. This system can prevent tax refund fraud before it occurs. The generating AI can anticipate countless possible scenarios for becoming a victim and quickly issue an alert, ensuring the safety of the owner.

[0029] A refund fraud detection system according to an embodiment includes an acquisition unit, an analysis unit, a notification unit, and a provision unit. The acquisition unit acquires conversation content. The conversation content may include, but is not limited to, voice data, text data, or a part or all of a conversation. The acquisition unit, for example, constantly listens to the owner's conversation. The acquisition unit may detect, for example, the content of the owner's phone conversation or phrases in everyday conversation that may be related to refund fraud. The analysis unit analyzes the conversation content acquired by the acquisition unit and determines whether or not refund fraud has occurred. The analysis unit detects conversation content containing keywords such as "refund," "transfer," and "bank account." The analysis unit analyzes the conversation content using a generation AI. The generation AI may be a text generation AI (e.g., LLM) or a multimodal generation AI, and determines the possibility of refund fraud based on the conversation content. The notification unit sends an alert based on the result determined by the analysis unit. For example, if the notification unit detects a possibility of fraud, it sends an alert to the owner's family or the police. The notification unit can send an alert using the generation AI. The provision unit provides the alert sent by the notification unit. The provision unit provides the alert, for example, to family members or the police. The provision unit can provide the alert using the generation AI. As a result, the refund fraud detection system according to the embodiment can prevent damage caused by refund fraud. For example, the system can constantly listen to the owner's conversations, detect conversations that may be related to refund fraud, and send alerts to family members or the police, allowing for prompt countermeasures to be taken. Privacy protection is thorough, and when the generation AI analyzes the content of the conversation, it is designed to prevent personal information from leaking to the outside. For example, the content of the conversation is analyzed only within the device and not sent to an external server. This mechanism can prevent damage caused by refund fraud. The generation AI can anticipate countless possible scenarios for victimization and quickly issue alerts, ensuring the owner's safety.

[0030] The acquisition unit can continuously listen to the owner's conversation. For example, the acquisition unit constantly listens to the owner's conversation. For example, the acquisition unit can detect phrases that may be tax refund fraud in the content of the owner's phone conversation or in everyday conversation. The acquisition unit can acquire the content of the conversation using a generation AI. The generation AI is a text generation AI (e.g., LLM) or a multimodal generation AI, and determines the possibility of tax refund fraud based on the content of the conversation. In this way, by constantly listening to the owner's conversation, the possibility of tax refund fraud can be quickly detected. Some or all of the above-mentioned processing in the acquisition unit may be performed using, or without, the generation AI. For example, the acquisition unit can input the owner's conversation into the generation AI and cause the generation AI to acquire the content of the conversation.

[0031] The analysis unit can detect conversation content containing keywords such as "refund," "transfer," and "bank account." The analysis unit can detect conversation content containing keywords such as "refund," "transfer," and "bank account." The analysis unit can analyze the conversation content using a generation AI. The generation AI is a text generation AI (e.g., LLM) or a multimodal generation AI, and determines the possibility of refund fraud based on the conversation content. For example, the generation AI detects conversation content containing keywords such as "refund," "transfer," and "bank account" and evaluates the possibility of fraud. In this way, by detecting conversation content containing specific keywords, the possibility of refund fraud can be evaluated highly. Some or all of the above-mentioned processing in the analysis unit can be performed using, for example, the generation AI, or can be performed without using the generation AI. For example, the analysis unit can input the conversation content into the generation AI and have the generation AI analyze the possibility of refund fraud.

[0032] The notification unit can send an alert to family or the police when fraud is detected. For example, the notification unit sends an alert to family or the police when the possibility of fraud is detected. The notification unit can send the alert using a generation AI. The generation AI is a text generation AI (e.g., LLM) or a multimodal generation AI, and sends the alert. For example, the generation AI detects the possibility of fraud and sends an alert to family or the police. This makes it possible to prevent damage by quickly sending an alert to family or the police when the possibility of fraud is detected. Some or all of the above-mentioned processing in the notification unit may be performed using, or without, the generation AI. For example, the notification unit can input the possibility of fraud into the generation AI and have the generation AI send an alert.

[0033] The providing unit can send the alert to the family or the police. The providing unit, for example, provides the alert to the family or the police. The providing unit can provide the alert using a generation AI. The generation AI is a text generation AI (e.g., LLM) or a multimodal generation AI, and provides the alert. For example, the generation AI provides the alert to the family or the police. By providing the alert to the family or the police, a prompt response is possible. Some or all of the above-mentioned processing in the providing unit may be performed, for example, using the generation AI, or may be performed without using the generation AI. For example, the providing unit can input the alert to the generation AI and have the generation AI provide the alert.

[0034] The analysis unit can analyze the conversation content only within the terminal and not send it to an external server. For example, the analysis unit can analyze the conversation content only within the terminal and not send it to an external server. The analysis unit can analyze the conversation content using a generation AI. The generation AI can be a text generation AI (e.g., LLM) or a multimodal generation AI, and determine the possibility of tax refund fraud based on the conversation content. For example, the generation AI analyzes the conversation content within the terminal and does not send it to an external server. By analyzing the conversation content only within the terminal, privacy is thoroughly protected. Some or all of the above-mentioned processing in the analysis unit can be performed, for example, using the generation AI, or can be performed without using the generation AI. For example, the analysis unit can input the conversation content to the generation AI and have the generation AI analyze the conversation content.

[0035] The acquisition unit can analyze the user's past conversation history and select an appropriate acquisition method. For example, the acquisition unit analyzes the user's past conversation history and selects the optimal acquisition method. The acquisition unit can analyze the user's past conversation history using a generation AI. The generation AI is a text generation AI (e.g., LLM) or a multimodal generation AI, and selects the optimal acquisition method based on the user's past conversation history. For example, the generation AI analyzes the user's past conversation history and prioritizes acquisition of frequently used phrases. The generation AI can also analyze the user's past conversation history to find a tendency for important conversations to occur during specific time periods and focus acquisition on those time periods. The generation AI can also prioritize acquisition of conversations with specific people based on the user's past conversation history. This allows the analysis of the past conversation history to select the optimal acquisition method, enabling efficient conversation acquisition. Some or all of the above-mentioned processing in the acquisition unit may be performed using, for example, the generation AI, or may be performed without using the generation AI. For example, the acquisition unit can input the user's past conversation history into the generation AI and have the generation AI select the optimal acquisition method.

[0036] The acquisition unit can filter the conversations based on the user's current situation or areas of interest when acquiring the conversations. For example, the acquisition unit can filter the conversations based on the user's current situation or areas of interest when acquiring the conversations. The acquisition unit can analyze the user's current situation or areas of interest using a generation AI. The generation AI is a text generation AI (e.g., LLM) or a multimodal generation AI, and filters the conversations based on the user's current situation or areas of interest. For example, the generation AI can analyze the user's current situation and acquire only important conversations. The generation AI can also analyze the user's areas of interest and prioritize acquisition of related conversations. Furthermore, if the user is in a specific location, the generation AI can prioritize acquisition of conversations related to that location. In this way, by filtering the conversations based on the user's situation or areas of interest, important conversations can be prioritized and acquired. Some or all of the above-described processing in the acquisition unit can be performed using, for example, the generation AI, or can be performed without using the generation AI. For example, the acquisition unit can input the user's current situation or areas of interest to the generation AI and have the generation AI perform conversation filtering.

[0037] The acquisition unit can select an appropriate acquisition means according to the user's input method when acquiring a conversation. For example, the acquisition unit selects the optimal acquisition means according to the user's input method (voice, text, image, etc.) when acquiring a conversation. The acquisition unit can analyze the user's input method using a generation AI. The generation AI is a text generation AI (e.g., LLM) or a multimodal generation AI, and selects the optimal acquisition means based on the user's input method. For example, if the generation AI is using voice input, it can prioritize acquiring voice data. Also, if the user is using text input, it can prioritize acquiring text data. Also, if the user is using image input, it can prioritize acquiring image data. This enables efficient conversation acquisition by selecting the optimal acquisition means according to the user's input method. Some or all of the above-mentioned processing in the acquisition unit can be performed using, for example, the generation AI, or can be performed without using the generation AI. For example, the acquisition unit can input the user's input method to the generation AI and cause the generation AI to select the optimal acquisition means.

[0038] The acquisition unit can prioritize acquisition of highly relevant conversations based on the user's geographical location information when acquiring a conversation. For example, the acquisition unit prioritizes acquisition of highly relevant conversations by taking the user's geographical location information into consideration when acquiring a conversation. The acquisition unit can analyze the user's geographical location information using a generation AI. The generation AI is a text generation AI (e.g., LLM) or a multimodal generation AI, and prioritizes acquisition of highly relevant conversations based on the user's geographical location information. For example, when the user is in a specific location, the generation AI can prioritize acquisition of conversations related to that location. Furthermore, when the user is traveling, the generation AI can prioritize acquisition of conversations related to the user's destination. Furthermore, when the user is in a specific area, the generation AI can prioritize acquisition of conversations related to that area. In this way, by taking the user's geographical location information into consideration, highly relevant conversations can be prioritized. Some or all of the above-described processing in the acquisition unit may be performed using the generation AI, for example, or may be performed without using the generation AI. For example, the acquisition unit can input the user's geographical location information to the generation AI and cause the generation AI to acquire highly relevant conversations.

[0039] The acquisition unit can analyze the user's social media usage status and acquire related conversations when acquiring a conversation. For example, the acquisition unit can analyze the user's social media activity and acquire related conversations when acquiring a conversation. The acquisition unit can analyze the user's social media activity using a generation AI. The generation AI is a text generation AI (e.g., LLM) or a multimodal generation AI, and acquires related conversations based on the user's social media activity. For example, the generation AI can prioritize acquiring conversations related to places the user has checked in to on social media. The generation AI can also analyze the content of the user's social media posts and acquire related conversations. The generation AI can also acquire related conversations by referring to the activities of the user's friends on social media. In this way, related conversations can be efficiently acquired by analyzing social media activity. Some or all of the above-mentioned processing in the acquisition unit can be performed using, for example, the generation AI, or can be performed without using the generation AI. For example, the acquisition unit can input the user's social media activity into the generation AI and cause the generation AI to acquire related conversations.

[0040] The acquisition unit can adjust the acquisition method by reflecting the user's past feedback when acquiring a conversation. For example, the acquisition unit customizes the acquisition method by reflecting the user's past feedback when acquiring a conversation. The acquisition unit can analyze the user's past feedback using a generation AI. The generation AI is a text generation AI (e.g., LLM) or a multimodal generation AI, and adjusts the acquisition method based on the user's past feedback. For example, the generation AI prioritizes acquisition of conversations that the user previously evaluated as important. The generation AI can also adjust the acquisition method based on the user's past feedback. The generation AI can also exclude and acquire conversations that the user previously evaluated as unnecessary. In this way, by reflecting past feedback, the acquisition method can be customized, enabling efficient conversation acquisition. Some or all of the above-mentioned processing in the acquisition unit may be performed using, for example, the generation AI, or may be performed without using the generation AI. For example, the acquisition unit can input the user's past feedback into the generation AI and cause the generation AI to adjust the acquisition method.

[0041] The analysis unit can adjust the level of analysis detail based on the priority of the conversation during analysis. For example, the analysis unit can adjust the level of analysis detail based on the importance of the conversation during analysis. The analysis unit can use a generation AI to evaluate the importance of the conversation. The generation AI is a text generation AI (e.g., LLM) or a multimodal generation AI, and adjusts the level of analysis detail based on the importance of the conversation. For example, the generation AI performs a detailed analysis of a conversation that is highly important. The generation AI can also perform a concise analysis of a conversation that is less important. The generation AI can also determine the priority of the analysis according to the importance of the conversation. This enables efficient analysis by adjusting the level of analysis detail based on the importance of the conversation. Some or all of the above-mentioned processing in the analysis unit may be performed using, or without, the generation AI. For example, the analysis unit can input the importance of the conversation to the generation AI and have the generation AI adjust the level of analysis detail.

[0042] The analysis unit can apply different analysis algorithms depending on the type of conversation during analysis. For example, the analysis unit can apply different analysis algorithms depending on the category of the conversation during analysis. The analysis unit can classify the category of the conversation using a generation AI. The generation AI can be a text generation AI (e.g., LLM) or a multimodal generation AI, and selects an appropriate analysis algorithm depending on the category of the conversation. For example, the generation AI can apply a specific analysis algorithm to finance-related conversations. The generation AI can also apply a different analysis algorithm to health-related conversations. The generation AI can also apply a general analysis algorithm to everyday conversations. In this way, applying different analysis algorithms depending on the category of the conversation enables more accurate analysis. Some or all of the above-mentioned processing in the analysis unit can be performed using, for example, the generation AI, or can be performed without using the generation AI. For example, the analysis unit can input the category of the conversation into the generation AI and have the generation AI select the analysis algorithm.

