system

Abstract

A system is provided.SOLUTION: A system comprising: means for monitoring answers and files generated by other artificial intelligence systems to check for suspicious content; means for monitoring reference URLs to detect fraudulent and malicious sites; means for monitoring generated program code to check for malicious programs; and means for issuing warnings when malicious content is detected.SELECTED DRAWING: Figure 1

Description

[Technical Field]

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

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

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

[0004] New AI technologies are being released frequently, and due to their convenience, many users are installing and using them without even thinking about it. However, these technologies may contain malicious AI technologies, which pose a risk of generating malicious answers, illegally collecting user data, or installing malicious programs. These risks pose a serious problem, as they could lead to security breaches without the user's knowledge. Therefore, there is a need for technology that can monitor the behavior and creation of malicious AI and mitigate these risks. [Means for solving the problem]

[0005] To solve this problem, the present invention provides the following means.

[0006] First, it includes a means to monitor answers and files generated by other AI systems and check for suspicious content, specifically by analyzing the content of the generated results using predefined keywords.

[0007] Second, it provides a means to monitor referring URLs and detect fraudulent or malicious sites by sending HTTP requests to the URLs and analyzing the page content to detect suspicious patterns.

[0008] Furthermore, a method is added to monitor the generated program code and check whether it contains malicious code, which makes it possible to detect unauthorized operations in the code in advance.

[0009] Finally, it provides a means to alert users when malicious content is detected, ensuring immediate notification to users.

[0010] A system that combines these measures can reduce the risk of malicious AI technology and increase security for users.

[0011] An "artificial intelligence system" is an information processing system that mimics human intelligence and has functions such as learning, reasoning, recognition, and decision-making.

[0012] "Monitoring" means continuously observing and analyzing specific behaviors and data, and notifying you when a problem occurs.

[0013] An "answer" is the text or information that an AI system generates in response to a user's input or question.

[0014] A "file" is a unit for recording digital data, and is a collection of data that an artificial intelligence system generates or manipulates.

[0015] "Suspicious content" refers to content that may contain malicious behavior or information and poses a safety risk.

[0016] A "referring URL" is an address (URL) that specifies a resource on the web and indicates the link destination that users can access.

[0017] A "fraudulent website" is a website designed to deceive users and operated with the intent of fraudulently obtaining personal information or money.

[0018] A "malicious site" is a website whose purpose is to harm users, such as by distributing malware or conducting phishing attacks.

[0019] "Program code" refers to scripts or source code that contain instructions for a computer to execute.

[0020] A "malicious program" is software designed with the intent to misuse or harm a user or system.

[0021] A "warning" is a notification or message that notifies you when an abnormal or dangerous situation occurs. [Brief explanation of the drawings]

[0022] [Figure 1] 1 is a conceptual diagram showing an example of the configuration of a data processing system according to a first embodiment. [Figure 2] 1 is a conceptual diagram showing an example of main functions of a data processing device and a smart device according to a first embodiment. [Figure 3] FIG. 10 is a conceptual diagram showing an example of the configuration of a data processing system according to a second embodiment. [Figure 4] FIG. 10 is a conceptual diagram showing an example of main functions of a data processing device and smart glasses according to a second embodiment. [Figure 5]FIG. 10 is a conceptual diagram showing an example of the configuration of a data processing system according to a third embodiment. [Figure 6] FIG. 11 is a conceptual diagram showing an example of main functions of a data processing device and a headset-type terminal according to a third embodiment. [Figure 7] FIG. 10 is a conceptual diagram showing an example of the configuration of a data processing system according to a fourth embodiment. [Figure 8] FIG. 10 is a conceptual diagram showing an example of main functions of a data processing device and a robot according to a fourth embodiment. [Figure 9] 1 shows an emotion map onto which multiple emotions are mapped. [Figure 10] 1 shows an emotion map onto which multiple emotions are mapped. [Figure 11] FIG. 3 is a sequence diagram showing a processing flow of the data processing system according to the first embodiment. [Figure 12] FIG. 10 is a sequence diagram showing the flow of processing in the data processing system in Application Example 1. [Figure 13] FIG. 10 is a sequence diagram showing the flow of processing in the data processing system according to the second embodiment when an emotion engine is combined. [Figure 14] FIG. 10 is a sequence diagram showing the flow of processing in the data processing system in Application Example 2 when an emotion engine is combined. DETAILED DESCRIPTION OF THE INVENTION

[0023] 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.

[0024] First, the terms used in the following description will be explained.

[0025] In the following embodiments, a coded processor (hereinafter simply referred to as a "processor") may be a single arithmetic device or a combination of multiple arithmetic devices. Furthermore, a processor may be a single type of arithmetic device or a combination of multiple types of arithmetic devices. Examples of arithmetic devices include a CPU (Central Processing Unit), a GPU (Graphics Processing Unit), a GPGPU (General-Purpose computing on Graphics Processing Units), and an APU (Accelerated Processing Unit).

[0026] 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.

[0027] 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.

[0028] In the following embodiments, a communication I / F (Interface) with a symbol is an interface including a communication processor, an antenna, etc. The communication I / F controls communication between multiple computers. Examples of communication standards applied to the communication I / F include wireless communication standards including 5G (5th Generation Mobile Communication System), Wi-Fi (registered trademark), Bluetooth (registered trademark), etc.

[0029] 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."

[0030] [First embodiment]

[0031] FIG. 1 shows an example of the configuration of a data processing system 10 according to the first embodiment.

[0032] 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.

[0033] The data processing device 12 includes a computer 22, a database 24, and a communication I / F 26. The computer 22 is an example of a "computer" according to the technology of the present disclosure. The computer 22 includes a processor 28, a RAM 30, and a storage 32. The processor 28, the RAM 30, and the storage 32 are connected to a bus 34. The database 24 and the communication I / F 26 are also connected to the bus 34. The communication I / F 26 is connected to a network 54. Examples of the network 54 include a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0034] 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.

[0035] The reception device 38 includes a touch panel 38A, a microphone 38B, and the like, and receives user input. The touch panel 38A detects contact with an indicator (for example, a pen or a finger) to receive user input by the touch of the indicator. The microphone 38B detects the user's voice to receive user input by voice. The control unit 46A transmits data indicating the user input received by the touch panel 38A and the microphone 38B to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the data indicating the user input.

[0036] The output device 40 includes a display 40A and a speaker 40B, and presents data to the user 20 by outputting the data in a form of expression that the user 20 can perceive (for example, audio and / or text). The display 40A displays visible information such as text and images in accordance with instructions from the processor 46. The speaker 40B outputs audio in accordance with instructions from the processor 46. The camera 42 is a compact digital camera equipped with an optical system including a lens, aperture, and shutter, and an imaging element such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor.

[0037] 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.

[0038] FIG. 2 shows an example of the main functions of the data processing device 12 and the smart device 14.

[0039] 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.

[0040] 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.

[0041] In the smart device 14, the processor 46 performs the reception output process. The storage 50 stores a reception output program 60. The reception output program 60 is used in conjunction with the specific processing program 56 by the data processing system 10. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output process is realized by the processor 46 operating as the control unit 46A in accordance with the reception output program 60 executed on the RAM 48.

[0042] Next, a description will be given of the specific processing performed by the specific processing unit 290 of the data processing device 12. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0043] The present invention is a system that monitors answers and files generated by other AI systems, detects suspicious content, and issues a warning to users. This system operates as software installed on a server, and plays a role in enhancing user security.

[0044] Basic system configuration

[0045] The system consists of the following components:

[0046] 1. Answer and file monitoring module: This module takes answers and files generated by other AI systems and analyzes them for suspicious content.

[0047] 2. URL Monitoring Module: Monitors referenced URLs in answers and files to check whether they are fraudulent or malicious sites.

[0048] 3. Program code monitoring module: Checks the program code generated by other AI systems to monitor whether it contains malicious code.

[0049] 4. Alert module: Issues an alert to the user if a problem is detected as a result of monitoring.

[0050] Program processing

[0051] 1. Monitoring responses and files

[0052] Server: The server receives answers and files generated by other AI systems, treating the received content as strings or data.

[0053] Server: The analysis module checks the content using predefined keywords (e.g., phishing, malware, suspicious) and issues a warning if any suspicious items are included.

[0054] 2. URL Monitoring

[0055] Server: Parses messages and content containing referral URLs and extracts the URLs.

[0056] Server: Sends an HTTP request to the extracted URL and retrieves the page content.

[0057] Server: Analyzes the retrieved page content and checks the text against predefined patterns to detect fraudulent or malicious sites, and issues a warning if a problem is detected.

[0058] 3. Program code monitoring

[0059] Server: Obtains program code generated by other AI systems, and analyzes the instructions in the code as strings.

[0060] Server: Uses regular expressions to check whether a program contains suspicious code patterns such as "import os", "import sys", or "exec(", and issues a warning if they are found.

[0061] 4. Issuing a warning

[0062] Server: A warning module runs and notifies the user of detailed information about detected problems. Warnings are mainly output to the console log, email, or as messages on the GUI.

[0063] Specific examples

[0064] For example, when another AI system recommends a link to a user, the server must verify whether the link is a fraudulent site. If the server analyzes the link and determines that it is likely to be a fraudulent site, it warns the user that "this link is not safe." Also, if the program code provided to the user by the other AI system contains fraudulent operations, the server detects this and issues a warning that "this code is dangerous."

[0065] This will allow users to minimize the risk from malicious AI technology and use the system with peace of mind, and this system will play an important role in improving user security.

[0066] The processing flow will be explained below.

[0067] Step 1:

[0068] The server receives answers and files generated by other AI systems, and the data is stored on the server as a string or file.

[0069] Step 2:

[0070] The server analyzes the responses and files it receives and checks their content against predefined keywords (e.g., phishing, malware, suspicious). The analysis module is responsible for this and checks whether any suspicious words are included.

[0071] Step 3:

[0072] The server extracts the reference URLs contained in the response or file, using regular expressions to extract the URLs from the text, and creates a list.

[0073] Step 4:

[0074] The server sends an HTTP request to the extracted URL to retrieve the content of the target web page. The request module issues an HTTP request and receives the response from the server.

[0075] Step 5:

[0076] The server performs text analysis on the retrieved web page content, checking for patterns (e.g., phishing terms, suspicious links) to detect fraudulent or malicious sites. The content is analyzed as text to see if it contains any suspicious patterns.

[0077] Step 6:

[0078] The server acquires program code generated by other AI systems and analyzes the code content. At this time, the code is acquired as a string and is the subject of inspection.

[0079] Step 7:

[0080] The server uses regular expressions to check whether the acquired program code contains suspicious code patterns such as "import os", "import sys", or "exec(". The analysis module checks each line of the code and records any suspicious patterns it finds.

[0081] Step 8:

[0082] The server combines the analysis results from each module and, if any suspicious content is detected, generates a warning message that contains detailed information about the detected problem and is presented in a user-friendly format.

[0083] Step 9:

[0084] The server notifies the user of any generated warning messages via console logs, email, or messages on the user interface, allowing the user to quickly identify suspicious areas.

[0085] This allows the server to continuously monitor the outputs of other artificial intelligence systems, creating a system that ensures user security.

[0086] Example 1

[0087] Next, a description will be given of Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0088] Answers and files generated by conventional AI systems sometimes contain malicious content, fraudulent websites, or malicious code, posing a threat to user security. Furthermore, there is a lack of systems that can efficiently detect such malicious content and quickly warn users. Therefore, there is a need for a reliable monitoring system that allows users of AI systems to use them with peace of mind.

[0089] The specific processing by the specific processing unit 290 of the data processing device 12 in the first embodiment is realized by the following means.

[0090] In this invention, the server includes means for monitoring answers and files generated by other artificial intelligence systems and analyzing suspicious content using predefined keywords, means for monitoring referred URLs and analyzing page content obtained by sending HTTP requests to detect fraudulent or malicious sites, means for monitoring generated program code and checking specific code patterns with regular expressions to detect whether malicious programs are included, and means for notifying users of detailed information about detected problems via console logs, email, or GUI messages. This makes it possible to effectively detect security risks contained in content generated by artificial intelligence systems and promptly notify users.

[0091] "Answers and files" refers to text data and document data generated by other artificial intelligence systems.

[0092] "Monitoring" means the act of continuously or periodically checking or auditing a particular object.

[0093] "Keywords" refers to predefined important phrases or words for a specific purpose.

[0094] "Referring URL" refers to a string of characters that indicates a link to a web page or resource.

[0095] "HTTP request" refers to a request sent by a web browser or client application to retrieve information from a web server.

[0096] "Page Content" refers to the text, images, scripts, and other information displayed on a web page.

[0097] "Fraudulent Site" refers to a fraudulent website designed to deceive users.

[0098] "Malicious Site" means a website whose purpose is to steal user information, cause damage to systems, or perform other fraudulent activities.

[0099] "Program code" refers to the source code for writing a computer program.

[0100] A "code pattern" refers to a set of notations or instructions that have a specific purpose or function.

[0101] "Regular expression" refers to a formal language for describing specific patterns.

[0102] "Console log" refers to character-based output that displays system or application status and error messages.

[0103] "Mail" means electronic mail, a means of sending and receiving messages in digital form.

[0104] "GUI Message" means a warning or informational message displayed on a graphical user interface.

[0105] The present invention is a system that monitors answers and files generated by other AI systems, detects suspicious content, and issues a warning to the user. Specific embodiments for implementing this system are described below.