[0043] The analysis unit can improve the accuracy of the analysis based on the user's past analysis results during analysis. For example, the analysis unit can improve the accuracy of the analysis by referring to the user's past analysis results during analysis. The analysis unit can analyze the user's past analysis results using a generation AI. The generation AI is a text generation AI (e.g., LLM) or a multimodal generation AI, and improves the accuracy of the analysis based on the past analysis results. For example, the generation AI adjusts the analysis algorithm based on the user's past analysis results. The generation AI can also learn specific patterns from the user's past analysis results to improve the analysis accuracy. The generation AI can also determine analysis priorities by referring to the user's past analysis results. In this way, the accuracy of the analysis can be improved by referring to the past analysis results. Some or all of the above-mentioned processing in the analysis unit may be performed using, for example, the generation AI, or may be performed without using the generation AI. For example, the analysis unit can input the user's past analysis results into the generation AI and have the generation AI improve the accuracy of the analysis.

[0044] The analysis unit can determine the analysis priority based on the date and time of the conversation during analysis. For example, the analysis unit determines the analysis priority based on the time of the conversation during analysis. The analysis unit can analyze the date and time of the conversation using a generation AI. The generation AI is a text generation AI (e.g., LLM) or a multimodal generation AI, and determines the analysis priority based on the date and time of the conversation. For example, the generation AI prioritizes analysis of recent conversations. The generation AI can also prioritize analysis of conversations that occurred during a specific time period. The generation AI can also determine the analysis priority by referring to past conversations. This enables efficient analysis by determining the analysis priority based on the time of the conversation. Some or all of the above-mentioned processing in the analysis unit may be performed using, for example, the generation AI, or may be performed without using the generation AI. For example, the analysis unit can input the date and time of the conversation into the generation AI and have the generation AI determine the analysis priority.

[0045] The analysis unit can adjust the order of analysis based on the relevance of the conversation during analysis. For example, the analysis unit adjusts the order of analysis based on the relevance of the conversation during analysis. The analysis unit can use a generation AI to evaluate the relevance of the conversation. The generation AI is a text generation AI (e.g., LLM) or a multimodal generation AI, and adjusts the order of analysis based on the relevance of the conversation. For example, the generation AI prioritizes analysis of highly relevant conversations. The generation AI can also postpone less relevant conversations. The generation AI can also adjust the order of analysis according to the relevance of the conversation. In this way, by adjusting the order of analysis based on the relevance of the conversation, important conversations can be prioritized for analysis. Some or all of the above-mentioned processing in the analysis unit may be performed using, for example, the generation AI, or may be performed without using the generation AI. For example, the analysis unit can input the relevance of the conversation to the generation AI and have the generation AI adjust the order of analysis.

[0046] The analysis unit can adjust the use of technical terms in the analysis according to the user's level of knowledge during analysis. For example, the analysis unit can adjust the use of technical terms in the analysis according to the user's level of expertise during analysis. The analysis unit can use a generation AI to evaluate the user's level of knowledge. The generation AI can be a text generation AI (e.g., LLM) or a multimodal generation AI, and adjust the use of technical terms in the analysis based on the user's level of knowledge. For example, if the user has technical knowledge, the generation AI uses a lot of technical terms. Also, if the user does not have technical knowledge, the generation AI can provide analysis results in simpler terms. The generation AI can also adjust the use of technical terms in the analysis according to the user's level of expertise. In this way, by adjusting the use of technical terms in the analysis according to the user's level of expertise, it is possible to provide analysis results that are easier to understand. Some or all of the above-mentioned processing in the analysis unit can be performed using, for example, a generation AI, or can be performed without using a generation AI. For example, the analysis unit can input the user's level of knowledge into the generation AI and have the generation AI adjust the use of technical terms in the analysis.

[0047] The notification unit can adjust the level of detail of the notification based on the priority of the possibility of fraud when issuing a notification. For example, the notification unit can adjust the level of detail of the notification based on the importance of the possibility of fraud when issuing a notification. The notification unit can use a generation AI to evaluate the possibility of fraud. The generation AI can be a text generation AI (e.g., LLM) or a multimodal generation AI, and adjust the level of detail of the notification based on the importance of the possibility of fraud. For example, if the generation AI determines that there is a high possibility of fraud, it can provide a detailed notification. Also, if the generation AI determines that there is a low possibility of fraud, it can provide a concise notification. The generation AI can also adjust the level of detail of the notification according to the importance of the possibility of fraud. In this way, adjusting the level of detail of the notification based on the importance of the possibility of fraud enables efficient notification. Some or all of the above-mentioned processing in the notification unit may be performed using, for example, the generation AI, or may be performed without using the generation AI. For example, the notification unit can input the possibility of fraud into the generation AI and have the generation AI adjust the level of detail of the notification.

[0048] The notification unit can apply different notification algorithms depending on the type of fraud when notifying. For example, the notification unit can apply different notification algorithms depending on the category of fraud when notifying. The notification unit can use a generation AI to classify the type of fraud. The generation AI can be a text generation AI (e.g., LLM) or a multimodal generation AI, and selects an appropriate notification algorithm depending on the type of fraud. For example, if the generation AI is a financial fraud, it can apply a specific notification algorithm. Also, if the generation AI is a health fraud, it can apply a different notification algorithm. Also, if the generation AI is an everyday fraud, it can apply a general notification algorithm. In this way, applying different notification algorithms depending on the category of fraud enables more accurate notifications. Some or all of the above-mentioned processing in the notification unit can be performed using, for example, the generation AI, or can be performed without using the generation AI. For example, the notification unit can input the type of fraud into the generation AI and have the generation AI select the notification algorithm.

[0049] The notification unit can improve the accuracy of notifications based on the user's past notification results when notifying the user. For example, the notification unit can improve the accuracy of notifications by referring to the user's past notification results when notifying the user. The notification unit can analyze the user's past notification results using a generation AI. The generation AI is a text generation AI (e.g., LLM) or a multimodal generation AI, and improves the accuracy of notifications based on the past notification results. For example, the generation AI adjusts the notification algorithm based on the user's past notification results. The generation AI can also learn specific patterns from the user's past notification results to improve the accuracy of notifications. The generation AI can also determine the priority of notifications by referring to the user's past notification results. In this way, the accuracy of notifications can be improved by referring to the past notification results. Some or all of the above-mentioned processing in the notification unit may be performed using, for example, the generation AI, or may be performed without using the generation AI. For example, the notification unit can input the user's past notification results into the generation AI and have the generation AI improve the accuracy of notifications.

[0050] The notification unit can determine the priority of notifications based on the date and time of fraud occurrence at the time of notification. The notification unit, for example, determines the priority of notifications based on the time of fraud occurrence at the time of notification. The notification unit can use a generation AI to analyze the date and time of fraud occurrence. The generation AI is a text generation AI (e.g., LLM) or a multimodal generation AI, and determines the priority of notifications based on the date and time of fraud occurrence. For example, if the generation AI determines that a fraud is likely to have occurred recently, it will prioritize notifications. Also, if the generation AI determines that a fraud is likely to have occurred during a specific time period, it can concentrate notifications during that time period. The generation AI can also determine the priority of notifications based on the time of past fraud occurrences. This enables efficient notifications by determining the priority of notifications based on the time of fraud occurrence. Some or all of the above-mentioned processing in the notification unit may be performed using, for example, the generation AI, or may be performed without using the generation AI. For example, the notification unit can input the date and time of fraud occurrence into the generation AI and have the generation AI determine the priority of notifications.

[0051] The notification unit can adjust the order of notifications based on the relevance of the fraud when notifying. For example, the notification unit adjusts the order of notifications based on the relevance of the fraud when notifying. The notification unit can use a generation AI to evaluate the relevance of the fraud. The generation AI is a text generation AI (e.g., LLM) or a multimodal generation AI, and adjusts the order of notifications based on the relevance of the fraud. For example, if the generation AI determines that a fraud is highly relevant, it will prioritize the notification. Also, if the generation AI determines that a fraud is less relevant, it can postpone the notification. The generation AI can also adjust the order of notifications based on the relevance of the fraud. In this way, by adjusting the order of notifications based on the relevance of the fraud, important notifications can be prioritized. Some or all of the above-mentioned processing in the notification unit may be performed using, for example, the generation AI, or may be performed without using the generation AI. For example, the notification unit can input the relevance of the fraud into the generation AI and have the generation AI adjust the order of the notifications.

[0052] The notification unit can adjust the use of technical terms in the notification according to the user's level of knowledge at the time of notification. For example, the notification unit can adjust the use of technical terms in the notification according to the user's level of expertise at the time of notification. The notification unit can use a generation AI to evaluate the user's level of knowledge. The generation AI can be a text generation AI (e.g., LLM) or a multimodal generation AI, and adjust the use of technical terms in the notification based on the user's level of knowledge. For example, if the user has technical knowledge, the generation AI can use a lot of technical terms. Also, if the user does not have technical knowledge, the generation AI can notify in simple language. The generation AI can also adjust the use of technical terms in the notification according to the user's level of expertise. This allows for notifications that are easier to understand by adjusting the use of technical terms in the notification according to the user's level of expertise. Some or all of the above-described processing in the notification unit can be performed using, for example, a generation AI, or can be performed without using a generation AI. For example, the notification unit can input the user's level of knowledge into the generation AI and cause the generation AI to adjust the use of technical terms in the notification.

[0053] The providing unit can adjust the level of detail of the alert provided based on the priority of the possibility of fraud when providing an alert. For example, the providing unit adjusts the level of detail of the alert provided based on the importance of the possibility of fraud when providing an alert. The providing unit can use a generation AI to evaluate the possibility of fraud. The generation AI is a text generation AI (e.g., LLM) or a multimodal generation AI, and adjusts the level of detail of the alert provided based on the importance of the possibility of fraud. For example, if the generation AI determines the possibility of fraud is high, it provides a detailed alert. Also, if the generation AI determines the possibility of fraud is low, it can provide a concise alert. The generation AI can also adjust the level of detail of the alert according to the importance of the possibility of fraud. In this way, adjusting the level of detail of the alert based on the importance of the possibility of fraud enables efficient alert provision. Some or all of the above-mentioned processing in the providing unit may be performed using, for example, the generation AI, or may be performed without using the generation AI. For example, the providing unit can input the possibility of fraud into the generation AI and cause the generation AI to adjust the level of detail of the alert provided.

[0054] The providing unit can apply different providing algorithms depending on the type of fraud when providing an alert. For example, when providing an alert, the providing unit applies different providing algorithms depending on the category of fraud. The providing unit can classify the type of fraud using a generation AI. The generation AI can be a text generation AI (e.g., LLM) or a multimodal generation AI, and selects an appropriate providing algorithm depending on the type of fraud. For example, if the generation AI is a financial fraud, it can apply a specific providing algorithm. Also, if the generation AI is a health fraud, it can apply a different providing algorithm. Also, if the generation AI is an everyday fraud, it can apply a general providing algorithm. In this way, applying different providing algorithms depending on the category of fraud enables more accurate alert provision. Some or all of the above-mentioned processing in the providing unit can be performed using, for example, the generation AI, or can be performed without using the generation AI. For example, the providing unit can input the type of fraud into the generation AI and have the generation AI select the providing algorithm.

[0055] The providing unit can improve the accuracy of the provision of an alert based on the user's past provision results when providing an alert. For example, when providing an alert, the providing unit improves the accuracy of the provision by referring to the user's past provision results. The providing unit can use the generation AI to analyze the user's past provision results. The generation AI is a text generation AI (e.g., LLM) or a multimodal generation AI, and improves the accuracy of the provision based on the past provision results. For example, the generation AI adjusts the provision algorithm based on the user's past provision results. The generation AI can also learn specific patterns from the user's past provision results and improve the provision accuracy. The generation AI can also determine provision priorities by referring to the user's past provision results. In this way, the accuracy of the provision can be improved by referring to the past provision results. Some or all of the above-described processing in the providing unit may be performed using, for example, the generation AI, or may be performed without using the generation AI. For example, the providing unit can input the user's past provision results into the generation AI and cause the generation AI to improve the accuracy of the provision.

[0056] The providing unit can determine the priority of alert provision based on the date and time of fraud occurrence when providing an alert. For example, the providing unit determines the priority of alert provision based on the time of fraud occurrence when providing an alert. The providing unit can use the generation AI to analyze the date and time of fraud occurrence. The generation AI is a text generation AI (e.g., LLM) or a multimodal generation AI, and determines the priority of alert provision based on the date and time of fraud occurrence. For example, if the generation AI determines that a fraud is likely to have occurred recently, it can provide an alert preferentially. Furthermore, if the generation AI determines that a fraud is likely to have occurred during a specific time period, it can also concentrate alerts on that time period. Furthermore, the generation AI can determine the priority of alerts by referring to the time of past fraud occurrences. This enables efficient alert provision by determining the priority of alert provision based on the time of fraud occurrence. Some or all of the above-described processing in the providing unit may be performed using, for example, the generation AI, or may be performed without using the generation AI. For example, the providing unit can input the date and time of fraud occurrence into the generation AI and have the generation AI determine the priority of alert provision.

[0057] The providing unit can adjust the order of providing alerts based on the relevance of the frauds when providing the alerts. For example, the providing unit can adjust the order of providing alerts based on the relevance of the frauds when providing the alerts. The providing unit can use a generation AI to evaluate the relevance of the frauds. The generation AI can be a text generation AI (e.g., LLM) or a multimodal generation AI, and adjust the order of providing alerts based on the relevance of the frauds. For example, if the generation AI determines that an alert is likely to be highly relevant, it can provide the alert preferentially. Also, if the generation AI determines that an alert is likely to be less relevant, it can postpone providing the alert. The generation AI can also adjust the order of providing alerts based on the relevance of the frauds. In this way, by adjusting the order of providing alerts based on the relevance of the frauds, important alerts can be provided preferentially. Some or all of the above-mentioned processing in the providing unit may be performed using, or without, the generation AI. For example, the providing unit can input the relevance of the frauds into the generation AI and cause the generation AI to adjust the order of providing alerts.

[0058] The providing unit can adjust the use of provided terminology according to the user's level of knowledge when providing an alert. For example, the providing unit adjusts the use of provided terminology according to the user's level of expertise when providing an alert. The providing unit can use a generation AI to evaluate the user's level of knowledge. The generation AI can be a text generation AI (e.g., LLM) or a multimodal generation AI, and adjusts the use of provided terminology based on the user's level of knowledge. For example, if the user has specialized knowledge, the generation AI uses a lot of specialized terminology. Also, if the user does not have specialized knowledge, the generation AI can provide an alert in simple language. The generation AI can also adjust the use of provided terminology according to the user's level of expertise. This makes it possible to provide alerts that are easier to understand by adjusting the use of provided terminology according to the user's level of expertise. Some or all of the above-described processing in the providing unit can be performed using, for example, the generation AI, or can be performed without using the generation AI. For example, the providing unit can input the user's level of knowledge into the generation AI and cause the generation AI to adjust the use of provided terminology.

[0059] The system according to the embodiment is not limited to the above-described example, and various modifications are possible, for example, as follows.

[0060] The acquisition unit can detect the user's current activity status and adjust the conversation acquisition method according to the activity. For example, if the user is exercising, the acquisition unit can set the conversation acquisition frequency low and acquire only important conversations. Alternatively, if the user is reading in a quiet environment, the acquisition unit can set the conversation acquisition frequency high and acquire all conversations in detail. Furthermore, if the user is in a meeting, the acquisition unit can temporarily stop conversation acquisition and resume it after the meeting ends. This allows for more appropriate conversation acquisition by adjusting the conversation acquisition method according to the user's activity status.

[0061] The analysis unit can understand the context of the conversation and assess the likelihood of fraud based on that context. For example, it can analyze the context before and after the conversation to determine whether specific keywords are used in a fraudulent context. It can also analyze the tone of the conversation and the speaker's intention to assess the likelihood of fraud. It can also detect inconsistencies or unnatural points in the conversation and assess the likelihood of fraud based on those. This allows for more accurate fraud detection by performing analysis that takes context into account.

[0062] The notification unit can select the optimal notification method based on the user's location information. For example, when the user is at home, the notification unit prioritizes voice notification. Also, when the user is out, the notification unit can prioritize text message or push notification. Furthermore, when the user is in a meeting, the notification unit can send a vibration notification and then send a detailed notification after the meeting ends. This allows for effective notification by selecting the optimal notification method based on the user's location information.

[0063] The alert delivery unit can analyze the user's past alert history and optimize the alert delivery method. For example, it can prioritize the use of an alert delivery method that the user has responded to quickly in the past. It can also improve the alert delivery method that the user has ignored in the past and select a more effective method. Furthermore, it can customize the alert delivery method according to specific time periods or situations based on the user's past alert history. This makes it possible to optimize the alert delivery method and provide effective alerts by referring to the past alert history.

[0064] When providing an alert, the providing unit can determine the priority of providing the alert based on the date and time of the fraud occurrence. For example, if there is a high possibility that the fraud occurred recently, the alert will be provided first. Also, if there is a high possibility that the fraud occurred during a specific time period, the alerts can be concentrated on that time period. Furthermore, the priority of the alert can be determined by referring to the time when past frauds occurred. In this way, by determining the priority of providing the alert based on the time when the fraud occurred, it is possible to provide alerts efficiently.

[0065] The processing flow of the first embodiment will be briefly explained below.

[0066] Step 1: The acquisition unit acquires the conversation content. The conversation content includes, for example, voice data, text data, and part or all of the conversation. The acquisition unit constantly listens to the owner's conversation and can detect phrases that may be related to tax refund fraud in the content of phone conversations and everyday conversations. Step 2: The analysis unit analyzes the conversation content acquired by the acquisition unit and determines whether or not there is any refund fraud. The analysis unit detects conversation content containing keywords such as "refund," "transfer," and "bank account," and uses the generation AI to analyze the conversation content to determine the possibility of refund fraud. Step 3: The notification unit sends an alert based on the results determined by the analysis unit. If a possible fraud is detected, the notification unit sends an alert to family members or the police. The notification unit can use generation AI to send alerts. Step 4: The sending unit provides the alert sent by the notification unit. The sending unit provides the alert to family members or the police. The sending unit can provide the alert using the generation AI.

[0067] (Example 2) A refund fraud detection system according to an embodiment of the present invention acquires conversation content, analyzes the possibility of refund fraud, and sends an alert. The refund fraud detection system acquires conversation content, analyzes the possibility of refund fraud, and sends an alert, thereby preventing fraud before it occurs. For example, the refund fraud detection system installs a dedicated app on a dedicated terminal or smartphone. This app is equipped with a generation AI and can constantly listen to the owner's conversations. The generation AI then analyzes the conversation content and determines whether there is a possibility of refund fraud. For example, if keywords such as "refund," "transfer," and "bank account" are included, the generation AI will evaluate the possibility of fraud. If a possible fraud is detected, an alert is sent to the owner's family or the police. For example, the dedicated app automatically notifies the owner's family or the police, allowing them to take measures before the owner is defrauded. In this way, the refund fraud detection system can prevent refund fraud before it occurs. For example, the system constantly listens to the owner's conversations, detects conversations that may be tax refund fraud, and sends an alert to family members or the police, allowing for swift countermeasures to be taken. Privacy is thoroughly protected, and when the generating AI analyzes the content of conversations, it is designed to prevent personal information from leaking to the outside. For example, the content of conversations is analyzed only within the device and is not sent to an external server. This system can prevent tax refund fraud before it occurs. The generating AI can anticipate countless possible scenarios for becoming a victim and quickly issue an alert, ensuring the safety of the owner.

[0068] A refund fraud detection system according to an embodiment includes an acquisition unit, an analysis unit, a notification unit, and a provision unit. The acquisition unit acquires conversation content. The conversation content may include, but is not limited to, voice data, text data, or a part or all of a conversation. The acquisition unit, for example, constantly listens to the owner's conversation. The acquisition unit may detect, for example, the content of the owner's phone conversation or phrases in everyday conversation that may be related to refund fraud. The analysis unit analyzes the conversation content acquired by the acquisition unit and determines whether or not refund fraud has occurred. The analysis unit detects conversation content containing keywords such as "refund," "transfer," and "bank account." The analysis unit analyzes the conversation content using a generation AI. The generation AI may be a text generation AI (e.g., LLM) or a multimodal generation AI, and determines the possibility of refund fraud based on the conversation content. The notification unit sends an alert based on the result determined by the analysis unit. For example, if the notification unit detects a possibility of fraud, it sends an alert to the owner's family or the police. The notification unit can send an alert using the generation AI. The provision unit provides the alert sent by the notification unit. The provision unit provides the alert, for example, to family members or the police. The provision unit can provide the alert using the generation AI. As a result, the refund fraud detection system according to the embodiment can prevent damage caused by refund fraud. For example, the system can constantly listen to the owner's conversations, detect conversations that may be related to refund fraud, and send alerts to family members or the police, allowing for prompt countermeasures to be taken. Privacy protection is thorough, and when the generation AI analyzes the content of the conversation, it is designed to prevent personal information from leaking to the outside. For example, the content of the conversation is analyzed only within the device and not sent to an external server. This mechanism can prevent damage caused by refund fraud. The generation AI can anticipate countless possible scenarios for victimization and quickly issue alerts, ensuring the owner's safety.

[0069] The acquisition unit can continuously listen to the owner's conversation. For example, the acquisition unit constantly listens to the owner's conversation. For example, the acquisition unit can detect phrases that may be tax refund fraud in the content of the owner's phone conversation or in everyday conversation. The acquisition unit can acquire the content of the conversation using a generation AI. The generation AI is a text generation AI (e.g., LLM) or a multimodal generation AI, and determines the possibility of tax refund fraud based on the content of the conversation. In this way, by constantly listening to the owner's conversation, the possibility of tax refund fraud can be quickly detected. Some or all of the above-mentioned processing in the acquisition unit may be performed using, or without, the generation AI. For example, the acquisition unit can input the owner's conversation into the generation AI and cause the generation AI to acquire the content of the conversation.

[0070] The analysis unit can detect conversation content containing keywords such as "refund," "transfer," and "bank account." The analysis unit can detect conversation content containing keywords such as "refund," "transfer," and "bank account." The analysis unit can analyze the conversation content using a generation AI. The generation AI is a text generation AI (e.g., LLM) or a multimodal generation AI, and determines the possibility of refund fraud based on the conversation content. For example, the generation AI detects conversation content containing keywords such as "refund," "transfer," and "bank account" and evaluates the possibility of fraud. In this way, by detecting conversation content containing specific keywords, the possibility of refund fraud can be evaluated highly. Some or all of the above-mentioned processing in the analysis unit can be performed using, for example, the generation AI, or can be performed without using the generation AI. For example, the analysis unit can input the conversation content into the generation AI and have the generation AI analyze the possibility of refund fraud.

[0071] The notification unit can send an alert to family or the police when fraud is detected. For example, the notification unit sends an alert to family or the police when the possibility of fraud is detected. The notification unit can send the alert using a generation AI. The generation AI is a text generation AI (e.g., LLM) or a multimodal generation AI, and sends the alert. For example, the generation AI detects the possibility of fraud and sends an alert to family or the police. This makes it possible to prevent damage by quickly sending an alert to family or the police when the possibility of fraud is detected. Some or all of the above-mentioned processing in the notification unit may be performed using, or without, the generation AI. For example, the notification unit can input the possibility of fraud into the generation AI and have the generation AI send an alert.

[0072] The providing unit can send the alert to the family or the police. The providing unit, for example, provides the alert to the family or the police. The providing unit can provide the alert using a generation AI. The generation AI is a text generation AI (e.g., LLM) or a multimodal generation AI, and provides the alert. For example, the generation AI provides the alert to the family or the police. By providing the alert to the family or the police, a prompt response is possible. Some or all of the above-mentioned processing in the providing unit may be performed, for example, using the generation AI, or may be performed without using the generation AI. For example, the providing unit can input the alert to the generation AI and have the generation AI provide the alert.

[0073] The analysis unit can analyze the conversation content only within the terminal and not send it to an external server. For example, the analysis unit can analyze the conversation content only within the terminal and not send it to an external server. The analysis unit can analyze the conversation content using a generation AI. The generation AI can be a text generation AI (e.g., LLM) or a multimodal generation AI, and determine the possibility of tax refund fraud based on the conversation content. For example, the generation AI analyzes the conversation content within the terminal and does not send it to an external server. By analyzing the conversation content only within the terminal, privacy is thoroughly protected. Some or all of the above-mentioned processing in the analysis unit can be performed, for example, using the generation AI, or can be performed without using the generation AI. For example, the analysis unit can input the conversation content to the generation AI and have the generation AI analyze the conversation content.

[0074] The acquisition unit can estimate the user's emotions and adjust the conversation acquisition timing based on the estimated user's emotions. For example, the acquisition unit can estimate the user's emotions and adjust the conversation acquisition timing based on the estimated user's emotions. The acquisition unit can estimate the user's emotions using a generation AI. The generation AI is a text generation AI (e.g., LLM) or a multimodal generation AI, and adjusts the conversation acquisition timing based on the user's emotions. For example, the generation AI can estimate the user's emotions and, if the user is relaxed, set the conversation acquisition frequency low and acquire only important conversations. Furthermore, the generation AI can estimate the user's emotions and, if the user is nervous, set the conversation acquisition frequency high and acquire all conversations in detail. Furthermore, the generation AI can estimate the user's emotions and, if the user is in a hurry, temporarily stop conversation acquisition and resume it later. This allows for more appropriate conversation acquisition by adjusting the conversation acquisition timing according to the user's emotions. Some or all of the above-described processing in the acquisition unit can be performed, for example, using the generation AI, or can be performed without using the generation AI. For example, the acquisition unit can input the user's emotions into the generation AI and cause the generation AI to adjust the timing of conversation acquisition.