[0106] Basic system configuration

[0107] The system runs as software installed on a server and consists of the following components:

[0108] 1. Answer and file monitoring module

[0109] 2. URL Monitoring Module

[0110] 3. Program Code Monitoring Module

[0111] 4. Warning Module

[0112] Monitor responses and files

[0113] server

[0114] The server retrieves answers and files generated by other AI systems, using API requests and file system monitoring.

[0115] The retrieved answers and file contents are analyzed as text data, and the server searches the text using predefined keywords (e.g., phishing, malware, suspicious) to extract suspicious content.

[0116] URL monitoring

[0117] server

[0118] The server extracts URLs from the responses and files and generates a list of URLs using regular expressions.

[0119] Send an HTTP request to the extracted URL and get the response.

[0120] The text content of the response is analyzed and matched against patterns (e.g., characteristics of phishing pages) to detect fraudulent or malicious websites.

[0121] Program code monitoring

[0122] server

[0123] The server retrieves program code generated by other artificial intelligence systems from specific folders or APIs.

[0124] The commands in the acquired program code are analyzed as strings and checked for specific patterns (e.g., "import os", "import sys", "exec()") using regular expressions.

[0125] Sending a warning

[0126] server

[0127] The server generates detailed warning messages about any problems it detects, including the type, location, and detailed description of the problem detected.

[0128] The generated warning messages are not only output to the console log, but also notified to the user by email and displayed as a popup message on the GUI.

[0129] Specific examples

[0130] A specific example of operation is shown below.

[0131] Example 1: If the server retrieves a new answer and it contains the keyword "phishing", the server will notify the user "Warning: This answer contains suspicious content."

[0132] Example 2: If the server sends an HTTP request to a URL extracted from a file and the result has characteristics of a phishing site, the server notifies the user with "Warning: This URL is not safe."

[0133] Example 3: When the server parses a new code file and it contains the instruction "exec(", the server will notify the user "Warning: this code is dangerous".

[0134] Prompt Sentence Examples

[0135] Check if this link is safe

[0136] "Check this code for illegal instructions"

[0137] By inputting such prompt sentences into the generative AI model, the server will initiate a specific analysis process, allowing users to obtain reliable information and minimizing security risks.

[0138] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0139] Step 1:

[0140] Get answers and files

[0141] Server: Retrieves answers and files generated by other AI systems, using API requests and file system monitoring.

[0142] Input: Answers or files generated by other AI systems.

[0143] Output: Text data of the obtained answers and files.

[0144] Step 2:

[0145] Check the contents

[0146] Server: Analyzes the obtained answers and file contents as text data. Analyzes suspicious content using predefined keywords (e.g., phishing, malware, suspicious).

[0147] Input: Text data from the answers or files obtained.

[0148] Output: Result on whether or not the content is suspicious.

[0149] What happens: The server searches for text against a list of keywords and flags it as suspicious if it contains a matching keyword.

[0150] Step 3:

[0151] URL extraction

[0152] Server: Extract URLs from responses and files using regular expressions.

[0153] Input: The text data from the parsed answer or file.

[0154] Output: The extracted URL list.

[0155] What it does: The server searches for strings that start with "http: / / " or "https: / / " and lists all URLs.

[0156] Step 4:

[0157] URL validation

[0158] Server: Sends an HTTP request to the extracted URL and obtains the response data.

[0159] Input: The extracted URL list.

[0160] Output: Response data for the URL.

[0161] Specific operation: The server sends an HTTP request for each URL and saves the page content of the response it receives.

[0162] Step 5:

[0163] Analyzing page content

[0164] Server: Analyzes the response data and matches it with patterns (e.g., characteristics of phishing sites) to detect fraudulent or malicious sites.

[0165] Input: Response data for the URL.

[0166] Output: Detected fraudulent and malicious sites.

[0167] What it does: The server performs pattern matching on the text content of the response and flags any fraudulent or malicious features detected.

[0168] Step 6:

[0169] Obtaining the program code

[0170] Server: Retrieves program code generated by other AI systems from specific folders or APIs.

[0171] Input: Program code generated by another artificial intelligence system.

[0172] Output: Text data of the retrieved program code.

[0173] What it does: The server monitors newly added code files and incorporates them one by one.

[0174] Step 7:

[0175] Code content analysis

[0176] Server: Analyzes the program code and checks for specific code patterns (e.g., "import os", "import sys", "exec()") using regular expressions.

[0177] Input: Text data of the retrieved program code.

[0178] Output: Suspicious code pattern detection results.

[0179] What it does: The server parses instructions in your code using regular expressions and flags when certain patterns are detected.

[0180] Step 8:

[0181] Generate a warning message

[0182] Server: Generates detailed warning messages about detected problems.

[0183] Input: Suspicious content, URL, or code pattern detection results.

[0184] Output: A detailed warning message.

[0185] What happens: The server generates a message containing the type, location, and detailed description of the problem detected.

[0186] Step 9:

[0187] User Notification

[0188] Server: Generated warning messages are output to the console log and notified to the user via email or GUI message.

[0189] Input: A detailed warning message.

[0190] Output: The warning message that was sent to the user.

[0191] Specific behavior: The server not only logs the warning to the console, but also sends an email to the user and displays a warning as a popup message in the GUI.

[0192] (Application example 1)

[0193] Next, a description will be given of Application Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0194] With the development of modern information and communication technology, security risks are increasing for answers, files, and program code generated by other artificial intelligence systems. Furthermore, the potential for suspicious URLs and malicious programs contained in this content to cause harm to users is also increasing. In particular, with the widespread use of smart devices, real-time monitoring and alerts are becoming increasingly important. However, current systems lack the means to effectively address these multifaceted threats.

[0195] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 1 is realized by the following means.

[0196] In this invention, the server includes means for monitoring answers and files generated by other artificial intelligence systems and checking for suspicious content, means for monitoring referred URLs and detecting fraudulent or malicious sites, means for monitoring generated program code and checking whether it contains malicious programs, means for issuing a warning when malicious content is detected, and means for operating on the smart device and monitoring the user's digital content in real time, thereby enabling real-time monitoring and warning of the digital content used by the user.

[0197] "Other artificial intelligence systems" refers to artificial intelligence technologies operated or utilized by users to analyze and generate data.

[0198] "Answer or file" refers to text, documents, images, audio, or other forms of data generated by other artificial intelligence systems.

[0199] "Suspicious Content" refers to elements within a file or message that may indicate phishing, malware, spam, or other malicious content.

[0200] "Referring URL" refers to a website address or link contained in a response or file.

[0201] A "fraudulent site" refers to a malicious website that attempts to fraudulently obtain users' personal information or money.

[0202] "Malicious Sites" refers to websites that install malware on users' devices or perform other malicious activities.

[0203] "Program Code" refers to the source code of software generated by other artificial intelligence systems.

[0204] "Malicious Program" refers to source code or a program designed to perform unauthorized operations or undesirable behavior.

[0205] "Warning" refers to a notification or alert issued to a user when the system detects suspicious or malicious content.

[0206] "Smart devices" refers to electronic devices such as mobile phones, tablets, and smart glasses that have internet connectivity.

[0207] "Digital content" refers to electronically stored information or data, such as web pages, messages, files, and program code.

[0208] "Real-time" refers to data being processed immediately at the moment it is generated or acquired.

[0209] This invention provides a security system called "AI Guard," which monitors answers and files generated by other artificial intelligence systems and analyzes whether they contain suspicious content. It also monitors referred URLs to determine whether they are fraudulent or malicious sites, and checks generated program code to monitor whether they contain malicious programs. Based on these monitoring results, it issues a warning to the user when malicious content is detected. Furthermore, this system runs on smart devices and can monitor users' digital content in real time.

[0210] System Configuration

[0211] The system consists of the following elements:

[0212] 1. Answer and file monitoring module:

[0213] The server receives answers and files generated by other AI systems, whose contents are processed as strings or data.

[0214] The server analyzes the content using predefined keywords (e.g., phishing, malware, suspicious) and issues a warning if it contains any suspicious items.

[0215] 2. URL Monitoring Module:

[0216] Parse messages and content that contain URLs and extract them.

[0217] An HTTP request is sent to the extracted URL to retrieve the page content.

[0218] It analyzes the retrieved page content, checks the text against defined patterns to detect fraudulent or malicious sites, and issues a warning if a problem is detected.

[0219] 3. Program code monitoring module:

[0220] It acquires program code generated by other AI systems and analyzes the instructions in the code as strings of characters.

[0221] It uses regular expressions to check whether a program contains suspicious code patterns such as "import os", "import sys", or "exec(", and issues a warning if they are found.

[0222] 4. Alert sending module:

[0223] The server includes a module to notify the user with detailed information about detected problems. Warnings are typically output to the console log, email, or as messages in the graphical user interface (GUI).

[0224] Hardware and software used

[0225] Hardware: Smartphones (e.g., iPhone® and ANDROID® devices), smart glasses (e.g., Google® Glass®, etc.).

[0226] Software: Python programming language, re module (regular expression processing), requests module (HTTP request processing).

[0227] Example of a system

[0228] The "AI Guard" app creates warning messages for users by inputting prompt sentences such as the following into a generative AI model:

[0229] Prompt Sentence Examples

[0230] "The site you are trying to visit, URL: {url}, may be fraudulent or malicious. We do not recommend visiting this site. More information: {details}"

[0231] Specific examples

[0232] For example, if a user receives a link via SMS to "http: / / phishing-site.com," the following steps will occur:

[0233] The server parses the URL and checks it against a blacklist.

[0234] Get the page content and detect the keyword "phishing".

[0235] It warns users that "this link is not safe."

[0236] In this way, the "AI Guard" system of the present invention plays a role in keeping users' digital lives safe by monitoring their digital content in real time and early detecting and warning of potential security risks.

[0237] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0238] Step 1:

[0239] The server receives the answers and files generated by other AI systems. The input is the answers and files generated by other AI systems, which are provided in the form of strings and data. The server receives this and analyzes it in the next step.

[0240] Step 2:

[0241] The server analyzes the retrieved answers and file contents using keyword filtering technology to detect suspicious keywords (e.g., phishing, malware, suspicious). The input is text data, and the output is the analysis result, indicating whether or not suspicious keywords are present. Through this process, the server makes an initial detection of suspicious content.

[0242] Step 3:

[0243] The server extracts the referrer URLs contained in the parsed answers and files. The input is the text data of the answers and files, and the output is the extracted URLs. This is then passed to the URL monitoring module.

[0244] Step 4:

[0245] The server sends an HTTP request to the URL extracted using the URL monitoring module to obtain the web page content. The input is the extracted URL, and the output is the HTML content of the web page corresponding to the URL. In this process, the server communicates over the network to obtain the page content of the target URL.

[0246] Step 5:

[0247] The server analyzes the retrieved HTML content and checks the text against predefined patterns (e.g., phishing keywords, malware download links) to detect fraudulent or malicious sites. The input is the HTML data of the web page, and the output is a judgment result on whether the page is suspicious or not. This allows the server to evaluate the safety of the URL.

[0248] Step 6:

[0249] The server retrieves program code generated by other AI systems and analyzes its contents as strings. The input is the program code, and the output is the analysis result, indicating whether or not there are suspicious code patterns. In this process, the server scans the strings of the program code with regular expressions to detect malicious code.

[0250] Step 7:

[0251] The server issues a warning to the user based on the detection results. The input is the analysis results and judgment results, and the output is a warning message. The warning is notified to the user via email, log, or GUI message. This process allows the user to receive warnings about security risks in real time.

[0252] Through the above processing steps, the server can monitor the user's digital content in real time and issue early warnings of potential security risks.

[0253] Furthermore, an emotion engine that estimates the user's emotion may be combined. That is, the identification processing unit 290 may estimate the user's emotion using the emotion identification model 59 and perform identification processing using the user's emotion.

[0254] This invention integrates an emotion engine into a system that monitors answers and files generated by other AI systems, detects suspicious content, and issues warnings to users. This system operates as software installed on a server, and aims to improve user security and experience by adjusting the content and presentation of warnings according to the user's emotional state.

[0255] Basic system configuration

[0256] The system consists of the following components:

[0257] 1. Answer and file monitoring module: This module takes answers and files generated by other AI systems and analyzes them for suspicious content.

[0258] 2. URL Monitoring Module: Monitors referenced URLs in answers and files to check whether they are fraudulent or malicious sites.

[0259] 3. Program code monitoring module: Checks the program code generated by other AI systems to monitor whether it contains malicious code.

[0260] 4. Alert module: Issues an alert to the user if a problem is detected as a result of monitoring.

[0261] 5. Emotion Engine: Recognizes user emotions and adjusts the content and presentation of warning messages based on those emotions.

[0262] Program processing

[0263] 1. Monitoring responses and files

[0264] Server: The server receives answers and files generated by other AI systems, treating the received content as strings or data.

[0265] Server: The analysis module checks the content using predefined keywords (e.g., phishing, malware, suspicious) to see if it contains any suspicious items.

[0266] 2. URL Monitoring

[0267] Server: Parses messages and content containing referral URLs and extracts the URLs.

[0268] Server: Sends an HTTP request to the extracted URL and retrieves the content of the target web page.

[0269] Server: Analyzes the retrieved page content and checks the text against predefined patterns to detect fraudulent or malicious sites, and issues a warning if a problem is detected.