[0075] The acquisition unit can analyze the user's past conversation history and select an appropriate acquisition method. For example, the acquisition unit analyzes the user's past conversation history and selects the optimal acquisition method. The acquisition unit can analyze the user's past conversation history using a generation AI. The generation AI is a text generation AI (e.g., LLM) or a multimodal generation AI, and selects the optimal acquisition method based on the user's past conversation history. For example, the generation AI analyzes the user's past conversation history and prioritizes acquisition of frequently used phrases. The generation AI can also analyze the user's past conversation history to find a tendency for important conversations to occur during specific time periods and focus acquisition on those time periods. The generation AI can also prioritize acquisition of conversations with specific people based on the user's past conversation history. This allows the analysis of the past conversation history to select the optimal acquisition method, enabling efficient conversation acquisition. Some or all of the above-mentioned processing in the acquisition unit may be performed using, for example, the generation AI, or may be performed without using the generation AI. For example, the acquisition unit can input the user's past conversation history into the generation AI and have the generation AI select the optimal acquisition method.

[0076] The acquisition unit can filter the conversations based on the user's current situation or areas of interest when acquiring the conversations. For example, the acquisition unit can filter the conversations based on the user's current situation or areas of interest when acquiring the conversations. The acquisition unit can analyze the user's current situation or areas of interest using a generation AI. The generation AI is a text generation AI (e.g., LLM) or a multimodal generation AI, and filters the conversations based on the user's current situation or areas of interest. For example, the generation AI can analyze the user's current situation and acquire only important conversations. The generation AI can also analyze the user's areas of interest and prioritize acquisition of related conversations. Furthermore, if the user is in a specific location, the generation AI can prioritize acquisition of conversations related to that location. In this way, by filtering the conversations based on the user's situation or areas of interest, important conversations can be prioritized and acquired. Some or all of the above-described processing in the acquisition unit can be performed using, for example, the generation AI, or can be performed without using the generation AI. For example, the acquisition unit can input the user's current situation or areas of interest to the generation AI and have the generation AI perform conversation filtering.

[0077] The acquisition unit can select an appropriate acquisition means according to the user's input method when acquiring a conversation. For example, the acquisition unit selects the optimal acquisition means according to the user's input method (voice, text, image, etc.) when acquiring a conversation. The acquisition unit can analyze the user's input method using a generation AI. The generation AI is a text generation AI (e.g., LLM) or a multimodal generation AI, and selects the optimal acquisition means based on the user's input method. For example, if the generation AI is using voice input, it can prioritize acquiring voice data. Also, if the user is using text input, it can prioritize acquiring text data. Also, if the user is using image input, it can prioritize acquiring image data. This enables efficient conversation acquisition by selecting the optimal acquisition means according to the user's input method. Some or all of the above-mentioned processing in the acquisition unit can be performed using, for example, the generation AI, or can be performed without using the generation AI. For example, the acquisition unit can input the user's input method to the generation AI and cause the generation AI to select the optimal acquisition means.

[0078] The acquisition unit can estimate the user's emotions and determine the priority of conversations to be acquired based on the estimated user's emotions. The acquisition unit, for example, estimates the user's emotions and determines the priority of conversations to be acquired based on the estimated user's emotions. The acquisition unit can estimate the user's emotions using a generation AI. The generation AI is a text generation AI (e.g., LLM) or a multimodal generation AI, and determines the priority of conversations based on the user's emotions. For example, the generation AI estimates the user's emotions and, if the user is relaxed, postpones conversations of lower importance. Furthermore, the generation AI can estimate the user's emotions and, if the user is nervous, prioritize acquiring conversations of higher importance. Furthermore, the generation AI estimates the user's emotions and, if the user is in a hurry, prioritize acquiring conversations containing important information in a short time. In this way, by determining the priority of conversations based on the user's emotions, important conversations can be prioritized and acquired. Some or all of the above-mentioned processing in the acquisition unit may be performed, for example, using the generation AI, or may be performed without using the generation AI. For example, the acquisition unit can input the user's emotions into the generation AI and have the generation AI determine the priority of the conversation.

[0079] The acquisition unit can prioritize acquisition of highly relevant conversations based on the user's geographical location information when acquiring a conversation. For example, the acquisition unit prioritizes acquisition of highly relevant conversations by taking the user's geographical location information into consideration when acquiring a conversation. The acquisition unit can analyze the user's geographical location information using a generation AI. The generation AI is a text generation AI (e.g., LLM) or a multimodal generation AI, and prioritizes acquisition of highly relevant conversations based on the user's geographical location information. For example, when the user is in a specific location, the generation AI can prioritize acquisition of conversations related to that location. Furthermore, when the user is traveling, the generation AI can prioritize acquisition of conversations related to the user's destination. Furthermore, when the user is in a specific area, the generation AI can prioritize acquisition of conversations related to that area. In this way, by taking the user's geographical location information into consideration, highly relevant conversations can be prioritized. Some or all of the above-described processing in the acquisition unit may be performed using the generation AI, for example, or may be performed without using the generation AI. For example, the acquisition unit can input the user's geographical location information to the generation AI and cause the generation AI to acquire highly relevant conversations.

[0080] The acquisition unit can analyze the user's social media usage status and acquire related conversations when acquiring a conversation. For example, the acquisition unit can analyze the user's social media activity and acquire related conversations when acquiring a conversation. The acquisition unit can analyze the user's social media activity using a generation AI. The generation AI is a text generation AI (e.g., LLM) or a multimodal generation AI, and acquires related conversations based on the user's social media activity. For example, the generation AI can prioritize acquiring conversations related to places the user has checked in to on social media. The generation AI can also analyze the content of the user's social media posts and acquire related conversations. The generation AI can also acquire related conversations by referring to the activities of the user's friends on social media. In this way, related conversations can be efficiently acquired by analyzing social media activity. Some or all of the above-mentioned processing in the acquisition unit can be performed using, for example, the generation AI, or can be performed without using the generation AI. For example, the acquisition unit can input the user's social media activity into the generation AI and cause the generation AI to acquire related conversations.

[0081] The acquisition unit can adjust the acquisition method by reflecting the user's past feedback when acquiring a conversation. For example, the acquisition unit customizes the acquisition method by reflecting the user's past feedback when acquiring a conversation. The acquisition unit can analyze the user's past feedback using a generation AI. The generation AI is a text generation AI (e.g., LLM) or a multimodal generation AI, and adjusts the acquisition method based on the user's past feedback. For example, the generation AI prioritizes acquisition of conversations that the user previously evaluated as important. The generation AI can also adjust the acquisition method based on the user's past feedback. The generation AI can also exclude and acquire conversations that the user previously evaluated as unnecessary. In this way, by reflecting past feedback, the acquisition method can be customized, enabling efficient conversation acquisition. Some or all of the above-mentioned processing in the acquisition unit may be performed using, for example, the generation AI, or may be performed without using the generation AI. For example, the acquisition unit can input the user's past feedback into the generation AI and cause the generation AI to adjust the acquisition method.

[0082] The analysis unit can estimate the user's emotions and adjust the presentation method of the analysis based on the estimated user's emotions. For example, the analysis unit can estimate the user's emotions and adjust the presentation method of the analysis based on the estimated user's emotions. The analysis unit can estimate the user's emotions using a generation AI. The generation AI can be a text generation AI (e.g., LLM) or a multimodal generation AI, and adjust the presentation method of the analysis based on the user's emotions. For example, the generation AI can estimate the user's emotions and provide detailed analysis results if the user is relaxed. Furthermore, the generation AI can estimate the user's emotions and provide concise and to-the-point analysis results if the user is nervous. Furthermore, the generation AI can estimate the user's emotions and provide quick analysis results if the user is in a hurry. This allows for more appropriate analysis results to be provided by adjusting the presentation method of the analysis based on the user's emotions. Some or all of the above-described processing in the analysis unit may be performed using, or without, the generation AI. For example, the analysis unit can input the user's emotions into the generation AI and have the generation AI adjust the presentation method of the analysis.

[0083] The analysis unit can adjust the level of analysis detail based on the priority of the conversation during analysis. For example, the analysis unit can adjust the level of analysis detail based on the importance of the conversation during analysis. The analysis unit can use a generation AI to evaluate the importance of the conversation. The generation AI is a text generation AI (e.g., LLM) or a multimodal generation AI, and adjusts the level of analysis detail based on the importance of the conversation. For example, the generation AI performs a detailed analysis of a conversation that is highly important. The generation AI can also perform a concise analysis of a conversation that is less important. The generation AI can also determine the priority of the analysis according to the importance of the conversation. This enables efficient analysis by adjusting the level of analysis detail based on the importance of the conversation. Some or all of the above-mentioned processing in the analysis unit may be performed using, or without, the generation AI. For example, the analysis unit can input the importance of the conversation to the generation AI and have the generation AI adjust the level of analysis detail.

[0084] The analysis unit can apply different analysis algorithms depending on the type of conversation during analysis. For example, the analysis unit can apply different analysis algorithms depending on the category of the conversation during analysis. The analysis unit can classify the category of the conversation using a generation AI. The generation AI can be a text generation AI (e.g., LLM) or a multimodal generation AI, and selects an appropriate analysis algorithm depending on the category of the conversation. For example, the generation AI can apply a specific analysis algorithm to finance-related conversations. The generation AI can also apply a different analysis algorithm to health-related conversations. The generation AI can also apply a general analysis algorithm to everyday conversations. In this way, applying different analysis algorithms depending on the category of the conversation enables more accurate analysis. Some or all of the above-mentioned processing in the analysis unit can be performed using, for example, the generation AI, or can be performed without using the generation AI. For example, the analysis unit can input the category of the conversation into the generation AI and have the generation AI select the analysis algorithm.

[0085] The analysis unit can improve the accuracy of the analysis based on the user's past analysis results during analysis. For example, the analysis unit can improve the accuracy of the analysis by referring to the user's past analysis results during analysis. The analysis unit can analyze the user's past analysis results using a generation AI. The generation AI is a text generation AI (e.g., LLM) or a multimodal generation AI, and improves the accuracy of the analysis based on the past analysis results. For example, the generation AI adjusts the analysis algorithm based on the user's past analysis results. The generation AI can also learn specific patterns from the user's past analysis results to improve the analysis accuracy. The generation AI can also determine analysis priorities by referring to the user's past analysis results. In this way, the accuracy of the analysis can be improved by referring to the past analysis results. Some or all of the above-mentioned processing in the analysis unit may be performed using, for example, the generation AI, or may be performed without using the generation AI. For example, the analysis unit can input the user's past analysis results into the generation AI and have the generation AI improve the accuracy of the analysis.

[0086] The analysis unit can estimate the user's emotions and adjust the length of the analysis based on the estimated user's emotions. For example, the analysis unit can estimate the user's emotions and adjust the length of the analysis based on the estimated user's emotions. The analysis unit can estimate the user's emotions using a generation AI. The generation AI can be a text generation AI (e.g., LLM) or a multimodal generation AI, and adjust the length of the analysis based on the user's emotions. For example, the generation AI can estimate the user's emotions and perform a detailed analysis if the user is relaxed. The generation AI can also estimate the user's emotions and perform a concise analysis if the user is nervous. The generation AI can also estimate the user's emotions and perform a quick analysis if the user is in a hurry. This allows for more appropriate analysis results by adjusting the length of the analysis based on the user's emotions. Some or all of the above-described processing in the analysis unit can be performed using, or without, the generation AI. For example, the analysis unit can input the user's emotions into the generation AI and have the generation AI adjust the length of the analysis.

[0087] The analysis unit can determine the analysis priority based on the date and time of the conversation during analysis. For example, the analysis unit determines the analysis priority based on the time of the conversation during analysis. The analysis unit can analyze the date and time of the conversation using a generation AI. The generation AI is a text generation AI (e.g., LLM) or a multimodal generation AI, and determines the analysis priority based on the date and time of the conversation. For example, the generation AI prioritizes analysis of recent conversations. The generation AI can also prioritize analysis of conversations that occurred during a specific time period. The generation AI can also determine the analysis priority by referring to past conversations. This enables efficient analysis by determining the analysis priority based on the time of the conversation. Some or all of the above-mentioned processing in the analysis unit may be performed using, for example, the generation AI, or may be performed without using the generation AI. For example, the analysis unit can input the date and time of the conversation into the generation AI and have the generation AI determine the analysis priority.

[0088] The analysis unit can adjust the order of analysis based on the relevance of the conversation during analysis. For example, the analysis unit adjusts the order of analysis based on the relevance of the conversation during analysis. The analysis unit can use a generation AI to evaluate the relevance of the conversation. The generation AI is a text generation AI (e.g., LLM) or a multimodal generation AI, and adjusts the order of analysis based on the relevance of the conversation. For example, the generation AI prioritizes analysis of highly relevant conversations. The generation AI can also postpone less relevant conversations. The generation AI can also adjust the order of analysis according to the relevance of the conversation. In this way, by adjusting the order of analysis based on the relevance of the conversation, important conversations can be prioritized for analysis. Some or all of the above-mentioned processing in the analysis unit may be performed using, for example, the generation AI, or may be performed without using the generation AI. For example, the analysis unit can input the relevance of the conversation to the generation AI and have the generation AI adjust the order of analysis.

[0089] The analysis unit can adjust the use of technical terms in the analysis according to the user's level of knowledge during analysis. For example, the analysis unit can adjust the use of technical terms in the analysis according to the user's level of expertise during analysis. The analysis unit can use a generation AI to evaluate the user's level of knowledge. The generation AI can be a text generation AI (e.g., LLM) or a multimodal generation AI, and adjust the use of technical terms in the analysis based on the user's level of knowledge. For example, if the user has technical knowledge, the generation AI uses a lot of technical terms. Also, if the user does not have technical knowledge, the generation AI can provide analysis results in simpler terms. The generation AI can also adjust the use of technical terms in the analysis according to the user's level of expertise. In this way, by adjusting the use of technical terms in the analysis according to the user's level of expertise, it is possible to provide analysis results that are easier to understand. Some or all of the above-mentioned processing in the analysis unit can be performed using, for example, a generation AI, or can be performed without using a generation AI. For example, the analysis unit can input the user's level of knowledge into the generation AI and have the generation AI adjust the use of technical terms in the analysis.

[0090] The notification unit can estimate the user's emotions and adjust the notification method based on the estimated user's emotions. For example, the notification unit can estimate the user's emotions and adjust the notification method based on the estimated user's emotions. The notification unit can estimate the user's emotions using a generation AI. The generation AI can be a text generation AI (e.g., LLM) or a multimodal generation AI, and adjust the notification method based on the user's emotions. For example, the generation AI can estimate the user's emotions and provide a detailed notification if the user is relaxed. The generation AI can also estimate the user's emotions and provide a concise notification if the user is nervous. The generation AI can also estimate the user's emotions and provide a quick notification if the user is in a hurry. This enables more appropriate notifications by adjusting the notification method based on the user's emotions. Some or all of the above-mentioned processing in the notification unit may be performed using, or without, the generation AI. For example, the notification unit can input the user's emotions into the generation AI and have the generation AI adjust the notification method.

[0091] The notification unit can adjust the level of detail of the notification based on the priority of the possibility of fraud when issuing a notification. For example, the notification unit can adjust the level of detail of the notification based on the importance of the possibility of fraud when issuing a notification. The notification unit can use a generation AI to evaluate the possibility of fraud. The generation AI can be a text generation AI (e.g., LLM) or a multimodal generation AI, and adjust the level of detail of the notification based on the importance of the possibility of fraud. For example, if the generation AI determines that there is a high possibility of fraud, it can provide a detailed notification. Also, if the generation AI determines that there is a low possibility of fraud, it can provide a concise notification. The generation AI can also adjust the level of detail of the notification according to the importance of the possibility of fraud. In this way, adjusting the level of detail of the notification based on the importance of the possibility of fraud enables efficient notification. Some or all of the above-mentioned processing in the notification unit may be performed using, for example, the generation AI, or may be performed without using the generation AI. For example, the notification unit can input the possibility of fraud into the generation AI and have the generation AI adjust the level of detail of the notification.

[0092] The notification unit can apply different notification algorithms depending on the type of fraud when notifying. For example, the notification unit can apply different notification algorithms depending on the category of fraud when notifying. The notification unit can use a generation AI to classify the type of fraud. The generation AI can be a text generation AI (e.g., LLM) or a multimodal generation AI, and selects an appropriate notification algorithm depending on the type of fraud. For example, if the generation AI is a financial fraud, it can apply a specific notification algorithm. Also, if the generation AI is a health fraud, it can apply a different notification algorithm. Also, if the generation AI is an everyday fraud, it can apply a general notification algorithm. In this way, applying different notification algorithms depending on the category of fraud enables more accurate notifications. Some or all of the above-mentioned processing in the notification unit can be performed using, for example, the generation AI, or can be performed without using the generation AI. For example, the notification unit can input the type of fraud into the generation AI and have the generation AI select the notification algorithm.

[0093] The notification unit can improve the accuracy of notifications based on the user's past notification results when notifying the user. For example, the notification unit can improve the accuracy of notifications by referring to the user's past notification results when notifying the user. The notification unit can analyze the user's past notification results using a generation AI. The generation AI is a text generation AI (e.g., LLM) or a multimodal generation AI, and improves the accuracy of notifications based on the past notification results. For example, the generation AI adjusts the notification algorithm based on the user's past notification results. The generation AI can also learn specific patterns from the user's past notification results to improve the accuracy of notifications. The generation AI can also determine the priority of notifications by referring to the user's past notification results. In this way, the accuracy of notifications can be improved by referring to the past notification results. Some or all of the above-mentioned processing in the notification unit may be performed using, for example, the generation AI, or may be performed without using the generation AI. For example, the notification unit can input the user's past notification results into the generation AI and have the generation AI improve the accuracy of notifications.

[0094] The notification unit can estimate the user's emotions and adjust the length of the notification based on the estimated user's emotions. For example, the notification unit can estimate the user's emotions and adjust the length of the notification based on the estimated user's emotions. The notification unit can estimate the user's emotions using a generation AI. The generation AI can be a text generation AI (e.g., LLM) or a multimodal generation AI, and adjust the length of the notification based on the user's emotions. For example, the generation AI can estimate the user's emotions and provide a detailed notification if the user is relaxed. The generation AI can also estimate the user's emotions and provide a concise notification if the user is nervous. The generation AI can also estimate the user's emotions and provide a quick notification if the user is in a hurry. This allows for more appropriate notifications by adjusting the length of the notification based on the user's emotions. Some or all of the above-described processing in the notification unit can be performed using, or without, the generation AI. For example, the notification unit can input the user's emotions into the generation AI and have the generation AI adjust the length of the notification.

[0095] The notification unit can determine the priority of notifications based on the date and time of fraud occurrence at the time of notification. The notification unit, for example, determines the priority of notifications based on the time of fraud occurrence at the time of notification. The notification unit can use a generation AI to analyze the date and time of fraud occurrence. The generation AI is a text generation AI (e.g., LLM) or a multimodal generation AI, and determines the priority of notifications based on the date and time of fraud occurrence. For example, if the generation AI determines that a fraud is likely to have occurred recently, it will prioritize notifications. Also, if the generation AI determines that a fraud is likely to have occurred during a specific time period, it can concentrate notifications during that time period. The generation AI can also determine the priority of notifications based on the time of past fraud occurrences. This enables efficient notifications by determining the priority of notifications based on the time of fraud occurrence. Some or all of the above-mentioned processing in the notification unit may be performed using, for example, the generation AI, or may be performed without using the generation AI. For example, the notification unit can input the date and time of fraud occurrence into the generation AI and have the generation AI determine the priority of notifications.

[0096] The notification unit can adjust the order of notifications based on the relevance of the fraud when notifying. For example, the notification unit adjusts the order of notifications based on the relevance of the fraud when notifying. The notification unit can use a generation AI to evaluate the relevance of the fraud. The generation AI is a text generation AI (e.g., LLM) or a multimodal generation AI, and adjusts the order of notifications based on the relevance of the fraud. For example, if the generation AI determines that a fraud is highly relevant, it will prioritize the notification. Also, if the generation AI determines that a fraud is less relevant, it can postpone the notification. The generation AI can also adjust the order of notifications based on the relevance of the fraud. In this way, by adjusting the order of notifications based on the relevance of the fraud, important notifications can be prioritized. Some or all of the above-mentioned processing in the notification unit may be performed using, for example, the generation AI, or may be performed without using the generation AI. For example, the notification unit can input the relevance of the fraud into the generation AI and have the generation AI adjust the order of the notifications.

[0097] The notification unit can adjust the use of technical terms in the notification according to the user's level of knowledge at the time of notification. For example, the notification unit can adjust the use of technical terms in the notification according to the user's level of expertise at the time of notification. The notification unit can use a generation AI to evaluate the user's level of knowledge. The generation AI can be a text generation AI (e.g., LLM) or a multimodal generation AI, and adjust the use of technical terms in the notification based on the user's level of knowledge. For example, if the user has technical knowledge, the generation AI can use a lot of technical terms. Also, if the user does not have technical knowledge, the generation AI can notify in simple language. The generation AI can also adjust the use of technical terms in the notification according to the user's level of expertise. This allows for notifications that are easier to understand by adjusting the use of technical terms in the notification according to the user's level of expertise. Some or all of the above-described processing in the notification unit can be performed using, for example, a generation AI, or can be performed without using a generation AI. For example, the notification unit can input the user's level of knowledge into the generation AI and cause the generation AI to adjust the use of technical terms in the notification.

[0098] The providing unit can estimate the user's emotions and adjust the alert delivery method based on the estimated user's emotions. For example, the providing unit can estimate the user's emotions and adjust the alert delivery method based on the estimated user's emotions. The providing unit can estimate the user's emotions using a generation AI. The generation AI is a text generation AI (e.g., LLM) or a multimodal generation AI, and adjusts the alert delivery method based on the user's emotions. For example, the generation AI can estimate the user's emotions and provide a detailed alert if the user is relaxed. The generation AI can also estimate the user's emotions and provide a concise alert if the user is nervous. The generation AI can also estimate the user's emotions and provide a quick alert if the user is in a hurry. This allows for more appropriate alert delivery by adjusting the alert delivery method based on the user's emotions. Some or all of the above-described processing in the providing unit may be performed using, or without, the generation AI. For example, the providing unit can input the user's emotions into the generation AI and cause the generation AI to adjust the alert delivery method.

[0099] The providing unit can adjust the level of detail of the alert provided based on the priority of the possibility of fraud when providing an alert. For example, the providing unit adjusts the level of detail of the alert provided based on the importance of the possibility of fraud when providing an alert. The providing unit can use a generation AI to evaluate the possibility of fraud. The generation AI is a text generation AI (e.g., LLM) or a multimodal generation AI, and adjusts the level of detail of the alert provided based on the importance of the possibility of fraud. For example, if the generation AI determines the possibility of fraud is high, it provides a detailed alert. Also, if the generation AI determines the possibility of fraud is low, it can provide a concise alert. The generation AI can also adjust the level of detail of the alert according to the importance of the possibility of fraud. In this way, adjusting the level of detail of the alert based on the importance of the possibility of fraud enables efficient alert provision. Some or all of the above-mentioned processing in the providing unit may be performed using, for example, the generation AI, or may be performed without using the generation AI. For example, the providing unit can input the possibility of fraud into the generation AI and cause the generation AI to adjust the level of detail of the alert provided.

[0100] The providing unit can apply different providing algorithms depending on the type of fraud when providing an alert. For example, when providing an alert, the providing unit applies different providing algorithms depending on the category of fraud. The providing unit can classify the type of fraud using a generation AI. The generation AI can be a text generation AI (e.g., LLM) or a multimodal generation AI, and selects an appropriate providing algorithm depending on the type of fraud. For example, if the generation AI is a financial fraud, it can apply a specific providing algorithm. Also, if the generation AI is a health fraud, it can apply a different providing algorithm. Also, if the generation AI is an everyday fraud, it can apply a general providing algorithm. In this way, applying different providing algorithms depending on the category of fraud enables more accurate alert provision. Some or all of the above-mentioned processing in the providing unit can be performed using, for example, the generation AI, or can be performed without using the generation AI. For example, the providing unit can input the type of fraud into the generation AI and have the generation AI select the providing algorithm.

[0101] The providing unit can improve the accuracy of the provision of an alert based on the user's past provision results when providing an alert. For example, when providing an alert, the providing unit improves the accuracy of the provision by referring to the user's past provision results. The providing unit can use the generation AI to analyze the user's past provision results. The generation AI is a text generation AI (e.g., LLM) or a multimodal generation AI, and improves the accuracy of the provision based on the past provision results. For example, the generation AI adjusts the provision algorithm based on the user's past provision results. The generation AI can also learn specific patterns from the user's past provision results and improve the provision accuracy. The generation AI can also determine provision priorities by referring to the user's past provision results. In this way, the accuracy of the provision can be improved by referring to the past provision results. Some or all of the above-described processing in the providing unit may be performed using, for example, the generation AI, or may be performed without using the generation AI. For example, the providing unit can input the user's past provision results into the generation AI and cause the generation AI to improve the accuracy of the provision.