[0270] 3. Program code monitoring

[0271] Server: Obtains program code generated by other AI systems. At this time, the code is obtained as a string and is subject to inspection.

[0272] Server: Regular expressions are used to check whether the program contains suspicious code patterns such as "import os", "import sys", or "exec(". The analysis module checks each line of the code and records any suspicious patterns found.

[0273] 4. Emotion recognition

[0274] Server: The emotion engine receives the user's voice, text, and facial expression data and performs emotion analysis. The emotion analysis module analyzes the user's tone of voice, words used, and facial expression changes to assess the user's emotional state (e.g., stress, anxiety, satisfaction).

[0275] Server: The emotion engine adjusts the display and content of warning messages based on the detected user emotional state. For example, if the user is feeling anxious, the warning message will be presented in a more understandable and less urgent manner.

[0276] 5. Issuing a warning

[0277] Server: A warning module runs and notifies the user of detailed information about detected problems. Warnings are mainly output to the console log, email, or as messages on the GUI.

[0278] Server: The emotion engine optimizes the priority and detail of the warning message based on the user's current emotional state, providing the user with an appropriate warning.

[0279] Specific examples

[0280] For example, when another AI system recommends a link to a user, it must verify whether the link is a fraudulent site. The server analyzes the link and, if it determines that it is likely to be a fraudulent site, warns the user, "This link is not safe." If the emotion engine detects anxiety from the user's tone of voice or facial expression, it provides a softer warning message and adds a help link if necessary. Furthermore, if the program code provided to the user by the other AI system contains fraudulent operations, the server detects this and issues a warning, saying, "This code is dangerous." If the emotion engine detects the user's state of stress, it simplifies the warning message and presents additional support options.

[0281] This allows the server to continuously monitor the outputs of other artificial intelligence systems and provide appropriate warnings based on the user's emotional state, improving both security and user experience.

[0282] The processing flow will be explained below.

[0283] Step 1:

[0284] The server receives answers and files generated by other AI systems, and the data is stored on the server as a string or file.

[0285] Step 2:

[0286] The server analyzes the responses and files it receives and checks their content against predefined keywords (e.g., phishing, malware, suspicious). The analysis module is responsible for this and checks whether any suspicious words are included.

[0287] Step 3:

[0288] The server extracts the reference URLs contained in the response or file, using regular expressions to extract the URLs from the text, and creates a list.

[0289] Step 4:

[0290] The server sends an HTTP request to the extracted URL to retrieve the content of the target web page. The request module issues an HTTP request, and the server receives the response.

[0291] Step 5:

[0292] The server performs text analysis on the retrieved web page content, checking for patterns (e.g., phishing terms, suspicious links) to detect fraudulent or malicious sites. The content is analyzed as text to see if it contains any suspicious patterns.

[0293] Step 6:

[0294] The server acquires program code generated by other AI systems and analyzes the code content. At this time, the code is acquired as a string and is the subject of inspection.

[0295] Step 7:

[0296] The server uses regular expressions to check whether the acquired program code contains suspicious code patterns such as "import os", "import sys", or "exec(". The analysis module checks each line of the code and records any suspicious patterns it finds.

[0297] Step 8:

[0298] The server uses an emotion engine to recognize the user's emotions. It analyzes the voice, text, and facial expression data obtained from the user to evaluate their emotional state (e.g., stress, anxiety, satisfaction). This analysis includes voice tone analysis, emotional keyword analysis in the text, and facial expression analysis.

[0299] Step 9:

[0300] The server adjusts the content and presentation of the warning message based on the emotional state recognized by the emotion engine. For example, if the user expresses anxiety, the server changes the warning message to be more gentle and detailed.

[0301] Step 10:

[0302] The server combines the analysis results from each module and, if any suspicious content is detected, generates a warning message that contains detailed information about the detected problem and is presented in an easy-to-understand format to the user.

[0303] Step 11:

[0304] The server notifies the user of the generated warning message. Notification methods include console log, email, or a message on the user interface. The priority and detail of the message are adjusted according to the analysis results of the emotion engine. This allows the user to quickly understand which parts are suspicious.

[0305] This allows the server to continuously monitor the outputs of other artificial intelligence systems, ensuring user security and providing warnings that take into account the user's emotional state.

[0306] Example 2

[0307] Next, a description will be given of Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0308] Conventional monitoring systems primarily focus on monitoring answers, files, reference links, and programs generated by other AI systems. However, because they do not take the user's emotional state into consideration, the content and display method of warning messages are uniform, which can lead to poor user experience. In particular, appropriate responses are required for users experiencing stress or anxiety, but these responses are insufficient.

[0309] The identification process by the identification processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means. In this invention, the server includes a means for monitoring answers and files generated by other artificial intelligence systems and checking for suspicious content, a means for monitoring reference links and detecting fraudulent or malicious pages, a means for monitoring generated programs and checking whether they contain malicious software, a means for issuing a warning when malicious content is detected, and a means for recognizing the user's emotional state and adjusting the display method and content of the warning message. This makes it possible to improve security and the user experience by providing appropriate warnings according to the user's emotional state.

[0310] "Suspicious content" is content in network responses or files that indicates security threats such as phishing or malware.

[0311] A "reference link" is a URL or web address contained in a response or file.

[0312] A "scam page" is a malicious website designed to deceive users and obtain their personal information.

[0313] A "malicious page" is a website that contains computer viruses or malware and is designed to harm a user's system.

[0314] "Generated Program" refers to program code created by another artificial intelligence system.

[0315] "Malicious software" refers to programs designed to cause damage to systems or data, including viruses, malware, and spyware.

[0316] A "warning message" is a message intended to notify the user of problems or potential risks detected by the system.

[0317] "Emotional state" refers to a user's psychological state, such as stress, anxiety, or satisfaction, and is recognized through voice, text, and facial expression data.

[0318] An "emotion recognition engine" is a software module for analyzing and evaluating a user's emotional state.

[0319] This invention integrates an emotion engine into a system that monitors answers and files generated by other AI systems, detects suspicious content, and issues warnings to users. This system operates as software installed on a server, and aims to improve user security and experience by adjusting the content and presentation of warnings according to the user's emotional state.

[0320] Basic system configuration

[0321] The system consists of the following components:

[0322] 1. Answer and file monitoring module: This module takes answers and files generated by other AI systems and analyzes them for suspicious content.

[0323] 2. URL Monitoring Module: Monitors referral links in answers and files to ensure they are not fraudulent or malicious.

[0324] 3. Program code monitoring module: Checks the program code generated by other AI systems to ensure it does not contain malicious software.

[0325] 4. Alert module: Issues an alert to the user if a problem is detected as a result of monitoring.

[0326] 5. Emotion Recognition Engine: Recognizes user emotions and adjusts the content and presentation of warning messages based on those emotions.

[0327] System processing flow and specific examples

[0328] Watch for answers and files:

[0329] Server: Periodically retrieves answers and files generated by other AI systems via API calls and database queries.

[0330] Server: Saves the obtained answers and files as string or binary data.

[0331] Server: The analysis module runs and performs text analysis on the stored data, checking for suspicious content using predefined keywords (e.g., phishing, malware, suspicious).

[0332] Extract and validate the URL:

[0333] Server: Analyzes the text data extracted from the responses and files and extracts all the URLs contained within.

[0334] Server: Sends an HTTP request to each extracted URL to retrieve the web page content, including timeout and error handling.

[0335] Server: Analyzes the HTML content of retrieved web pages and filters them to detect fraudulent or malicious sites by checking for specific keywords and patterns.

[0336] Program code monitoring:

[0337] Server: Retrieves program code generated by other AI systems in the form of strings, via API calls or database queries.

[0338] Server: Analyzes the acquired program code line by line and checks for suspicious code patterns using regular expressions. For example, it performs pattern matching on "import os", "import sys", "exec(", etc.

[0339] Emotion Recognition and Analysis:

[0340] Device: Captures the user's voice, text, and facial expression data using input devices such as a microphone, keyboard, and camera.

[0341] Server: Receives data sent from the device and passes it to the emotion engine.

[0342] Server: The emotion engine analyzes the tone of voice, words used, and facial expressions to assess the user's emotional state.

[0343] Sending a warning:

[0344] Server: The alert module is launched to collect detailed information about the problem detected in the previous step.

[0345] Server: Using the evaluation results of the emotion engine, generate a warning message according to the user's emotional state. If the user is feeling anxious, use a less urgent message such as "This link is not safe. Please rest assured, we will take action immediately."

[0346] Server: Notifies the user of a warning message in a variety of formats, including console log, email, or a GUI popup message.

[0347] Examples of concrete examples and prompts

[0348] For example, suppose a user is recommended a link and needs to verify whether it is a fraudulent page. The server analyzes the link and, if it determines that it is likely a fraudulent page, warns the user that "This link is not safe." If the emotion engine detects anxiety in the user's tone of voice or facial expression, it will provide a softer warning message and, if necessary, add a help link.

[0349] Example prompt sentence:

[0350] "Please explain how your system verifies the possibility that other AI-generated links may be fraudulent sites and issues an appropriate warning to the user if it finds them. Also, please explain how you would recognize the user's emotions and adjust the warning message accordingly."

[0351] In this way, the server, terminal, and user work together to carry out the process, creating a system that improves the user experience while ensuring the security of the information provided by the generative AI model.

[0352] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0353] Step 1: Get answers and files

[0354] Input: Answers generated from other artificial intelligence systems or data from files.

[0355] Server: Periodically retrieves generated answers and files from other AI systems using REST API calls and database queries.

[0356] Server: Saves the retrieved data as string data or binary data.

[0357] Output: Saved answers and file data.

[0358] Step 2: Monitor and analyze responses and files

[0359] Input: Saved answers and file data.

[0360] Server: The analysis module runs and performs text analysis on the stored data, checking for suspicious content using predefined keywords (e.g., phishing, malware, suspicious).

[0361] Server: If suspicious content is detected as a result of text analysis, it will record the content in a log file and notify the warning module.

[0362] Output: Analysis results and log file updates.

[0363] Step 3: Extracting the URL

[0364] Input: Saved answers and file data.

[0365] Server: Analyzes the text data extracted from the responses and files and extracts all the URLs contained within.

[0366] Server: Use text analysis tools and libraries to identify and catalog URL patterns in responses and files.

[0367] Output: A list of extracted URLs.

[0368] Step 4: Verify the URL

[0369] Input: A list of extracted URLs.

[0370] Server: Sends HTTP requests to each URL to retrieve the web page content, including timeout and error handling.

[0371] Server: Analyzes the HTML of the retrieved page content and filters it to detect fraudulent or malicious sites, checking for specific keywords and suspicious patterns.

[0372] Server: Records the results of URL validation in a log file and notifies the alert module if a problem is detected.

[0373] Output: URL validation result and log file update.

[0374] Step 5: Monitoring the program code

[0375] Input: Program code generated by another artificial intelligence system.

[0376] Server: Retrieves the generated program code in string form, either via an API call or a database query.

[0377] Server: Analyzes the acquired program code line by line and checks for suspicious code patterns (e.g., import os, import sys, exec()) using regular expressions.

[0378] Server: If pattern matching detects suspicious code, identify the line and log it with a detailed comment.

[0379] Output: Code analysis results and log file updates.

[0380] Step 6: Emotion Recognition and Analysis

[0381] Input: User voice, text, and facial expression data.

[0382] Device: Captures the user's voice, text, and facial expression data using input devices such as a microphone, keyboard, and camera.

[0383] Server: Receives data sent from the device and passes it to the emotion engine.

[0384] Server: The emotion engine analyzes the tone of voice, words used, and facial expressions to assess the user's emotional state.

[0385] Server: Quantifies or categorizes the emotion evaluation results and notifies the warning module.

[0386] Output: Emotion data reflected in the emotion rating results and warnings.

[0387] Step 7: Sending an alert

[0388] Input: Analysis results and emotion evaluation results.

[0389] Server: The warning module is launched to collect detailed information about the problem detected in the previous step. Using the evaluation results of the emotion engine, a warning message is generated based on the user's emotional state.

[0390] Server: Notifies each user of warning messages in a format suitable for them, such as console logs, emails, or GUI popup messages.

[0391] User: Receives a warning message and takes action. Users can also view the provided help links and support options.

[0392] Output: Warning message provided to user and log file updates.

[0393] In this way, by having the server, terminal, and user work together to carry out processing, we have created a system that improves the user experience while ensuring the security of the information provided by the generative AI model.

[0394] (Application example 2)

[0395] Next, a description will be given of Application Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0396] The monitoring of answers, files, reference URLs, and program code generated by conventional AI systems has limited ability to detect suspicious content and malicious sites, making it difficult to fully guarantee user security. Furthermore, these systems issue warnings without taking into account the user's emotional state, which can lead to a poor user experience.

[0397] The identification processing by the identification processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means. In this invention, the server includes means for monitoring answers and files generated by other artificial intelligence systems and checking for suspicious content, means for monitoring reference URLs and detecting fraudulent or malicious sites, means for monitoring generated program code and checking whether it contains malicious programs, and means for issuing a warning based on the user's emotional state when malicious content is detected. This makes it possible to improve user security and the user experience.

[0398] "Answers and files generated by other artificial intelligence systems" refers to responses and document data automatically created by external artificial intelligence systems.

[0399] "Means for checking for suspicious content" refers to a method or device that has the functionality to analyze and detect fraudulent or malicious content contained in messages or files.