[0102] The providing unit can estimate the user's emotions and adjust the length of the alert provided based on the estimated user's emotions. For example, the providing unit can estimate the user's emotions and adjust the length of the alert provided based on the estimated user's emotions. The providing unit can estimate the user's emotions using a generation AI. The generation AI is a text generation AI (e.g., LLM) or a multimodal generation AI, and adjusts the length of the alert provided based on the user's emotions. For example, the generation AI can estimate the user's emotions and provide a detailed alert if the user is relaxed. The generation AI can also estimate the user's emotions and provide a concise alert if the user is nervous. The generation AI can also estimate the user's emotions and provide a quick alert if the user is in a hurry. This enables more appropriate alert provision by adjusting the length of the alert provided based on the user's emotions. Some or all of the above-described processing in the providing unit may be performed using, or without, the generation AI. For example, the providing unit can input the user's emotions into the generation AI and cause the generation AI to adjust the length of the alert provided.

[0103] The providing unit can determine the priority of alert provision based on the date and time of fraud occurrence when providing an alert. For example, the providing unit determines the priority of alert provision based on the time of fraud occurrence when providing an alert. The providing unit can use the generation AI to analyze the date and time of fraud occurrence. The generation AI is a text generation AI (e.g., LLM) or a multimodal generation AI, and determines the priority of alert provision based on the date and time of fraud occurrence. For example, if the generation AI determines that a fraud is likely to have occurred recently, it can provide an alert preferentially. Furthermore, if the generation AI determines that a fraud is likely to have occurred during a specific time period, it can also concentrate alerts on that time period. Furthermore, the generation AI can determine the priority of alerts by referring to the time of past fraud occurrences. This enables efficient alert provision by determining the priority of alert provision based on the time of fraud occurrence. Some or all of the above-described processing in the providing unit may be performed using, for example, the generation AI, or may be performed without using the generation AI. For example, the providing unit can input the date and time of fraud occurrence into the generation AI and have the generation AI determine the priority of alert provision.

[0104] The providing unit can adjust the order of providing alerts based on the relevance of the frauds when providing the alerts. For example, the providing unit can adjust the order of providing alerts based on the relevance of the frauds when providing the alerts. The providing unit can use a generation AI to evaluate the relevance of the frauds. The generation AI can be a text generation AI (e.g., LLM) or a multimodal generation AI, and adjust the order of providing alerts based on the relevance of the frauds. For example, if the generation AI determines that an alert is likely to be highly relevant, it can provide the alert preferentially. Also, if the generation AI determines that an alert is likely to be less relevant, it can postpone providing the alert. The generation AI can also adjust the order of providing alerts based on the relevance of the frauds. In this way, by adjusting the order of providing alerts based on the relevance of the frauds, important alerts can be provided preferentially. Some or all of the above-mentioned processing in the providing unit may be performed using, or without, the generation AI. For example, the providing unit can input the relevance of the frauds into the generation AI and cause the generation AI to adjust the order of providing alerts.

[0105] The providing unit can adjust the use of provided terminology according to the user's level of knowledge when providing an alert. For example, the providing unit adjusts the use of provided terminology according to the user's level of expertise when providing an alert. The providing unit can use a generation AI to evaluate the user's level of knowledge. The generation AI can be a text generation AI (e.g., LLM) or a multimodal generation AI, and adjusts the use of provided terminology based on the user's level of knowledge. For example, if the user has specialized knowledge, the generation AI uses a lot of specialized terminology. Also, if the user does not have specialized knowledge, the generation AI can provide an alert in simple language. The generation AI can also adjust the use of provided terminology according to the user's level of expertise. This makes it possible to provide alerts that are easier to understand by adjusting the use of provided terminology according to the user's level of expertise. Some or all of the above-described processing in the providing unit can be performed using, for example, the generation AI, or can be performed without using the generation AI. For example, the providing unit can input the user's level of knowledge into the generation AI and cause the generation AI to adjust the use of provided terminology. === Hard Collateral 1-1 === Each of the multiple elements including the acquisition unit, analysis unit, notification unit, and provision unit described above is realized, for example, by at least one of the smart device 14 and the data processing device 12. For example, the acquisition unit acquires conversation content using the microphone 38B of the smart device 14 and transmits it to the data processing device 12 by the control unit 46A. The analysis unit is realized by the specific processing unit 290 of the data processing device 12 and analyzes the acquired conversation content to determine the possibility of refund fraud. The notification unit is realized by the specific processing unit 290 of the data processing device 12 and generates an alert when a possibility of fraud is detected. The provision unit is realized by the control unit 46A of the smart device 14 and provides the generated alert to family members or the police. === Hard Collateral 1-2 === Each of the multiple elements, including the acquisition unit, analysis unit, notification unit, and provision unit, described above, is realized, for example, by at least one of the smart glasses 214 and the data processing device 12. For example, the acquisition unit acquires conversation content using the microphone 238 of the smart glasses 214 and transmits it to the data processing device 12 by the control unit 46A. The analysis unit is realized by the specific processing unit 290 of the data processing device 12 and analyzes the acquired conversation content to determine the possibility of refund fraud. The notification unit is realized by the specific processing unit 290 of the data processing device 12 and generates an alert when a possibility of fraud is detected. The provision unit is realized by the control unit 46A of the smart glasses 214 and provides the generated alert to family members or the police. === Hard Collateral 1-3 === Each of the multiple elements including the above-mentioned acquisition unit, analysis unit, notification unit, and provision unit is realized, for example, in at least one of the headset type terminal 314 and the data processing device 12. For example, the acquisition unit acquires the conversation content using the microphone 238 of the headset type terminal 314 and transmits it to the data processing device 12 by the control unit 46A. The analysis unit is realized by the specific processing unit 290 of the data processing device 12 and analyzes the acquired conversation content to determine the possibility of refund fraud. The notification unit is realized by the specific processing unit 290 of the data processing device 12 and generates an alert when the possibility of fraud is detected. The provision unit is realized by the control unit 46A of the headset type terminal 314 and provides the generated alert to family members or the police. === Hard Collateral 1-4 === Each of the multiple elements including the acquisition unit, analysis unit, notification unit, and provision unit described above is realized, for example, by at least one of the robot 414 and the data processing device 12. For example, the acquisition unit acquires the conversation content using the microphone 238 of the robot 414 and transmits it to the data processing device 12 by the control unit 46A. The analysis unit is realized by the specific processing unit 290 of the data processing device 12 and analyzes the acquired conversation content to determine the possibility of refund fraud. The notification unit is realized by the specific processing unit 290 of the data processing device 12 and generates an alert when the possibility of fraud is detected. The provision unit is realized by the control unit 46A of the robot 414 and provides the generated alert to family members or the police.

[0106] The system according to the embodiment is not limited to the above-described example, and various modifications are possible, for example, as follows.

[0107] The acquisition unit can detect the user's current activity status and adjust the conversation acquisition method according to the activity. For example, if the user is exercising, the acquisition unit can set the conversation acquisition frequency low and acquire only important conversations. Alternatively, if the user is reading in a quiet environment, the acquisition unit can set the conversation acquisition frequency high and acquire all conversations in detail. Furthermore, if the user is in a meeting, the acquisition unit can temporarily stop conversation acquisition and resume it after the meeting ends. This allows for more appropriate conversation acquisition by adjusting the conversation acquisition method according to the user's activity status.

[0108] The analysis unit can understand the context of the conversation and assess the likelihood of fraud based on that context. For example, it can analyze the context before and after the conversation to determine whether specific keywords are used in a fraudulent context. It can also analyze the tone of the conversation and the speaker's intention to assess the likelihood of fraud. It can also detect inconsistencies or unnatural points in the conversation and assess the likelihood of fraud based on those. This allows for more accurate fraud detection by performing analysis that takes context into account.

[0109] The notification unit can select the optimal notification method based on the user's location information. For example, when the user is at home, the notification unit prioritizes voice notification. Also, when the user is out, the notification unit can prioritize text message or push notification. Furthermore, when the user is in a meeting, the notification unit can send a vibration notification and then send a detailed notification after the meeting ends. This allows for effective notification by selecting the optimal notification method based on the user's location information.

[0110] The alert delivery unit can analyze the user's past alert history and optimize the alert delivery method. For example, it can prioritize the use of an alert delivery method that the user has responded to quickly in the past. It can also improve the alert delivery method that the user has ignored in the past and select a more effective method. Furthermore, it can customize the alert delivery method according to specific time periods or situations based on the user's past alert history. This makes it possible to optimize the alert delivery method and provide effective alerts by referring to the past alert history.

[0111] The acquisition unit can estimate the user's emotions and adjust the conversation acquisition method based on the estimated user's emotions. For example, if the user is relaxed, the acquisition unit can set the conversation acquisition frequency low and acquire only important conversations. Alternatively, if the user is nervous, the acquisition unit can set the conversation acquisition frequency high and acquire all conversations in detail. Furthermore, if the user is in a hurry, the acquisition unit can temporarily stop conversation acquisition and resume it later. This allows for more appropriate conversation acquisition by adjusting the conversation acquisition method according to the user's emotions.

[0112] The analysis unit can estimate the user's emotions and determine the analysis priority based on the estimated user's emotions. For example, if the user is relaxed, conversations of low importance can be postponed. Also, if the user is nervous, conversations of high importance can be analyzed with priority. Furthermore, if the user is in a hurry, conversations containing important information can be analyzed with priority in a short time. In this way, by determining the analysis priority based on the user's emotions, important conversations can be analyzed with priority.

[0113] The notification unit can estimate the user's emotions and adjust the content of the notification based on the estimated user's emotions. For example, if the user is relaxed, a detailed notification can be provided. If the user is nervous, a brief notification can be provided. Furthermore, if the user is in a hurry, a quick notification can be provided. In this way, by adjusting the content of the notification based on the user's emotions, more appropriate notifications can be provided.

[0114] The providing unit can estimate the user's emotion and adjust the alert providing method based on the estimated user's emotion. For example, if the user is relaxed, a detailed alert can be provided. If the user is nervous, a concise alert can be provided. Furthermore, if the user is in a hurry, a quick alert can be provided. In this way, by adjusting the alert providing method based on the user's emotion, more appropriate alerts can be provided.

[0115] The providing unit can estimate the user's emotion and adjust the length of the alert provided based on the estimated user's emotion. For example, if the user is relaxed, a detailed alert can be provided. If the user is nervous, a concise alert can be provided. Furthermore, if the user is in a hurry, a quick alert can be provided. In this way, by adjusting the length of the alert provided based on the user's emotion, more appropriate alerts can be provided.

[0116] When providing an alert, the providing unit can determine the priority of providing the alert based on the date and time of the fraud occurrence. For example, if there is a high possibility that the fraud occurred recently, the alert will be provided first. Also, if there is a high possibility that the fraud occurred during a specific time period, the alerts can be concentrated on that time period. Furthermore, the priority of the alert can be determined by referring to the time when past frauds occurred. In this way, by determining the priority of providing the alert based on the time when the fraud occurred, it is possible to provide alerts efficiently.

[0117] The processing flow of the second embodiment will be briefly explained below.

[0118] Step 1: The acquisition unit acquires the conversation content. The conversation content includes, for example, voice data, text data, and part or all of the conversation. The acquisition unit constantly listens to the owner's conversation and can detect phrases that may be related to tax refund fraud in the content of phone conversations and everyday conversations. Step 2: The analysis unit analyzes the conversation content acquired by the acquisition unit and determines whether or not there is any refund fraud. The analysis unit detects conversation content containing keywords such as "refund," "transfer," and "bank account," and uses the generation AI to analyze the conversation content to determine the possibility of refund fraud. Step 3: The notification unit sends an alert based on the results determined by the analysis unit. If a possible fraud is detected, the notification unit sends an alert to family members or the police. The notification unit can use generation AI to send alerts. Step 4: The sending unit provides the alert sent by the notification unit. The sending unit provides the alert to family members or the police. The sending unit can provide the alert using the generation AI.

[0119] The specific processing unit 290 transmits the result of the specific processing to the smart device 14. In the smart device 14, the control unit 46A causes the output device 40 to output the result of the specific processing. The microphone 38B acquires audio indicating a user input regarding the result of the specific processing. The control unit 46A transmits audio data indicating the user input acquired by the microphone 38B to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the audio data.

[0120] The data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of the data generation model 58 is ChatGPT (registered trademark) (Internet search engine).<URL: https: / / openai.com / blog / chatgpt> Examples of generative AIs include the data generation model 58, such as a neural network model (e.g., a neural network model), and a neural network model (e.g., a neural network model). The data generation model 58 is obtained by performing deep learning on a neural network. A prompt including an instruction is input to the data generation model 58, and inference data such as voice data indicating speech, text data indicating text, and image data indicating an image is also input to the data generation model 58. The data generation model 58 performs inference on the input inference data in accordance with the instruction indicated by the prompt and outputs the inference result in a data format such as voice data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization. The specification processing unit 290 performs the above-mentioned specification processing using the data generation model 58. The data generation model 58 may be a fine-tuned model so as to output an inference result from a prompt that does not include an instruction. In this case, the data generation model 58 can output an inference result from a prompt that does not include an instruction. The data processing device 12 and the like include multiple types of data generation models 58, and the data generation model 58 includes AIs other than the generative AI. The AI ​​other than the generative AI may be, for example, linear regression, logistic regression, decision tree, random forest, support vector machine (SVM), k-means clustering, convolutional neural network (CNN), recurrent neural network (RNN), generative adversarial network (GAN), or naive Bayes, and can perform various processes, but is not limited to these examples. The AI ​​may also be an AI agent. When the processes of each of the above-mentioned parts are performed by AI, the processes may be performed in part or entirely by AI, but are not limited to these examples. The processes performed by AI, including the generative AI, may be replaced with rule-based processes.