[0400] "Means for monitoring referring URLs to detect fraudulent or malicious sites" refers to a method or device for monitoring URLs contained in responses or files and determining whether the linked site is a fraudulent or malicious site.

[0401] "Means for monitoring generated program code and checking whether it contains malicious code" refers to a method or device for monitoring program code generated by another artificial intelligence system and detecting malicious code contained therein.

[0402] "Means for issuing a warning based on the user's emotional state when malicious content is detected" refers to a method or device for issuing a warning in an optimal manner based on the results of an emotional analysis of the user in response to detected suspicious content.

[0403] "Means for assessing emotional state" means a method or device for analyzing a user's text or other information to determine whether the user is in a positive, negative, or neutral emotional state.

[0404] "Means for adjusting the display method and content of a warning message" refers to a method or device for appropriately adjusting a warning message according to the user's emotional state and displaying it in the most suitable format and content for the user.

[0405] The system of the present invention includes means for monitoring answers and files generated by other artificial intelligence systems and checking for suspicious content, means for monitoring referred URLs and detecting fraudulent or malicious sites, means for monitoring generated program code and checking whether it contains malicious programs, and means for issuing a warning based on the user's emotional state when malicious content is detected.

[0406] 1. Answers and file monitoring methods:

[0407] The server monitors answers and files generated by external AI systems in real time. This includes analyzing the content using predefined keywords (e.g., phishing, malware, suspicious) to check for suspicious content. For example, if the monitored text contains a keyword such as "phishing," the server will determine that the text is suspicious.

[0408] 2. URL monitoring methods:

[0409] The server extracts the URL contained in the response or file and sends an HTTP request to retrieve the content of the web page pointed to by the URL. The retrieved page content is then analyzed and checked against patterns designed to detect fraudulent or malicious sites. For example, a message containing the URL "http: / / example-phishing.com" is analyzed, and if the URL is determined to be a fraudulent site, an alert is issued.

[0410] 3. Program code monitoring methods:

[0411] The server monitors program code generated by external artificial intelligence systems, checking it using regular expressions to see if it contains malicious program patterns such as "import os," "import sys," or "exec(." If suspicious code is detected, it records its contents and issues a warning to the user.

[0412] 4. Emotion recognition means:

[0413] The server receives the user's voice, text, and facial expression data and performs emotion analysis using an emotion engine. This analyzes the user's tone of voice, words used, and facial expression changes to assess the user's emotional state (e.g., stress, anxiety, satisfaction), and adjusts the display method and content of warning messages descriptively based on the user's emotional state.

[0414] 5. Warning Method:

[0415] The server optimizes the content and presentation of warning messages based on the user's emotional state, as detected by the emotion engine. For example, if the user is feeling anxious, the server will present the warning message in a more understandable and less stressful format, and also add a help link with specific actions to take.

[0416] Hardware and software used

[0417] Server: Server hardware that includes a central processing unit (CPU), memory, and storage.

[0418] Emotion Recognition Engine: Uses TextBlob as a model for emotion analysis.

[0419] URL monitoring: Uses the Requests library to send HTTP requests.

[0420] Program code monitoring: Uses the re library for regular expression analysis.

[0421] Sending email: Uses the smtplib library to send alert messages.

[0422] Specific examples

[0423] For example, when a user receives a message, it may contain a link to a fraudulent website. The server analyzes the URL and, if it determines that the site is likely fraudulent, warns the user, "This link is not safe." At the same time, the emotion engine analyzes the user's text, "I have a bad feeling about this," and, if it detects a negative emotional state, it makes the warning message more gentle and adds a help link if necessary.

[0424] Prompt sentence for generative AI model

[0425] "Analyze text entered by users and assess its emotional state (positive, negative, neutral). Develop an application that monitors provided messages, URLs, and program code for suspicious content and issues alerts accordingly."

[0426] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[0427] Step 1:

[0428] The server receives answers or files generated by other AI systems. These answers or files are input as string data. The server analyzes the content using predefined keywords (e.g., phishing, malware, suspicious) to check for suspicious items. The results of this analysis are provided as output.

[0429] Step 2:

[0430] The server automatically extracts the URLs contained in the response or file. The list of URLs is input. The server sends an HTTP request to each URL to retrieve the web page content. This content is output as text data.

[0431] Step 3:

[0432] The server analyzes the content of the retrieved web page and checks the text against predefined patterns to detect fraudulent or malicious sites. If a fraudulent or malicious site is detected, the URL and reason for the detection are printed.

[0433] Step 4:

[0434] The server receives program code generated by other AI systems. The code is input as string data. The server uses regular expressions to check whether the program contains suspicious code patterns such as "import os," "import sys," or "exec(." The server then outputs the analysis results.

[0435] Step 5:

[0436] The server receives the user's voice, text, and facial expression data. These data are used as input. The emotion engine analyzes the user's tone of voice, the words used, and changes in facial expressions to assess the user's emotional state (e.g., stress, anxiety, satisfaction). This emotion assessment is output.

[0437] Step 6:

[0438] The server adjusts the display method and content of the warning message based on the user's emotional state detected by the emotion engine. The warning message template is used as input, and a warning message adapted to the user's emotional state is generated as the final output.

[0439] Step 7:

[0440] The device receives the optimized warning message sent from the server and displays it to the user, specifically, a warning message containing gentle language and additional support information according to the user's emotional state.

[0441] 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.

[0442] 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> ), Gemini (registered trademark) (Internet search <url: https: gemini.google.com ?hl="ja">) and other generation AIs. The data generation model 58 is obtained by performing deep learning on a neural network. A prompt including an instruction is input to the data generation model 58, and inference data such as voice data indicating voice, text data indicating text, and image data indicating an image is also input. The data generation model 58 performs inference on the input inference data in accordance with the instruction indicated by the prompt, and outputs the inference result in a data format such as voice data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[0443] In the above embodiment, an example in which the specific process is performed by the data processing device 12 has been given, but the technology of the present disclosure is not limited to this, and the specific process may be performed by the smart device 14.

[0444] [Second embodiment]

[0445] FIG. 3 shows an example of the configuration of a data processing system 210 according to the second embodiment.

[0446] 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.

[0447] The data processing device 12 includes a computer 22, a database 24, and a communication I / F 26. The computer 22 is an example of a "computer" according to the technology of the present disclosure. The computer 22 includes a processor 28, a RAM 30, and a storage 32. The processor 28, the RAM 30, and the storage 32 are connected to a bus 34. The database 24 and the communication I / F 26 are also connected to the bus 34. The communication I / F 26 is connected to a network 54. Examples of the network 54 include a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0448] 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.

[0449] The microphone 238 receives instructions and the like from the user 20 by receiving voice uttered by the user 20. The microphone 238 captures the voice uttered by the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio in accordance with instructions from the processor 46.

[0450] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an imaging element such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the surroundings of user 20 (for example, an imaging range defined by an angle of view equivalent to the field of vision of a typical healthy person).

[0451] 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.

[0452] 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.

[0453] The specific processing program 56 is an example of a "program" according to the technology of the present disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific processing is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific processing program 56 executed on the RAM 30.

[0454] 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.

[0455] In the smart glasses 214, the reception output process is performed by the processor 46. A reception output program 60 is stored in the storage 50. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output process is realized by the processor 46 operating as the control unit 46A in accordance with the reception output program 60 executed on the RAM 48.

[0456] Next, a description will be given of the identification process performed by the identification processing unit 290 of the data processing device 12. In the following description, the data processing device 12 will be referred to as the "server" and the smart glasses 214 will be referred to as the "terminal."

[0457] The present invention is a system that monitors answers and files generated by other AI systems, detects suspicious content, and issues a warning to users. This system operates as software installed on a server, and plays a role in enhancing user security.

[0458] Basic system configuration

[0459] The system consists of the following components:

[0460] 1. Answer and file monitoring module: This module takes answers and files generated by other AI systems and analyzes them for suspicious content.

[0461] 2. URL Monitoring Module: Monitors referenced URLs in answers and files to check whether they are fraudulent or malicious sites.

[0462] 3. Program code monitoring module: Checks the program code generated by other AI systems to monitor whether it contains malicious code.

[0463] 4. Alert module: Issues an alert to the user if a problem is detected as a result of monitoring.

[0464] Program processing

[0465] 1. Monitoring responses and files

[0466] Server: The server receives answers and files generated by other AI systems, treating the received content as strings or data.

[0467] Server: The analysis module checks the content using predefined keywords (e.g., phishing, malware, suspicious) and issues a warning if any suspicious items are included.

[0468] 2. URL Monitoring

[0469] Server: Parses messages and content containing referral URLs and extracts the URLs.

[0470] Server: Sends an HTTP request to the extracted URL and retrieves the page content.

[0471] Server: Analyzes the retrieved page content and checks the text against predefined patterns to detect fraudulent or malicious sites, and issues a warning if a problem is detected.

[0472] 3. Program code monitoring

[0473] Server: Obtains program code generated by other AI systems, and analyzes the instructions in the code as strings.

[0474] Server: Uses regular expressions to check whether a program contains suspicious code patterns such as "import os", "import sys", or "exec(", and issues a warning if they are found.

[0475] 4. Issuing a warning

[0476] Server: A warning module runs and notifies the user of detailed information about detected problems. Warnings are mainly output to the console log, email, or as messages on the GUI.

[0477] Specific examples

[0478] For example, when another AI system recommends a link to a user, the server must verify whether the link is a fraudulent site. If the server analyzes the link and determines that it is likely to be a fraudulent site, it warns the user that "this link is not safe." Also, if the program code provided to the user by the other AI system contains fraudulent operations, the server detects this and issues a warning that "this code is dangerous."

[0479] This will allow users to minimize the risk from malicious AI technology and use the system with peace of mind, and this system will play an important role in improving user security.

[0480] The processing flow will be explained below.

[0481] Step 1:

[0482] The server receives answers and files generated by other AI systems, and the data is stored on the server as a string or file.

[0483] Step 2:

[0484] The server analyzes the responses and files it receives and checks their content against predefined keywords (e.g., phishing, malware, suspicious). The analysis module is responsible for this and checks whether any suspicious words are included.

[0485] Step 3:

[0486] The server extracts the reference URLs contained in the response or file, using regular expressions to extract the URLs from the text, and creates a list.

[0487] Step 4:

[0488] The server sends an HTTP request to the extracted URL to retrieve the content of the target web page. The request module issues an HTTP request and receives the response from the server.

[0489] Step 5:

[0490] The server performs text analysis on the retrieved web page content, checking for patterns (e.g., phishing terms, suspicious links) to detect fraudulent or malicious sites. The content is analyzed as text to see if it contains any suspicious patterns.

[0491] Step 6:

[0492] The server acquires program code generated by other AI systems and analyzes the code content. At this time, the code is acquired as a string and is the subject of inspection.

[0493] Step 7:

[0494] The server uses regular expressions to check whether the acquired program code contains suspicious code patterns such as "import os", "import sys", or "exec(". The analysis module checks each line of the code and records any suspicious patterns it finds.

[0495] Step 8:

[0496] The server combines the analysis results from each module and, if any suspicious content is detected, generates a warning message that contains detailed information about the detected problem and is presented in a user-friendly format.

[0497] Step 9:

[0498] The server notifies the user of any generated warning messages via console logs, email, or messages on the user interface, allowing the user to quickly identify suspicious areas.

[0499] This allows the server to continuously monitor the outputs of other artificial intelligence systems, creating a system that ensures user security.

[0500] Example 1

[0501] Next, a description will be given of Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[0502] Answers and files generated by conventional AI systems sometimes contain malicious content, fraudulent websites, or malicious code, posing a threat to user security. Furthermore, there is a lack of systems that can efficiently detect such malicious content and quickly warn users. Therefore, there is a need for a reliable monitoring system that allows users of AI systems to use them with peace of mind.

[0503] The specific processing by the specific processing unit 290 of the data processing device 12 in the first embodiment is realized by the following means.

[0504] In this invention, the server includes means for monitoring answers and files generated by other artificial intelligence systems and analyzing suspicious content using predefined keywords, means for monitoring referred URLs and analyzing page content obtained by sending HTTP requests to detect fraudulent or malicious sites, means for monitoring generated program code and checking specific code patterns with regular expressions to detect whether malicious programs are included, and means for notifying users of detailed information about detected problems via console logs, email, or GUI messages. This makes it possible to effectively detect security risks contained in content generated by artificial intelligence systems and promptly notify users.

[0505] "Answers and files" refers to text data and document data generated by other artificial intelligence systems.

[0506] "Monitoring" means the act of continuously or periodically checking or auditing a particular object.

[0507] "Keywords" refers to predefined important phrases or words for a specific purpose.

[0508] "Referring URL" refers to a string of characters that indicates a link to a web page or resource.

[0509] "HTTP request" refers to a request sent by a web browser or client application to retrieve information from a web server.

[0510] "Page Content" refers to the text, images, scripts, and other information displayed on a web page.

[0511] "Fraudulent Site" refers to a fraudulent website designed to deceive users.

[0512] "Malicious Site" means a website whose purpose is to steal user information, cause damage to systems, or perform other fraudulent activities.

[0513] "Program code" refers to the source code for writing a computer program.

[0514] A "code pattern" refers to a set of notations or instructions that have a specific purpose or function.

[0515] "Regular expression" refers to a formal language for describing specific patterns.

[0516] "Console log" refers to character-based output that displays system or application status and error messages.

[0517] "Mail" means electronic mail, a means of sending and receiving messages in digital form.

[0518] "GUI Message" means a warning or informational message displayed on a graphical user interface.

[0519] The present invention is a system that monitors answers and files generated by other AI systems, detects suspicious content, and issues a warning to the user. Specific embodiments for implementing this system are described below.

[0520] Basic system configuration

[0521] The system runs as software installed on a server and consists of the following components:

[0522] 1. Answer and file monitoring module

[0523] 2. URL Monitoring Module

[0524] 3. Program Code Monitoring Module

[0525] 4. Warning Module

[0526] Monitor responses and files

[0527] server

[0528] The server retrieves answers and files generated by other AI systems, using API requests and file system monitoring.

[0529] The retrieved answers and file contents are analyzed as text data, and the server searches the text using predefined keywords (e.g., phishing, malware, suspicious) to extract suspicious content.

[0530] URL monitoring

[0531] server

[0532] The server extracts URLs from the responses and files and generates a list of URLs using regular expressions.

[0533] Send an HTTP request to the extracted URL and get the response.

[0534] The text content of the response is analyzed and matched against patterns (e.g., characteristics of phishing pages) to detect fraudulent or malicious websites.

[0535] Program code monitoring

[0536] server

[0537] The server retrieves program code generated by other artificial intelligence systems from specific folders or APIs.

[0538] The commands in the acquired program code are analyzed as strings and checked for specific patterns (e.g., "import os", "import sys", "exec()") using regular expressions.

[0539] Sending a warning

[0540] server

[0541] The server generates detailed warning messages about any problems it detects, including the type, location, and detailed description of the problem detected.

[0542] The generated warning messages are not only output to the console log, but also notified to the user by email and displayed as a popup message on the GUI.

[0543] Specific examples

[0544] A specific example of operation is shown below.

[0545] Example 1: If the server retrieves a new answer and it contains the keyword "phishing", the server will notify the user "Warning: This answer contains suspicious content."

[0546] Example 2: If the server sends an HTTP request to a URL extracted from a file and the result has characteristics of a phishing site, the server notifies the user with "Warning: This URL is not safe."

[0547] Example 3: When the server parses a new code file and it contains the instruction "exec(", the server will notify the user "Warning: this code is dangerous".

[0548] Prompt Sentence Examples

[0549] Check if this link is safe

[0550] "Check this code for illegal instructions"

[0551] By inputting such prompt sentences into the generative AI model, the server will initiate a specific analysis process, allowing users to obtain reliable information and minimizing security risks.

[0552] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0553] Step 1:

[0554] Get answers and files

[0555] Server: Retrieves answers and files generated by other AI systems, using API requests and file system monitoring.

[0556] Input: Answers or files generated by other AI systems.

[0557] Output: Text data of the obtained answers and files.

[0558] Step 2:

[0559] Check the contents

[0560] Server: Analyzes the obtained answers and file contents as text data. Analyzes suspicious content using predefined keywords (e.g., phishing, malware, suspicious).

[0561] Input: Text data from the answers or files obtained.

[0562] Output: Result on whether or not the content is suspicious.

[0563] What happens: The server searches for text against a list of keywords and flags it as suspicious if it contains a matching keyword.

[0564] Step 3:

[0565] URL extraction

[0566] Server: Extract URLs from responses and files using regular expressions.

[0567] Input: The text data from the parsed answer or file.

[0568] Output: The extracted URL list.

[0569] What it does: The server searches for strings that start with "http: / / " or "https: / / " and lists all URLs.

[0570] Step 4:

[0571] URL validation

[0572] Server: Sends an HTTP request to the extracted URL and obtains the response data.

[0573] Input: The extracted URL list.

[0574] Output: Response data for the URL.

[0575] Specific operation: The server sends an HTTP request for each URL and saves the page content of the response it receives.

[0576] Step 5:

[0577] Analyzing page content

[0578] Server: Analyzes the response data and matches it with patterns (e.g., characteristics of phishing sites) to detect fraudulent or malicious sites.

[0579] Input: Response data for the URL.

[0580] Output: Detected fraudulent and malicious sites.

[0581] What it does: The server performs pattern matching on the text content of the response and flags any fraudulent or malicious features detected.

[0582] Step 6:

[0583] Obtaining the program code

[0584] Server: Retrieves program code generated by other AI systems from specific folders or APIs.

[0585] Input: Program code generated by another artificial intelligence system.

[0586] Output: Text data of the retrieved program code.

[0587] What it does: The server monitors newly added code files and incorporates them one by one.

[0588] Step 7:

[0589] Code content analysis

[0590] Server: Analyzes the program code and checks for specific code patterns (e.g., "import os", "import sys", "exec()") using regular expressions.

[0591] Input: Text data of the retrieved program code.

[0592] Output: Suspicious code pattern detection results.

[0593] What it does: The server parses instructions in your code using regular expressions and flags when certain patterns are detected.

[0594] Step 8:

[0595] Generate a warning message

[0596] Server: Generates detailed warning messages about detected problems.

[0597] Input: Suspicious content, URL, or code pattern detection results.

[0598] Output: A detailed warning message.

[0599] What happens: The server generates a message containing the type, location, and detailed description of the problem detected.

[0600] Step 9:

[0601] User Notification

[0602] Server: Generated warning messages are output to the console log and notified to the user via email or GUI message.

[0603] Input: A detailed warning message.

[0604] Output: The warning message that was sent to the user.

[0605] Specific behavior: The server not only logs the warning to the console, but also sends an email to the user and displays a warning as a popup message in the GUI.

[0606] (Application example 1)

[0607] Next, a description will be given of Application Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[0608] With the development of modern information and communication technology, security risks are increasing for answers, files, and program code generated by other artificial intelligence systems. Furthermore, the potential for suspicious URLs and malicious programs contained in this content to cause harm to users is also increasing. In particular, with the widespread use of smart devices, real-time monitoring and alerts are becoming increasingly important. However, current systems lack the means to effectively address these multifaceted threats.

[0609] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 1 is realized by the following means.

[0610] In this invention, the server includes means for monitoring answers and files generated by other artificial intelligence systems and checking for suspicious content, means for monitoring referred URLs and detecting fraudulent or malicious sites, means for monitoring generated program code and checking whether it contains malicious programs, means for issuing a warning when malicious content is detected, and means for operating on the smart device and monitoring the user's digital content in real time, thereby enabling real-time monitoring and warning of the digital content used by the user.

[0611] "Other artificial intelligence systems" refers to artificial intelligence technologies operated or utilized by users to analyze and generate data.

[0612] "Answer or file" refers to text, documents, images, audio, or other forms of data generated by other artificial intelligence systems.

[0613] "Suspicious Content" refers to elements within a file or message that may indicate phishing, malware, spam, or other malicious content.

[0614] "Referring URL" refers to a website address or link contained in a response or file.

[0615] A "fraudulent site" refers to a malicious website that attempts to fraudulently obtain users' personal information or money.

[0616] "Malicious Sites" refers to websites that install malware on users' devices or perform other malicious activities.

[0617] "Program Code" refers to the source code of software generated by other artificial intelligence systems.

[0618] "Malicious Program" refers to source code or a program designed to perform unauthorized operations or undesirable behavior.

[0619] "Warning" refers to a notification or alert issued to a user when the system detects suspicious or malicious content.

[0620] "Smart devices" refers to electronic devices such as mobile phones, tablets, and smart glasses that have internet connectivity.

[0621] "Digital content" refers to electronically stored information or data, such as web pages, messages, files, and program code.

[0622] "Real-time" refers to data being processed immediately at the moment it is generated or acquired.

[0623] This invention provides a security system called "AI Guard," which monitors answers and files generated by other artificial intelligence systems and analyzes whether they contain suspicious content. It also monitors referred URLs to determine whether they are fraudulent or malicious sites, and checks generated program code to monitor whether they contain malicious programs. Based on these monitoring results, it issues a warning to the user when malicious content is detected. Furthermore, this system runs on smart devices and can monitor users' digital content in real time.

[0624] System Configuration

[0625] The system consists of the following elements:

[0626] 1. Answer and file monitoring module:

[0627] The server receives answers and files generated by other AI systems, whose contents are processed as strings or data.

[0628] The server analyzes the content using predefined keywords (e.g., phishing, malware, suspicious) and issues a warning if it contains any suspicious items.

[0629] 2. URL Monitoring Module:

[0630] Parse messages and content that contain URLs and extract them.

[0631] An HTTP request is sent to the extracted URL to retrieve the page content.

[0632] It analyzes the retrieved page content, checks the text against defined patterns to detect fraudulent or malicious sites, and issues a warning if a problem is detected.

[0633] 3. Program code monitoring module:

[0634] It acquires program code generated by other AI systems and analyzes the instructions in the code as strings of characters.

[0635] It uses regular expressions to check whether a program contains suspicious code patterns such as "import os", "import sys", or "exec(", and issues a warning if they are found.

[0636] 4. Alert sending module:

[0637] The server includes a module to notify the user with detailed information about detected problems. Warnings are typically output to the console log, email, or as messages in the graphical user interface (GUI).

[0638] Hardware and software used

[0639] Hardware: Smartphones (e.g., iPhones and Android devices), smart glasses (e.g., Google Glass).

[0640] Software: Python programming language, re module (regular expression processing), requests module (HTTP request processing).

[0641] Example of a system

[0642] The "AI Guard" app creates warning messages for users by inputting prompt sentences such as the following into a generative AI model:

[0643] Prompt Sentence Examples

[0644] "The site you are trying to visit, URL: {url}, may be fraudulent or malicious. We do not recommend visiting this site. More information: {details}"

[0645] Specific examples

[0646] For example, if a user receives a link via SMS to "http: / / phishing-site.com," the following steps will occur:

[0647] The server parses the URL and checks it against a blacklist.

[0648] Get the page content and detect the keyword "phishing".

[0649] It warns users that "this link is not safe."

[0650] In this way, the "AI Guard" system of the present invention plays a role in keeping users' digital lives safe by monitoring their digital content in real time and early detecting and warning of potential security risks.

[0651] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0652] Step 1:

[0653] The server receives the answers and files generated by other AI systems. The input is the answers and files generated by other AI systems, which are provided in the form of strings and data. The server receives this and analyzes it in the next step.

[0654] Step 2:

[0655] The server analyzes the retrieved answers and file contents using keyword filtering technology to detect suspicious keywords (e.g., phishing, malware, suspicious). The input is text data, and the output is the analysis result, indicating whether or not suspicious keywords are present. Through this process, the server makes an initial detection of suspicious content.

[0656] Step 3:

[0657] The server extracts the referrer URLs contained in the parsed answers and files. The input is the text data of the answers and files, and the output is the extracted URLs. This is then passed to the URL monitoring module.

[0658] Step 4:

[0659] The server sends an HTTP request to the URL extracted using the URL monitoring module to obtain the web page content. The input is the extracted URL, and the output is the HTML content of the web page corresponding to the URL. In this process, the server communicates over the network to obtain the page content of the target URL.

[0660] Step 5:

[0661] The server analyzes the retrieved HTML content and checks the text against predefined patterns (e.g., phishing keywords, malware download links) to detect fraudulent or malicious sites. The input is the HTML data of the web page, and the output is a judgment result on whether the page is suspicious or not. This allows the server to evaluate the safety of the URL.

[0662] Step 6:

[0663] The server retrieves program code generated by other AI systems and analyzes its contents as strings. The input is the program code, and the output is the analysis result, indicating whether or not there are suspicious code patterns. In this process, the server scans the strings of the program code with regular expressions to detect malicious code.

[0664] Step 7:

[0665] The server issues a warning to the user based on the detection results. The input is the analysis results and judgment results, and the output is a warning message. The warning is notified to the user via email, log, or GUI message. This process allows the user to receive warnings about security risks in real time.

[0666] Through the above processing steps, the server can monitor the user's digital content in real time and issue early warnings of potential security risks.

[0667] Furthermore, an emotion engine that estimates the user's emotion may be further combined. That is, the identification processing unit 290 may estimate the user's emotion using the emotion identification model 59, and perform identification processing using the user's emotion.

[0668] This invention integrates an emotion engine into a system that monitors answers and files generated by other AI systems, detects suspicious content, and issues warnings to users. This system operates as software installed on a server, and aims to improve user security and experience by adjusting the content and presentation of warnings according to the user's emotional state.

[0669] Basic system configuration

[0670] The system consists of the following components:

[0671] 1. Answer and file monitoring module: This module takes answers and files generated by other AI systems and analyzes them for suspicious content.

[0672] 2. URL Monitoring Module: Monitors referenced URLs in answers and files to check whether they are fraudulent or malicious sites.

[0673] 3. Program code monitoring module: Checks the program code generated by other AI systems to monitor whether it contains malicious code.

[0674] 4. Alert module: Issues an alert to the user if a problem is detected as a result of monitoring.

[0675] 5. Emotion Engine: Recognizes user emotions and adjusts the content and presentation of warning messages based on those emotions.

[0676] Program processing

[0677] 1. Monitoring responses and files

[0678] Server: The server receives answers and files generated by other AI systems, treating the received content as strings or data.

[0679] Server: The analysis module checks the content using predefined keywords (e.g., phishing, malware, suspicious) to see if it contains any suspicious items.

[0680] 2. URL Monitoring

[0681] Server: Parses messages and content containing referral URLs and extracts the URLs.