[0121] Furthermore, the processing by the data processing system 10 described above is executed by the specific processing unit 290 of the data processing device 12 or the control unit 46A of the smart device 14, but may also be executed by the specific processing unit 290 of the data processing device 12 and the control unit 46A of the smart device 14. Furthermore, the specific processing unit 290 of the data processing device 12 acquires or collects information necessary for processing from the smart device 14 or an external device, and the smart device 14 acquires or collects information necessary for processing from the data processing device 12 or an external device.

[0122] The correspondence between each part and the device or control part is not limited to the example described above, and various modifications are possible.

[0123] [Second embodiment] FIG. 3 shows an example of the configuration of a data processing system 210 according to the second embodiment.

[0124] 3, the data processing system 210 includes the data processing device 12 and smart glasses 214. An example of the data processing device 12 is a server.

[0125] The data processing device 12 includes a computer 22, a database 24, and a communication I / F 26. The computer 22 includes a processor 28, a RAM 30, and a storage 32. The processor 28, RAM 30, and storage 32 are connected to a bus 34. The database 24 and the communication I / F 26 are also connected to the bus 34. The communication I / F 26 is connected to a network 54. Examples of the network 54 include a WAN and / or a LAN.

[0126] The smart glasses 214 include a computer 36, a microphone 238, a speaker 240, a camera 42, and a communication I / F 44. The computer 36 includes a processor 46, a RAM 48, and a storage 50. The processor 46, the RAM 48, and the storage 50 are connected to a bus 52. The microphone 238, the speaker 240, and the camera 42 are also connected to the bus 52.

[0127] The microphone 238 receives instructions and the like from the user by receiving voice uttered by the user. The microphone 238 captures the voice uttered by the user, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio according to instructions from the processor 46.

[0128] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an imaging element such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the user's surroundings (for example, an imaging range defined by an angle of view equivalent to the field of vision of a typical healthy person).

[0129] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 are responsible for the exchange of various information between the processor 46 and the processor 28 via the network 54. The exchange of various information between the processor 46 and the processor 28 using the communication I / Fs 44 and 26 is carried out in a secure state.

[0130] Fig. 4 shows an example of the main functions of the data processing device 12 and the smart glasses 214. As shown in Fig. 4, in the data processing device 12, a specific process is performed by the processor 28. A specific process program 56 is stored in the storage 32.

[0131] 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.

[0132] The storage 32 stores a data generation model 58 and an emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290. The identification processing unit 290 can estimate a user's emotion using the emotion identification model 59 and perform identification processing using the user's emotion. The emotion estimation function (emotion identification function) using the emotion identification model 59 performs various estimations and predictions regarding the user's emotion, including estimation and prediction of the user's emotion, but is not limited to these examples. Furthermore, the estimation and prediction of emotion also includes, for example, emotion analysis.

[0133] In the smart glasses 214, the specific processing is performed by the processor 46. A specific processing program 60 is stored in the storage 50. The processor 46 reads the specific processing program 60 from the storage 50 and executes the read specific processing program 60 on the RAM 48. The specific processing is realized by the processor 46 operating as the control unit 46A in accordance with the specific processing program 60 executed on the RAM 48. The smart glasses 214 also have a data generation model and an emotion identification model similar to the data generation model 58 and the emotion identification model 59, and can perform processing similar to that of the specific processing unit 290 using these models.

[0134] Note that a device other than the data processing device 12 may have the data generation model 58. For example, a server device may have the data generation model 58. In this case, the data processing device 12 communicates with the server device having the data generation model 58 to obtain a processing result (such as a prediction result) using the data generation model 58. Furthermore, the data processing device 12 may be a server device, or may be a terminal device (for example, a mobile phone, a robot, a home appliance, etc.) owned by a user.

[0135] The specific processing unit 290 transmits the result of the specific processing to the smart glasses 214. In the smart glasses 214, the control unit 46A causes the speaker 240 to output the result of the specific processing. The microphone 238 acquires audio indicating a user input regarding the result of the specific processing. The control unit 46A transmits audio data indicating the user input acquired by the microphone 238 to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the audio data.

[0136] The data generation model 58 is a so-called generative AI. An example of the data generation model 58 is a generative AI such as ChatGPT. The data generation model 58 is obtained by performing deep learning on a neural network. The data generation model 58 receives a prompt containing an instruction, as well as inference data such as voice data representing speech, text data representing text, and image data representing an image. The data generation model 58 performs inference on the input inference data in accordance with the instruction indicated by the prompt and outputs the inference result in a data format such as voice data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization. The identification processing unit 290 performs the above-mentioned identification processing using the data generation model 58. The data generation model 58 may be a fine-tuned model so as to output an inference result from a prompt that does not include an instruction. In this case, the data generation model 58 can output an inference result from a prompt that does not include an instruction. The data processing device 12 and the like include multiple types of data generation models 58, and the data generation model 58 includes AIs other than the generative AI. The AI ​​other than the generative AI may be, for example, linear regression, logistic regression, decision tree, random forest, support vector machine (SVM), k-means clustering, convolutional neural network (CNN), recurrent neural network (RNN), generative adversarial network (GAN), or naive Bayes, and can perform various processes, but is not limited to these examples. The AI ​​may also be an AI agent. When the processes of each of the above-mentioned parts are performed by AI, the processes may be performed in part or entirely by AI, but are not limited to these examples. The processes performed by AI, including the generative AI, may be replaced with rule-based processes.

[0137] The data processing system 210 according to the second embodiment performs the same processing as the data processing system 10 according to the first embodiment. The processing by the data processing system 210 is executed by the specific processing unit 290 of the data processing device 12 or the control unit 46A of the smart glasses 214, but may also be executed by the specific processing unit 290 of the data processing device 12 and the control unit 46A of the smart glasses 214. Furthermore, the specific processing unit 290 of the data processing device 12 acquires or collects information required for processing from the smart glasses 214 or an external device, etc., and the smart glasses 214 acquires or collects information required for processing from the data processing device 12 or an external device, etc.

[0138] The correspondence between each part and the device or control part is not limited to the example described above, and various modifications are possible.

[0139] [Third embodiment] FIG. 5 shows an example of the configuration of a data processing system 310 according to the third embodiment.

[0140] 5, the data processing system 310 includes the data processing device 12 and a headset type terminal 314. An example of the data processing device 12 is a server.

[0141] The data processing device 12 includes a computer 22, a database 24, and a communication I / F 26. The computer 22 includes a processor 28, a RAM 30, and a storage 32. The processor 28, RAM 30, and storage 32 are connected to a bus 34. The database 24 and the communication I / F 26 are also connected to the bus 34. The communication I / F 26 is connected to a network 54. Examples of the network 54 include a WAN and / or a LAN.

[0142] The headset type terminal 314 includes a computer 36, a microphone 238, a speaker 240, a camera 42, a communication I / F 44, and a display 343. The computer 36 includes a processor 46, a RAM 48, and a storage 50. The processor 46, the RAM 48, and the storage 50 are connected to a bus 52. The microphone 238, the speaker 240, the camera 42, and the display 343 are also connected to the bus 52.

[0143] The microphone 238 receives instructions and the like from the user by receiving voice uttered by the user. The microphone 238 captures the voice uttered by the user, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio according to instructions from the processor 46.

[0144] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an imaging element such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the user's surroundings (for example, an imaging range defined by an angle of view equivalent to the field of vision of a typical healthy person).

[0145] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 are responsible for the exchange of various information between the processor 46 and the processor 28 via the network 54. The exchange of various information between the processor 46 and the processor 28 using the communication I / Fs 44 and 26 is carried out in a secure state.

[0146] Fig. 6 shows an example of the main functions of the data processing device 12 and the headset type terminal 314. As shown in Fig. 6, in the data processing device 12, a specific process is performed by the processor 28. A specific process program 56 is stored in the storage 32.

[0147] 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.

[0148] The storage 32 stores a data generation model 58 and an emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290. The identification processing unit 290 can estimate a user's emotion using the emotion identification model 59 and perform identification processing using the user's emotion. The emotion estimation function (emotion identification function) using the emotion identification model 59 performs various estimations and predictions regarding the user's emotion, including estimation and prediction of the user's emotion, but is not limited to these examples. Furthermore, the estimation and prediction of emotion also includes, for example, emotion analysis.

[0149] In the headset type terminal 314, the identification process is performed by the processor 46. A identification program 60 is stored in the storage 50. The processor 46 reads the identification program 60 from the storage 50 and executes the read identification program 60 on the RAM 48. The identification process is realized by the processor 46 operating as a control unit 46A in accordance with the identification program 60 executed on the RAM 48. Note that the headset type terminal 314 has a data generation model and an emotion identification model similar to the data generation model 58 and the emotion identification model 59, and can also perform processing similar to that of the identification processing unit 290 using these models.

[0150] Note that a device other than the data processing device 12 may have the data generation model 58. For example, a server device may have the data generation model 58. In this case, the data processing device 12 communicates with the server device having the data generation model 58 to obtain a processing result (such as a prediction result) using the data generation model 58. Furthermore, the data processing device 12 may be a server device, or may be a terminal device (for example, a mobile phone, a robot, a home appliance, etc.) owned by a user.

[0151] The specific processing unit 290 transmits the result of the specific processing to the headset type terminal 314. In the headset type terminal 314, the control unit 46A causes the speaker 240 and the display 343 to output the result of the specific processing. The microphone 238 acquires audio indicating a user input regarding the result of the specific processing. The control unit 46A transmits audio data indicating the user input acquired by the microphone 238 to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the audio data.

[0152] The data generation model 58 is a so-called generative AI. An example of the data generation model 58 is a generative AI such as ChatGPT. The data generation model 58 is obtained by performing deep learning on a neural network. The data generation model 58 receives a prompt containing an instruction, as well as inference data such as voice data representing speech, text data representing text, and image data representing an image. The data generation model 58 performs inference on the input inference data in accordance with the instruction indicated by the prompt and outputs the inference result in a data format such as voice data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization. The identification processing unit 290 performs the above-mentioned identification processing using the data generation model 58. The data generation model 58 may be a fine-tuned model so as to output an inference result from a prompt that does not include an instruction. In this case, the data generation model 58 can output an inference result from a prompt that does not include an instruction. The data processing device 12 and the like include multiple types of data generation models 58, and the data generation model 58 includes AIs other than the generative AI. The AI ​​other than the generative AI may be, for example, linear regression, logistic regression, decision tree, random forest, support vector machine (SVM), k-means clustering, convolutional neural network (CNN), recurrent neural network (RNN), generative adversarial network (GAN), or naive Bayes, and can perform various processes, but is not limited to these examples. The AI ​​may also be an AI agent. When the processes of each of the above-mentioned parts are performed by AI, the processes may be performed in part or entirely by AI, but are not limited to these examples. The processes performed by AI, including the generative AI, may be replaced with rule-based processes.

[0153] The data processing system 310 according to the third embodiment performs the same processing as the data processing system 10 according to the first embodiment. The processing by the data processing system 310 is executed by the specific processing unit 290 of the data processing device 12 or the control unit 46A of the headset type terminal 314, but may also be executed by the specific processing unit 290 of the data processing device 12 and the control unit 46A of the headset type terminal 314. Furthermore, the specific processing unit 290 of the data processing device 12 acquires or collects information required for processing from the headset type terminal 314 or an external device, etc., and the headset type terminal 314 acquires or collects information required for processing from the data processing device 12 or an external device, etc.

[0154] The correspondence between each part and the device or control part is not limited to the example described above, and various modifications are possible.

[0155] [Fourth embodiment] FIG. 7 shows an example of the configuration of a data processing system 410 according to the fourth embodiment.

[0156] 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.

[0157] The data processing device 12 includes a computer 22, a database 24, and a communication I / F 26. The computer 22 includes a processor 28, a RAM 30, and a storage 32. The processor 28, RAM 30, and storage 32 are connected to a bus 34. The database 24 and the communication I / F 26 are also connected to the bus 34. The communication I / F 26 is connected to a network 54. Examples of the network 54 include a WAN and / or a LAN.