[0682] Server: Sends an HTTP request to the extracted URL and retrieves the content of the target web page.

[0683] Server: Analyzes the retrieved page content and checks the text against predefined patterns to detect fraudulent or malicious sites, and issues a warning if a problem is detected.

[0684] 3. Program code monitoring

[0685] Server: Obtains program code generated by other AI systems. At this time, the code is obtained as a string and is subject to inspection.

[0686] Server: Regular expressions are used to check whether the program contains suspicious code patterns such as "import os", "import sys", or "exec(". The analysis module checks each line of the code and records any suspicious patterns found.

[0687] 4. Emotion recognition

[0688] Server: The emotion engine receives the user's voice, text, and facial expression data and performs emotion analysis. The emotion analysis module analyzes the user's tone of voice, words used, and facial expression changes to assess the user's emotional state (e.g., stress, anxiety, satisfaction).

[0689] Server: The emotion engine adjusts the display and content of warning messages based on the detected user emotional state. For example, if the user is feeling anxious, the warning message will be presented in a more understandable and less urgent manner.

[0690] 5. Issuing a warning

[0691] Server: A warning module runs and notifies the user of detailed information about detected problems. Warnings are mainly output to the console log, email, or as messages on the GUI.

[0692] Server: The emotion engine optimizes the priority and detail of the warning message based on the user's current emotional state, providing the user with an appropriate warning.

[0693] Specific examples

[0694] For example, when another AI system recommends a link to a user, it must verify whether the link is a fraudulent site. The server analyzes the link and, if it determines that it is likely to be a fraudulent site, warns the user, "This link is not safe." If the emotion engine detects anxiety from the user's tone of voice or facial expression, it provides a softer warning message and adds a help link if necessary. Furthermore, if the program code provided to the user by the other AI system contains fraudulent operations, the server detects this and issues a warning, saying, "This code is dangerous." If the emotion engine detects the user's state of stress, it simplifies the warning message and presents additional support options.

[0695] This allows the server to continuously monitor the outputs of other artificial intelligence systems and provide appropriate warnings based on the user's emotional state, improving both security and user experience.

[0696] The processing flow will be explained below.

[0697] Step 1:

[0698] The server receives answers and files generated by other AI systems, and the data is stored on the server as a string or file.

[0699] Step 2:

[0700] The server analyzes the responses and files it receives and checks their content against predefined keywords (e.g., phishing, malware, suspicious). The analysis module is responsible for this and checks whether any suspicious words are included.

[0701] Step 3:

[0702] The server extracts the reference URLs contained in the response or file, using regular expressions to extract the URLs from the text, and creates a list.

[0703] Step 4:

[0704] The server sends an HTTP request to the extracted URL to retrieve the content of the target web page. The request module issues an HTTP request, and the server receives the response.

[0705] Step 5:

[0706] The server performs text analysis on the retrieved web page content, checking for patterns (e.g., phishing terms, suspicious links) to detect fraudulent or malicious sites. The content is analyzed as text to see if it contains any suspicious patterns.

[0707] Step 6:

[0708] The server acquires program code generated by other AI systems and analyzes the code content. At this time, the code is acquired as a string and is the subject of inspection.

[0709] Step 7:

[0710] The server uses regular expressions to check whether the acquired program code contains suspicious code patterns such as "import os", "import sys", or "exec(". The analysis module checks each line of the code and records any suspicious patterns it finds.

[0711] Step 8:

[0712] The server uses an emotion engine to recognize the user's emotions. It analyzes the voice, text, and facial expression data obtained from the user to evaluate their emotional state (e.g., stress, anxiety, satisfaction). This analysis includes voice tone analysis, emotional keyword analysis in the text, and facial expression analysis.

[0713] Step 9:

[0714] The server adjusts the content and presentation of the warning message based on the emotional state recognized by the emotion engine. For example, if the user expresses anxiety, the server changes the warning message to be more gentle and detailed.

[0715] Step 10:

[0716] The server combines the analysis results from each module and, if any suspicious content is detected, generates a warning message that contains detailed information about the detected problem and is presented in an easy-to-understand format to the user.

[0717] Step 11:

[0718] The server notifies the user of the generated warning message. Notification methods include console log, email, or a message on the user interface. The priority and detail of the message are adjusted according to the analysis results of the emotion engine. This allows the user to quickly understand which parts are suspicious.

[0719] This allows the server to continuously monitor the outputs of other artificial intelligence systems, ensuring user security and providing warnings that take into account the user's emotional state.

[0720] Example 2

[0721] Next, a description will be given of Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[0722] Conventional monitoring systems primarily focus on monitoring answers, files, reference links, and programs generated by other AI systems. However, because they do not take the user's emotional state into consideration, the content and display method of warning messages are uniform, which can lead to poor user experience. In particular, appropriate responses are required for users experiencing stress or anxiety, but these responses are insufficient.

[0723] The identification process by the identification processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means. In this invention, the server includes a means for monitoring answers and files generated by other artificial intelligence systems and checking for suspicious content, a means for monitoring reference links and detecting fraudulent or malicious pages, a means for monitoring generated programs and checking whether they contain malicious software, a means for issuing a warning when malicious content is detected, and a means for recognizing the user's emotional state and adjusting the display method and content of the warning message. This makes it possible to improve security and the user experience by providing appropriate warnings according to the user's emotional state.

[0724] "Suspicious content" is content in network responses or files that indicates security threats such as phishing or malware.

[0725] A "reference link" is a URL or web address contained in a response or file.

[0726] A "scam page" is a malicious website designed to deceive users and obtain their personal information.

[0727] A "malicious page" is a website that contains computer viruses or malware and is designed to harm a user's system.

[0728] "Generated Program" refers to program code created by another artificial intelligence system.

[0729] "Malicious software" refers to programs designed to cause damage to systems or data, including viruses, malware, and spyware.

[0730] A "warning message" is a message intended to notify the user of problems or potential risks detected by the system.

[0731] "Emotional state" refers to a user's psychological state, such as stress, anxiety, or satisfaction, and is recognized through voice, text, and facial expression data.

[0732] An "emotion recognition engine" is a software module for analyzing and evaluating a user's emotional state.

[0733] This invention integrates an emotion engine into a system that monitors answers and files generated by other AI systems, detects suspicious content, and issues warnings to users. This system operates as software installed on a server, and aims to improve user security and experience by adjusting the content and presentation of warnings according to the user's emotional state.

[0734] Basic system configuration

[0735] The system consists of the following components:

[0736] 1. Answer and file monitoring module: This module takes answers and files generated by other AI systems and analyzes them for suspicious content.

[0737] 2. URL Monitoring Module: Monitors referral links in answers and files to ensure they are not fraudulent or malicious.

[0738] 3. Program code monitoring module: Checks the program code generated by other AI systems to ensure it does not contain malicious software.

[0739] 4. Alert module: Issues an alert to the user if a problem is detected as a result of monitoring.

[0740] 5. Emotion Recognition Engine: Recognizes user emotions and adjusts the content and presentation of warning messages based on those emotions.

[0741] System processing flow and specific examples

[0742] Watch for answers and files:

[0743] Server: Periodically retrieves answers and files generated by other AI systems via API calls and database queries.

[0744] Server: Saves the obtained answers and files as string or binary data.

[0745] Server: The analysis module runs and performs text analysis on the stored data, checking for suspicious content using predefined keywords (e.g., phishing, malware, suspicious).

[0746] Extract and validate the URL:

[0747] Server: Analyzes the text data extracted from the responses and files and extracts all the URLs contained within.

[0748] Server: Sends an HTTP request to each extracted URL to retrieve the web page content, including timeout and error handling.

[0749] Server: Analyzes the HTML content of retrieved web pages and filters them to detect fraudulent or malicious sites by checking for specific keywords and patterns.

[0750] Program code monitoring:

[0751] Server: Retrieves program code generated by other AI systems in the form of strings, via API calls or database queries.

[0752] Server: Analyzes the acquired program code line by line and checks for suspicious code patterns using regular expressions. For example, it performs pattern matching on "import os", "import sys", "exec(", etc.

[0753] Emotion Recognition and Analysis:

[0754] Device: Captures the user's voice, text, and facial expression data using input devices such as a microphone, keyboard, and camera.

[0755] Server: Receives data sent from the device and passes it to the emotion engine.

[0756] Server: The emotion engine analyzes the tone of voice, words used, and facial expressions to assess the user's emotional state.

[0757] Sending a warning:

[0758] Server: The alert module is launched to collect detailed information about the problem detected in the previous step.

[0759] Server: Using the evaluation results of the emotion engine, generate a warning message according to the user's emotional state. If the user is feeling anxious, use a less urgent message such as "This link is not safe. Please rest assured, we will take action immediately."

[0760] Server: Notifies the user of a warning message in a variety of formats, including console log, email, or a GUI popup message.

[0761] Examples of concrete examples and prompts

[0762] For example, suppose a user is recommended a link and needs to verify whether it is a fraudulent page. The server analyzes the link and, if it determines that it is likely a fraudulent page, warns the user that "This link is not safe." If the emotion engine detects anxiety in the user's tone of voice or facial expression, it will provide a softer warning message and, if necessary, add a help link.

[0763] Example prompt sentence:

[0764] "Please explain how your system verifies the possibility that other AI-generated links may be fraudulent sites and issues an appropriate warning to the user if it finds them. Also, please explain how you would recognize the user's emotions and adjust the warning message accordingly."

[0765] In this way, the server, terminal, and user work together to carry out the process, creating a system that improves the user experience while ensuring the security of the information provided by the generative AI model.

[0766] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0767] Step 1: Get answers and files

[0768] Input: Answers generated from other artificial intelligence systems or data from files.

[0769] Server: Periodically retrieves generated answers and files from other AI systems using REST API calls and database queries.

[0770] Server: Saves the retrieved data as string data or binary data.

[0771] Output: Saved answers and file data.

[0772] Step 2: Monitor and analyze responses and files

[0773] Input: Saved answers and file data.

[0774] Server: The analysis module runs and performs text analysis on the stored data, checking for suspicious content using predefined keywords (e.g., phishing, malware, suspicious).

[0775] Server: If suspicious content is detected as a result of text analysis, it will record the content in a log file and notify the warning module.

[0776] Output: Analysis results and log file updates.

[0777] Step 3: Extracting the URL

[0778] Input: Saved answers and file data.

[0779] Server: Analyzes the text data extracted from the responses and files and extracts all the URLs contained within.

[0780] Server: Use text analysis tools and libraries to identify and catalog URL patterns in responses and files.

[0781] Output: A list of extracted URLs.

[0782] Step 4: Verify the URL

[0783] Input: A list of extracted URLs.

[0784] Server: Sends HTTP requests to each URL to retrieve the web page content, including timeout and error handling.

[0785] Server: Analyzes the HTML of the retrieved page content and filters it to detect fraudulent or malicious sites, checking for specific keywords and suspicious patterns.

[0786] Server: Records the results of URL validation in a log file and notifies the alert module if a problem is detected.

[0787] Output: URL validation result and log file update.

[0788] Step 5: Monitoring the program code

[0789] Input: Program code generated by another artificial intelligence system.

[0790] Server: Retrieves the generated program code in string form, either via an API call or a database query.

[0791] Server: Analyzes the acquired program code line by line and checks for suspicious code patterns (e.g., import os, import sys, exec()) using regular expressions.

[0792] Server: If pattern matching detects suspicious code, identify the line and log it with a detailed comment.

[0793] Output: Code analysis results and log file updates.

[0794] Step 6: Emotion Recognition and Analysis

[0795] Input: User voice, text, and facial expression data.

[0796] Device: Captures the user's voice, text, and facial expression data using input devices such as a microphone, keyboard, and camera.

[0797] Server: Receives data sent from the device and passes it to the emotion engine.

[0798] Server: The emotion engine analyzes the tone of voice, words used, and facial expressions to assess the user's emotional state.

[0799] Server: Quantifies or categorizes the emotion evaluation results and notifies the warning module.

[0800] Output: Emotion data reflected in the emotion rating results and warnings.

[0801] Step 7: Sending an alert

[0802] Input: Analysis results and emotion evaluation results.

[0803] Server: The warning module is launched to collect detailed information about the problem detected in the previous step. Using the evaluation results of the emotion engine, a warning message is generated based on the user's emotional state.

[0804] Server: Notifies each user of warning messages in a format suitable for them, such as console logs, emails, or GUI popup messages.

[0805] User: Receives a warning message and takes action. Users can also view the provided help links and support options.

[0806] Output: Warning message provided to user and log file updates.

[0807] In this way, by having the server, terminal, and user work together to carry out processing, we have created a system that improves the user experience while ensuring the security of the information provided by the generative AI model.

[0808] (Application example 2)

[0809] Next, a description will be given of Application Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[0810] The monitoring of answers, files, reference URLs, and program code generated by conventional AI systems has limited ability to detect suspicious content and malicious sites, making it difficult to fully guarantee user security. Furthermore, these systems issue warnings without taking into account the user's emotional state, which can lead to a poor user experience.

[0811] The identification processing by the identification processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means. In this invention, the server includes means for monitoring answers and files generated by other artificial intelligence systems and checking for suspicious content, means for monitoring reference URLs and detecting fraudulent or malicious sites, means for monitoring generated program code and checking whether it contains malicious programs, and means for issuing a warning based on the user's emotional state when malicious content is detected. This makes it possible to improve user security and the user experience.