[0158] The robot 414 includes a computer 36, a microphone 238, a speaker 240, a camera 42, a communication I / F 44, and a control target 443. The computer 36 includes a processor 46, a RAM 48, and a storage 50. The processor 46, the RAM 48, and the storage 50 are connected to a bus 52. The microphone 238, the speaker 240, the camera 42, and the control target 443 are also connected to the bus 52.

[0159] The microphone 238 receives instructions and the like from the user by receiving voice uttered by the user. The microphone 238 captures the voice uttered by the user, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio according to instructions from the processor 46.

[0160] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an imaging element such as a CMOS image sensor or a CCD image sensor, and captures images of the user's surroundings (for example, an imaging range defined by an angle of view equivalent to the field of vision of a typical healthy person).

[0161] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 are responsible for the exchange of various information between the processor 46 and the processor 28 via the network 54. The exchange of various information between the processor 46 and the processor 28 using the communication I / Fs 44 and 26 is carried out in a secure state.

[0162] The control 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 emotions of the robot 414 can be expressed by controlling these motors. In addition, the facial expressions of the robot 414 can also be expressed by controlling the light emission state of the LEDs in the eyes of the robot 414.

[0163] Fig. 8 shows an example of the main functions of the data processing device 12 and the robot 414. As shown in Fig. 8, in the data processing device 12, a specific process is performed by the processor 28. A specific process program 56 is stored in the storage 32.

[0164] 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.

[0165] The storage 32 stores a data generation model 58 and an emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290. The identification processing unit 290 can estimate a user's emotion using the emotion identification model 59 and perform identification processing using the user's emotion. The emotion estimation function (emotion identification function) using the emotion identification model 59 performs various estimations and predictions regarding the user's emotion, including estimation and prediction of the user's emotion, but is not limited to these examples. Furthermore, the estimation and prediction of emotion also includes, for example, emotion analysis.

[0166] In the robot 414, the processor 46 performs the identification process. The storage 50 stores the identification program 60. The processor 46 reads the identification program 60 from the storage 50 and executes the read identification program 60 on the RAM 48. The identification process is realized by the processor 46 operating as the control unit 46A in accordance with the identification program 60 executed on the RAM 48. The robot 414 also has a data generation model and an emotion identification model similar to the data generation model 58 and the emotion identification model 59, and can perform the same process as the identification processing unit 290 using these models.

[0167] Note that a device other than the data processing device 12 may have the data generation model 58. For example, a server device may have the data generation model 58. In this case, the data processing device 12 communicates with the server device having the data generation model 58 to obtain a processing result (such as a prediction result) using the data generation model 58. Furthermore, the data processing device 12 may be a server device, or may be a terminal device (for example, a mobile phone, a robot, a home appliance, etc.) owned by a user.

[0168] The specific processing unit 290 transmits the result of the specific processing to the robot 414. In the robot 414, the control unit 46A causes the speaker 240 and the control target 443 to output the result of the specific processing. The microphone 238 acquires voice indicating a user input regarding the result of the specific processing. The control unit 46A transmits voice data indicating the user input acquired by the microphone 238 to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the voice data.

[0169] The data generation model 58 is a so-called generative AI. An example of the data generation model 58 is a generative AI such as ChatGPT. The data generation model 58 is obtained by performing deep learning on a neural network. The data generation model 58 receives a prompt containing an instruction, as well as inference data such as voice data representing speech, text data representing text, and image data representing an image. The data generation model 58 performs inference on the input inference data in accordance with the instruction indicated by the prompt and outputs the inference result in a data format such as voice data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization. The identification processing unit 290 performs the above-mentioned identification processing using the data generation model 58. The data generation model 58 may be a fine-tuned model so as to output an inference result from a prompt that does not include an instruction. In this case, the data generation model 58 can output an inference result from a prompt that does not include an instruction. The data processing device 12 and the like include multiple types of data generation models 58, and the data generation model 58 includes AIs other than the generative AI. The AI ​​other than the generative AI may be, for example, linear regression, logistic regression, decision tree, random forest, support vector machine (SVM), k-means clustering, convolutional neural network (CNN), recurrent neural network (RNN), generative adversarial network (GAN), or naive Bayes, and can perform various processes, but is not limited to these examples. The AI ​​may also be an AI agent. When the processes of each of the above-mentioned parts are performed by AI, the processes may be performed in part or entirely by AI, but are not limited to these examples. The processes performed by AI, including the generative AI, may be replaced with rule-based processes.

[0170] The data processing system 410 according to the fourth embodiment performs the same processing as the data processing system 10 according to the first embodiment. The processing by the data processing system 410 is executed by the specific processing unit 290 of the data processing device 12 or the control unit 46A of the robot 414, but may also be executed by the specific processing unit 290 of the data processing device 12 and the control unit 46A of the robot 414. Furthermore, the specific processing unit 290 of the data processing device 12 acquires or collects information required for processing from the robot 414 or an external device, etc., and the robot 414 acquires or collects information required for processing from the data processing device 12 or an external device, etc.

[0171] The correspondence between each part and the device or control part is not limited to the example described above, and various modifications are possible.

[0172] The emotion identification model 59 as an emotion engine may determine the user's emotion according to a specific mapping. Specifically, the emotion identification model 59 may determine the user's emotion according to an emotion map (see FIG. 9), which is a specific mapping. Similarly, the emotion identification model 59 may determine the robot's emotion, and the identification processing unit 290 may perform identification processing using the robot's emotion.

[0173] FIG. 9 illustrates an emotion map 400 on which multiple emotions are mapped. In the emotion map 400, emotions are arranged in concentric circles radiating from the center. Emotions closer to the center of the concentric circles are more primitive. Emotions representing states and behaviors arising from a state of mind are arranged on the outer edges of the concentric circles. The concept of emotion encompasses both emotions and mental states. Emotions generally generated from reactions occurring in the brain are arranged on the left side of the concentric circles. Emotions generally induced by situational judgment are arranged on the right side of the concentric circles. Emotions generally generated from reactions occurring in the brain and induced by situational judgment are arranged on the upper and lower sides of the concentric circles. Furthermore, the emotion of "pleasure" is arranged on the upper side of the concentric circles, and the emotion of "discomfort" is arranged on the lower side. In this way, in the emotion map 400, multiple emotions are mapped based on the structure by which emotions are generated, and emotions that tend to occur simultaneously are mapped close to each other.

[0174] These emotions are distributed in the 3 o'clock direction on emotion map 400, and typically fluctuate between relief and anxiety. In the right half of emotion map 400, situational awareness dominates over internal sensations, resulting in a sense of calm.

[0175] The inside of emotion map 400 represents what is going on in the mind, and the outside of emotion map 400 represents behavior, so the further you go outside emotion map 400, the more visible the emotions become (the more they are expressed in behavior).

[0176] Human emotions are based on various balances, such as posture and blood sugar levels. When these balances deviate from the ideal, a state of discomfort is expressed, and when they approach the ideal, a state of pleasure is expressed. Emotions can also be created for robots, cars, and motorcycles, based on various balances, such as posture and remaining battery life. When these balances deviate from the ideal, a state of discomfort is expressed, and when they approach the ideal, a state of pleasure is expressed. An emotion map can be generated, for example, based on Dr. Mitsuyoshi's emotion map (Research on speech emotion recognition and brain physiological signal analysis systems for emotions, Tokushima University, doctoral dissertation: https: / / ci.nii.ac.jp / naid / 500000375379). The left half of the emotion map lists emotions belonging to the area called "reaction," where sensation is dominant. The right half of the emotion map lists emotions belonging to the area called "situation," where situational awareness is dominant.

[0177] The emotion map defines two emotions that promote learning. One is a negative emotion on the situation side, around the middle of "repentance" or "reflection." In other words, this occurs when the robot experiences negative emotions such as "I never want to feel this way again" or "I don't want to be scolded again." The other is a positive emotion on the response side, around "desire." In other words, this occurs when the robot experiences positive feelings such as "I want more" or "I want to know more."

[0178] The emotion identification model 59 inputs user input into a pre-trained neural network, obtains emotion values ​​indicating each emotion shown in the emotion map 400, and determines the user's emotion. This neural network is pre-trained based on multiple pieces of training data that are combinations of user input and emotion values ​​indicating each emotion shown in the emotion map 400. Furthermore, this neural network is trained so that emotions that are located close to each other have similar values, as in the emotion map 900 shown in FIG. 10. FIG. 10 shows an example in which multiple emotions, "relieved," "calm," and "reassuring," have similar emotion values.

[0179] In the above embodiment, an example was given in which a specific process is performed by one computer 22, but the technology disclosed herein is not limited to this, and distributed processing of the specific process may be performed by multiple computers including computer 22.

[0180] In the above embodiment, an example in which the specific processing program 56 is stored in the storage 32 has been described, but the technology of the present disclosure is not limited to this. For example, the specific processing program 56 may be stored in a portable, computer-readable, non-transitory storage medium such as a USB (Universal Serial Bus) memory. The specific processing program 56 stored in the non-transitory storage medium is installed in the computer 22 of the data processing device 12. The processor 28 executes the specific processing in accordance with the specific processing program 56.

[0181] 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.

[0182] It is not necessary to store all of the specific processing program 56 in a storage device such as a server connected to the data processing device 12 via the network 54, or to store all of the specific processing program 56 in the storage 32; only a portion of the specific processing program 56 may be stored.

[0183] The hardware resource for executing a specific process can be any of the following types of processors: A processor, for example, is a CPU, which is a general-purpose processor that functions as a hardware resource for executing a specific process by executing software, i.e., a program. A processor also includes a dedicated electrical circuit, such as an FPGA (Field-Programmable Gate Array), a PLD (Programmable Logic Device), or an ASIC (Application Specific Integrated Circuit), which is a processor with a circuit configuration designed specifically for executing a specific process. Each processor has built-in or connected memory, and each processor uses the memory to execute the specific process.

[0184] The hardware resource that executes the specific process may be configured with one of these various processors, or may be configured with a combination of two or more processors of the same or different types (for example, a combination of multiple FPGAs, or a combination of a CPU and an FPGA). Also, the hardware resource that executes the specific process may be a single processor.

[0185] As an example of a system configured with a single processor, first, one processor is configured by combining one or more CPUs and software, and this processor functions as a hardware resource that executes a specific process. Second, there is a system that uses a processor that realizes the functions of an entire system including multiple hardware resources that execute a specific process on a single IC chip, as typified by SoC (System-on-a-chip). In this way, a specific process is realized using one or more of the above-mentioned various processors as hardware resources.

[0186] Furthermore, the hardware structure of these various processors can be, more specifically, an electric circuit that combines circuit elements such as semiconductor devices. The specific processing described above is merely an example. Therefore, it goes without saying that unnecessary steps may be deleted, new steps may be added, or the processing order may be rearranged, without departing from the spirit of the invention.

[0187] In the above example, the first to fourth embodiments have been described separately, but some or all of these embodiments may be combined. The smart device 14, smart glasses 214, headset terminal 314, and robot 414 are merely examples, and they may be combined, or other devices may be used. In the above example, the first and second embodiments have been described separately, but they may be combined.

[0188] The above-described description and illustrations are a detailed explanation of the parts related to the technology of the present disclosure and are merely an example of the technology of the present disclosure. For example, the above description of the configuration, functions, actions, and effects is an explanation of an example of the configuration, functions, actions, and effects of the parts related to the technology of the present disclosure. Therefore, it goes without saying that unnecessary parts may be deleted, new elements may be added, or replacements may be made to the above-described description and illustrations within the scope of the gist of the technology of the present disclosure. Furthermore, to avoid confusion and facilitate understanding of the parts related to the technology of the present disclosure, the above-described description and illustrations omit explanations of common technical knowledge that do not require particular explanation to enable the implementation of the technology of the present disclosure.

[0189] All publications, patent applications, and technical standards mentioned in this specification are herein incorporated by reference to the same extent as if each individual publication, patent application, or technical standard was specifically and individually indicated to be incorporated by reference.

[0190] [Explanation of symbols]

[0191] 10, 210, 310, 410 Data Processing Systems 12 Data Processing Device 14 Smart Devices 214 Smart Glasses 314 Headset-type terminal 414 Robot

Claims

1. an acquisition unit that acquires conversation content; an analysis unit that analyzes the conversation content acquired by the acquisition unit and determines whether or not there is any refund fraud; a notification unit that sends an alert based on the result determined by the analysis unit; a transmission unit that transmits the alert transmitted by the notification unit. A system characterized by:

2. The acquisition unit Continuously listens to your conversations 2. The system of claim 1.

3. The analysis unit Detect conversations containing specific keywords 2. The system of claim 1.

4. The notification unit Send alerts to family or police if fraud is detected 2. The system of claim 1.

5. The analysis unit Conversation content is analyzed only within the device and is not sent to an external server.

2. The system of claim 1.

6. The acquisition unit Estimate user emotions and adjust conversation capture timing based on the estimated user emotions 2. The system of claim 1.

7. The acquisition unit Analyze the user's past conversation history and select the appropriate acquisition method 2. The system of claim 1.

Images