[0812] "Answers and files generated by other artificial intelligence systems" refers to responses and document data automatically created by external artificial intelligence systems.

[0813] "Means for checking for suspicious content" refers to a method or device that has the functionality to analyze and detect fraudulent or malicious content contained in messages or files.

[0814] "Means for monitoring referring URLs to detect fraudulent or malicious sites" refers to a method or device for monitoring URLs contained in responses or files and determining whether the linked site is a fraudulent or malicious site.

[0815] "Means for monitoring generated program code and checking whether it contains malicious code" refers to a method or device for monitoring program code generated by another artificial intelligence system and detecting malicious code contained therein.

[0816] "Means for issuing a warning based on the user's emotional state when malicious content is detected" refers to a method or device for issuing a warning in an optimal manner based on the results of an emotional analysis of the user in response to detected suspicious content.

[0817] "Means for assessing emotional state" means a method or device for analyzing a user's text or other information to determine whether the user is in a positive, negative, or neutral emotional state.

[0818] "Means for adjusting the display method and content of a warning message" refers to a method or device for appropriately adjusting a warning message according to the user's emotional state and displaying it in the most suitable format and content for the user.

[0819] The system of the present invention includes means for monitoring answers and files generated by other artificial intelligence systems and checking for suspicious content, means for monitoring referred URLs and detecting fraudulent or malicious sites, means for monitoring generated program code and checking whether it contains malicious programs, and means for issuing a warning based on the user's emotional state when malicious content is detected.

[0820] 1. Answers and file monitoring methods:

[0821] The server monitors answers and files generated by external AI systems in real time. This includes analyzing the content using predefined keywords (e.g., phishing, malware, suspicious) to check for suspicious content. For example, if the monitored text contains a keyword such as "phishing," the server will determine that the text is suspicious.

[0822] 2. URL monitoring methods:

[0823] The server extracts the URL contained in the response or file and sends an HTTP request to retrieve the content of the web page pointed to by the URL. The retrieved page content is then analyzed and checked against patterns designed to detect fraudulent or malicious sites. For example, a message containing the URL "http: / / example-phishing.com" is analyzed, and if the URL is determined to be a fraudulent site, an alert is issued.

[0824] 3. Program code monitoring methods:

[0825] The server monitors program code generated by external artificial intelligence systems, checking it using regular expressions to see if it contains malicious program patterns such as "import os," "import sys," or "exec(." If suspicious code is detected, it records its contents and issues a warning to the user.

[0826] 4. Emotion recognition means:

[0827] The server receives the user's voice, text, and facial expression data and performs emotion analysis using an emotion engine. This analyzes the user's tone of voice, words used, and facial expression changes to assess the user's emotional state (e.g., stress, anxiety, satisfaction), and adjusts the display method and content of warning messages descriptively based on the user's emotional state.

[0828] 5. Warning Method:

[0829] The server optimizes the content and presentation of warning messages based on the user's emotional state, as detected by the emotion engine. For example, if the user is feeling anxious, the server will present the warning message in a more understandable and less stressful format, and also add a help link with specific actions to take.

[0830] Hardware and software used

[0831] Server: Server hardware that includes a central processing unit (CPU), memory, and storage.

[0832] Emotion Recognition Engine: Uses TextBlob as a model for emotion analysis.

[0833] URL monitoring: Uses the Requests library to send HTTP requests.

[0834] Program code monitoring: Uses the re library for regular expression analysis.

[0835] Sending email: Uses the smtplib library to send alert messages.

[0836] Specific examples

[0837] For example, when a user receives a message, it may contain a link to a fraudulent website. The server analyzes the URL and, if it determines that the site is likely fraudulent, warns the user, "This link is not safe." At the same time, the emotion engine analyzes the user's text, "I have a bad feeling about this," and, if it detects a negative emotional state, it makes the warning message more gentle and adds a help link if necessary.

[0838] Prompt sentence for generative AI model

[0839] "Analyze text entered by users and assess its emotional state (positive, negative, neutral). Develop an application that monitors provided messages, URLs, and program code for suspicious content and issues alerts accordingly."

[0840] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[0841] Step 1:

[0842] The server receives answers or files generated by other AI systems. These answers or files are input as string data. The server analyzes the content using predefined keywords (e.g., phishing, malware, suspicious) to check for suspicious items. The results of this analysis are provided as output.

[0843] Step 2:

[0844] The server automatically extracts the URLs contained in the response or file. The list of URLs is input. The server sends an HTTP request to each URL to retrieve the web page content. This content is output as text data.

[0845] Step 3:

[0846] The server analyzes the content of the retrieved web page and checks the text against predefined patterns to detect fraudulent or malicious sites. If a fraudulent or malicious site is detected, the URL and reason for the detection are printed.

[0847] Step 4:

[0848] The server receives program code generated by other AI systems. The code is input as string data. The server uses regular expressions to check whether the program contains suspicious code patterns such as "import os," "import sys," or "exec(." The server then outputs the analysis results.

[0849] Step 5:

[0850] The server receives the user's voice, text, and facial expression data. These data are used as input. The emotion engine analyzes the user's tone of voice, the words used, and changes in facial expressions to assess the user's emotional state (e.g., stress, anxiety, satisfaction). This emotion assessment is output.

[0851] Step 6:

[0852] The server adjusts the display method and content of the warning message based on the user's emotional state detected by the emotion engine. The warning message template is used as input, and a warning message adapted to the user's emotional state is generated as the final output.

[0853] Step 7:

[0854] The device receives the optimized warning message sent from the server and displays it to the user, specifically, a warning message containing gentle language and additional support information according to the user's emotional state.

[0855] 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.

[0856] The data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of the data generation model 58 is ChatGPT (Internet Search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search <url: https: gemini.google.com ?hl="ja">) and other generation AIs. The data generation model 58 is obtained by performing deep learning on a neural network. A prompt including an instruction is input to the data generation model 58, and inference data such as voice data indicating voice, text data indicating text, and image data indicating an image is also input. The data generation model 58 performs inference on the input inference data in accordance with the instruction indicated by the prompt, and outputs the inference result in a data format such as voice data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[0857] In the above embodiment, an example in which the specific processing is performed by the data processing device 12 has been given, but the technology of the present disclosure is not limited to this, and the specific processing may be performed by the smart glasses 214.

[0858] [Third embodiment]

[0859] FIG. 5 shows an example of the configuration of a data processing system 310 according to the third embodiment.

[0860] 5, the data processing system 310 includes the data processing device 12 and a headset terminal 314. An example of the data processing device 12 is a server.

[0861] The data processing device 12 includes a computer 22, a database 24, and a communication I / F 26. The computer 22 is an example of a "computer" according to the technology of the present disclosure. The computer 22 includes a processor 28, a RAM 30, and a storage 32. The processor 28, the RAM 30, and the storage 32 are connected to a bus 34. The database 24 and the communication I / F 26 are also connected to the bus 34. The communication I / F 26 is connected to a network 54. Examples of the network 54 include a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0862] 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.

[0863] The microphone 238 receives instructions and the like from the user 20 by receiving voice uttered by the user 20. The microphone 238 captures the voice uttered by the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio in accordance with instructions from the processor 46.

[0864] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an imaging element such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the surroundings of user 20 (for example, an imaging range defined by an angle of view equivalent to the field of vision of a typical healthy person).

[0865] 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.

[0866] 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.

[0867] The specific processing program 56 is an example of a "program" according to the technology of the present disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific processing is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific processing program 56 executed on the RAM 30.

[0868] 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.

[0869] In the headset type terminal 314, a reception output process is performed by the processor 46. A reception output program 60 is stored in the storage 50. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output process is realized by the processor 46 operating as the control unit 46A in accordance with the reception output program 60 executed on the RAM 48.

[0870] Next, a description will be given of the identification process performed by the identification processing unit 290 of the data processing device 12. In the following description, the data processing device 12 will be referred to as the "server" and the headset type terminal 314 will be referred to as the "terminal."

[0871] The present invention is a system that monitors answers and files generated by other AI systems, detects suspicious content, and issues a warning to users. This system operates as software installed on a server, and plays a role in enhancing user security.

[0872] Basic system configuration

[0873] The system consists of the following components:

[0874] 1. Answer and file monitoring module: This module takes answers and files generated by other AI systems and analyzes them for suspicious content.

[0875] 2. URL Monitoring Module: Monitors referenced URLs in answers and files to check whether they are fraudulent or malicious sites.

[0876] 3. Program code monitoring module: Checks the program code generated by other AI systems to monitor whether it contains malicious code.

[0877] 4. Alert module: Issues an alert to the user if a problem is detected as a result of monitoring.

[0878] Program processing

[0879] 1. Monitoring responses and files

[0880] Server: The server receives answers and files generated by other AI systems, treating the received content as strings or data.

[0881] Server: The analysis module checks the content using predefined keywords (e.g., phishing, malware, suspicious) and issues a warning if any suspicious items are included.

[0882] 2. URL Monitoring

[0883] Server: Parses messages and content containing referral URLs and extracts the URLs.

[0884] Server: Sends an HTTP request to the extracted URL and retrieves the page content.

[0885] Server: Analyzes the retrieved page content and checks the text against predefined patterns to detect fraudulent or malicious sites, and issues a warning if a problem is detected.

[0886] 3. Program code monitoring

[0887] Server: Obtains program code generated by other AI systems, and analyzes the instructions in the code as strings.

[0888] Server: Uses regular expressions to check whether a program contains suspicious code patterns such as "import os", "import sys", or "exec(", and issues a warning if they are found.

[0889] 4. Issuing a warning

[0890] Server: A warning module runs and notifies the user of detailed information about detected problems. Warnings are mainly output to the console log, email, or as messages on the GUI.

[0891] Specific examples

[0892] For example, when another AI system recommends a link to a user, the server must verify whether the link is a fraudulent site. If the server analyzes the link and determines that it is likely to be a fraudulent site, it warns the user that "this link is not safe." Also, if the program code provided to the user by the other AI system contains fraudulent operations, the server detects this and issues a warning that "this code is dangerous."

[0893] This will allow users to minimize the risk from malicious AI technology and use the system with peace of mind, and this system will play an important role in improving user security.

[0894] The processing flow will be explained below.

[0895] Step 1:

[0896] The server receives answers and files generated by other AI systems, and the data is stored on the server as a string or file.

[0897] Step 2:

[0898] The server analyzes the responses and files it receives and checks their content against predefined keywords (e.g., phishing, malware, suspicious). The analysis module is responsible for this and checks whether any suspicious words are included.

[0899] Step 3:

[0900] The server extracts the reference URLs contained in the response or file, using regular expressions to extract the URLs from the text, and creates a list.

[0901] Step 4:

[0902] The server sends an HTTP request to the extracted URL to retrieve the content of the target web page. The request module issues an HTTP request and receives the response from the server.

[0903] Step 5:

[0904] The server performs text analysis on the retrieved web page content, checking for patterns (e.g., phishing terms, suspicious links) to detect fraudulent or malicious sites. The content is analyzed as text to see if it contains any suspicious patterns.

[0905] Step 6:

[0906] The server acquires program code generated by other AI systems and analyzes the code content. At this time, the code is acquired as a string and is the subject of inspection.

[0907] Step 7:

[0908] The server uses regular expressions to check whether the acquired program code contains suspicious code patterns such as "import os", "import sys", or "exec(". The analysis module checks each line of the code and records any suspicious patterns it finds.

[0909] Step 8:

[0910] The server combines the analysis results from each module and, if any suspicious content is detected, generates a warning message that contains detailed information about the detected problem and is presented in a user-friendly format.

[0911] Step 9:

[0912] The server notifies the user of any generated warning messages via console logs, email, or messages on the user interface, allowing the user to quickly identify suspicious areas.

[0913] This allows the server to continuously monitor the outputs of other artificial intelligence systems, creating a system that ensures user security.

[0914] Example 1

[0915] Next, a description will be given of Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal."

[0916] Answers and files generated by conventional AI systems sometimes contain malicious content, fraudulent websites, or malicious code, posing a threat to user security. Furthermore, there is a lack of systems that can efficiently detect such malicious content and quickly warn users. Therefore, there is a need for a reliable monitoring system that allows users of AI systems to use them with peace of mind.

[0917] The specific processing by the specific processing unit 290 of the data processing device 12 in the first embodiment is realized by the following means.

[0918] In this invention, the server includes means for monitoring answers and files generated by other artificial intelligence systems and analyzing suspicious content using predefined keywords, means for monitoring referred URLs and analyzing page content obtained by sending HTTP requests to detect fraudulent or malicious sites, means for monitoring generated program code and checking specific code patterns with regular expressions to detect whether malicious programs are included, and means for notifying users of detailed information about detected problems via console logs, email, or GUI messages. This makes it possible to effectively detect security risks contained in content generated by artificial intelligence systems and promptly notify users.

[0919] "Answers and files" refers to text data and document data generated by other artificial intelligence systems.

[0920] "Monitoring" means the act of continuously or periodically checking or auditing a particular object.

[0921] "Keywords" refers to predefined important phrases or words for a specific purpose.

[0922] "Referring URL" refers to a string of characters that indicates a link to a web page or resource.

[0923] "HTTP request" refers to a request sent by a web browser or client application to retrieve information from a web server.

[0924] "Page Content" refers to the text, images, scripts, and other information displayed on a web page.

[0925] "Fraudulent Site" refers to a fraudulent website designed to deceive users.

[0926] "Malicious Site" means a website whose purpose is to steal user information, cause damage to systems, or perform other fraudulent activities.

[0927] "Program code" refers to the source code for writing a computer program.

[0928] A "code pattern" refers to a set of notations or instructions that have a specific purpose or function.

[0929] "Regular expression" refers to a formal language for describing specific patterns.

[0930] "Console log" refers to character-based output that displays system or application status and error messages.

[0931] "Mail" means electronic mail, a means of sending and receiving messages in digital form.

[0932] "GUI Message" means a warning or informational message displayed on a graphical user interface.

[0933] The present invention is a system that monitors answers and files generated by other AI systems, detects suspicious content, and issues a warning to the user. Specific embodiments for implementing this system are described below.

[0934] Basic system configuration

[0935] The system runs as software installed on a server and consists of the following components:

[0936] 1. Answer and file monitoring module

[0937] 2. URL Monitoring Module

[0938] 3. Program Code Monitoring Module

[0939] 4. Warning Module

[0940] Monitor responses and files

[0941] server

[0942] The server retrieves answers and files generated by other AI systems, using API requests and file system monitoring.

[0943] The retrieved answers and file contents are analyzed as text data, and the server searches the text using predefined keywords (e.g., phishing, malware, suspicious) to extract suspicious content.

[0944] URL monitoring

[0945] server

[0946] The server extracts URLs from the responses and files and generates a list of URLs using regular expressions.

[0947] Send an HTTP request to the extracted URL and get the response.

[0948] The text content of the response is analyzed and matched against patterns (e.g., characteristics of phishing pages) to detect fraudulent or malicious websites.

[0949] Program code monitoring

[0950] server

[0951] The server retrieves program code generated by other artificial intelligence systems from specific folders or APIs.

[0952] The commands in the acquired program code are analyzed as strings and checked for specific patterns (e.g., "import os", "import sys", "exec()") using regular expressions.

[0953] Sending a warning

[0954] server

[0955] The server generates detailed warning messages about any problems it detects, including the type, location, and detailed description of the problem detected.

[0956] The generated warning messages are not only output to the console log, but also notified to the user by email and displayed as a popup message on the GUI.

[0957] Specific examples

[0958] A specific example of operation is shown below.

[0959] Example 1: If the server retrieves a new answer and it contains the keyword "phishing", the server will notify the user "Warning: This answer contains suspicious content."

[0960] Example 2: If the server sends an HTTP request to a URL extracted from a file and the result has characteristics of a phishing site, the server notifies the user with "Warning: This URL is not safe."

[0961] Example 3: When the server parses a new code file and it contains the instruction "exec(", the server will notify the user "Warning: this code is dangerous".

[0962] Prompt Sentence Examples

[0963] Check if this link is safe

[0964] "Check this code for illegal instructions"

[0965] By inputting such prompt sentences into the generative AI model, the server will initiate a specific analysis process, allowing users to obtain reliable information and minimizing security risks.

[0966] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0967] Step 1:

[0968] Get answers and files

[0969] Server: Retrieves answers and files generated by other AI systems, using API requests and file system monitoring.

[0970] Input: Answers or files generated by other AI systems.

[0971] Output: Text data of the obtained answers and files.

[0972] Step 2:

[0973] Check the contents

[0974] Server: Analyzes the obtained answers and file contents as text data. Analyzes suspicious content using predefined keywords (e.g., phishing, malware, suspicious).

[0975] Input: Text data from the answers or files obtained.

[0976] Output: Result on whether or not the content is suspicious.

[0977] What happens: The server searches for text against a list of keywords and flags it as suspicious if it contains a matching keyword.

[0978] Step 3:

[0979] URL extraction

[0980] Server: Extract URLs from responses and files using regular expressions.

[0981] Input: The text data from the parsed answer or file.

[0982] Output: The extracted URL list.

[0983] What it does: The server searches for strings that start with "http: / / " or "https: / / " and lists all URLs.

[0984] Step 4:

[0985] URL validation

[0986] Server: Sends an HTTP request to the extracted URL and obtains the response data.

[0987] Input: The extracted URL list.

[0988] Output: Response data for the URL.

[0989] Specific operation: The server sends an HTTP request for each URL and saves the page content of the response it receives.

[0990] Step 5:

[0991] Analyzing page content

[0992] Server: Analyzes the response data and matches it with patterns (e.g., characteristics of phishing sites) to detect fraudulent or malicious sites.

[0993] Input: Response data for the URL.

[0994] Output: Detected fraudulent and malicious sites.

[0995] What it does: The server performs pattern matching on the text content of the response and flags any fraudulent or malicious features detected.

[0996] Step 6:

[0997] Obtaining the program code

[0998] Server: Retrieves program code generated by other AI systems from specific folders or APIs.

[0999] Input: Program code generated by another artificial intelligence system.

[1000] Output: Text data of the retrieved program code.

[1001] What it does: The server monitors newly added code files and incorporates them one by one.

[1002] Step 7:

[1003] Code content analysis

[1004] Server: Analyzes the program code and checks for specific code patterns (e.g., "import os", "import sys", "exec()") using regular expressions.

[1005] Input: Text data of the retrieved program code.

[1006] Output: Suspicious code pattern detection results.

[1007] What it does: The server parses instructions in your code using regular expressions and flags when certain patterns are detected.

[1008] Step 8:

[1009] Generate a warning message

[1010] Server: Generates detailed warning messages about detected problems.

[1011] Input: Suspicious content, URL, or code pattern detection results.

[1012] Output: A detailed warning message.

[1013] What happens: The server generates a message containing the type, location, and detailed description of the problem detected.

[1014] Step 9:

[1015] User Notification

[1016] Server: Generated warning messages are output to the console log and notified to the user via email or GUI message.

[1017] Input: A detailed warning message.

[1018] Output: The warning message that was sent to the user.

[1019] Specific behavior: The server not only logs the warning to the console, but also sends an email to the user and displays a warning as a popup message in the GUI.

[1020] (Application example 1)

[1021] Next, a description will be given of Application Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal."

[1022] With the development of modern information and communication technology, security risks are increasing for answers, files, and program code generated by other artificial intelligence systems. Furthermore, the potential for suspicious URLs and malicious programs contained in this content to cause harm to users is also increasing. In particular, with the widespread use of smart devices, real-time monitoring and alerts are becoming increasingly important. However, current systems lack the means to effectively address these multifaceted threats.

[1023] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 1 is realized by the following means.

[1024] In this invention, the server includes means for monitoring answers and files generated by other artificial intelligence systems and checking for suspicious content, means for monitoring referred URLs and detecting fraudulent or malicious sites, means for monitoring generated program code and checking whether it contains malicious programs, means for issuing a warning when malicious content is detected, and means for operating on the smart device and monitoring the user's digital content in real time, thereby enabling real-time monitoring and warning of the digital content used by the user.

[1025] "Other artificial intelligence systems" refers to artificial intelligence technologies operated or utilized by users to analyze and generate data.

[1026] "Answer or file" refers to text, documents, images, audio, or other forms of data generated by other artificial intelligence systems.

[1027] "Suspicious Content" refers to elements within a file or message that may indicate phishing, malware, spam, or other malicious content.

[1028] "Referring URL" refers to a website address or link contained in a response or file.

[1029] A "fraudulent site" refers to a malicious website that attempts to fraudulently obtain users' personal information or money.

[1030] "Malicious Sites" refers to websites that install malware on users' devices or perform other malicious activities.

[1031] "Program Code" refers to the source code of software generated by other artificial intelligence systems.

[1032] "Malicious Program" refers to source code or a program designed to perform unauthorized operations or undesirable behavior.

[1033] "Warning" refers to a notification or alert issued to a user when the system detects suspicious or malicious content.

[1034] "Smart devices" refers to electronic devices such as mobile phones, tablets, and smart glasses that have internet connectivity.

[1035] "Digital content" refers to electronically stored information or data, such as web pages, messages, files, and program code.

[1036] "Real-time" refers to data being processed immediately at the moment it is generated or acquired.

[1037] This invention provides a security system called "AI Guard," which monitors answers and files generated by other artificial intelligence systems and analyzes whether they contain suspicious content. It also monitors referred URLs to determine whether they are fraudulent or malicious sites, and checks generated program code to monitor whether they contain malicious programs. Based on these monitoring results, it issues a warning to the user when malicious content is detected. Furthermore, this system runs on smart devices and can monitor users' digital content in real time.

[1038] System Configuration

[1039] The system consists of the following elements:

[1040] 1. Answer and file monitoring module:

[1041] The server receives answers and files generated by other AI systems, whose contents are processed as strings or data.

[1042] The server analyzes the content using predefined keywords (e.g., phishing, malware, suspicious) and issues a warning if it contains any suspicious items.

[1043] 2. URL Monitoring Module:

[1044] Parse messages and content that contain URLs and extract them.

[1045] An HTTP request is sent to the extracted URL to retrieve the page content.

[1046] It analyzes the retrieved page content, checks the text against defined patterns to detect fraudulent or malicious sites, and issues a warning if a problem is detected.

[1047] 3. Program code monitoring module:

[1048] It acquires program code generated by other AI systems and analyzes the instructions in the code as strings of characters.

[1049] It uses regular expressions to check whether a program contains suspicious code patterns such as "import os", "import sys", or "exec(", and issues a warning if they are found.

[1050] 4. Alert sending module:

[1051] The server includes a module to notify the user with detailed information about detected problems. Warnings are typically output to the console log, email, or as messages in the graphical user interface (GUI).

[1052] Hardware and software used

[1053] Hardware: Smartphones (e.g., iPhones and Android devices), smart glasses (e.g., Google Glass).

[1054] Software: Python programming language, re module (regular expression processing), requests module (HTTP request processing).

[1055] Example of a system

[1056] The "AI Guard" app creates warning messages for users by inputting prompt sentences such as the following into a generative AI model:

[1057] Prompt Sentence Examples

[1058] "The site you are trying to visit, URL: {url}, may be fraudulent or malicious. We do not recommend visiting this site. More information: {details}"

[1059] Specific examples

[1060] For example, if a user receives a link via SMS to "http: / / phishing-site.com," the following steps will occur:

[1061] The server parses the URL and checks it against a blacklist.

[1062] Get the page content and detect the keyword "phishing".

[1063] It warns users that "this link is not safe."

[1064] In this way, the "AI Guard" system of the present invention plays a role in keeping users' digital lives safe by monitoring their digital content in real time and early detecting and warning of potential security risks.

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

[1066] Step 1:

[1067] The server receives the answers and files generated by other AI systems. The input is the answers and files generated by other AI systems, which are provided in the form of strings and data. The server receives this and analyzes it in the next step.

[1068] Step 2:

[1069] The server analyzes the retrieved answers and file contents using keyword filtering technology to detect suspicious keywords (e.g., phishing, malware, suspicious). The input is text data, and the output is the analysis result, indicating whether or not suspicious keywords are present. Through this process, the server makes an initial detection of suspicious content.

[1070] Step 3:

[1071] The server extracts the referrer URLs contained in the parsed answers and files. The input is the text data of the answers and files, and the output is the extracted URLs. This is then passed to the URL monitoring module.

[1072] Step 4:

[1073] The server sends an HTTP request to the URL extracted using the URL monitoring module to obtain the web page content. The input is the extracted URL, and the output is the HTML content of the web page corresponding to the URL. In this process, the server communicates over the network to obtain the page content of the target URL.

[1074] Step 5:

[1075] The server analyzes the retrieved HTML content and checks the text against predefined patterns (e.g., phishing keywords, malware download links) to detect fraudulent or malicious sites. The input is the HTML data of the web page, and the output is a judgment result on whether the page is suspicious or not. This allows the server to evaluate the safety of the URL.

[1076] Step 6:

[1077] The server retrieves program code generated by other AI systems and analyzes its contents as strings. The input is the program code, and the output is the analysis result, indicating whether or not there are suspicious code patterns. In this process, the server scans the strings of the program code with regular expressions to detect malicious code.

[1078] Step 7:

[1079] The server issues a warning to the user based on the detection results. The input is the analysis results and judgment results, and the output is a warning message. The warning is notified to the user via email, log, or GUI message. This process allows the user to receive warnings about security risks in real time.

[1080] Through the above processing steps, the server can monitor the user's digital content in real time and issue early warnings of potential security risks.

[1081] Furthermore, an emotion engine that estimates the user's emotion may be further combined. That is, the identification processing unit 290 may estimate the user's emotion using the emotion identification model 59, and perform identification processing using the user's emotion.

[1082] This invention integrates an emotion engine into a system that monitors answers and files generated by other AI systems, detects suspicious content, and issues warnings to users. This system operates as software installed on a server, and aims to improve user security and experience by adjusting the content and presentation of warnings according to the user's emotional state.

[1083] Basic system configuration

[1084] The system consists of the following components:

[1085] 1. Answer and file monitoring module: This module takes answers and files generated by other AI systems and analyzes them for suspicious content.

[1086] 2. URL Monitoring Module: Monitors referenced URLs in answers and files to check whether they are fraudulent or malicious sites.

[1087] 3. Program code monitoring module: Checks the program code generated by other AI systems to monitor whether it contains malicious code.

[1088] 4. Alert module: Issues an alert to the user if a problem is detected as a result of monitoring.

[1089] 5. Emotion Engine: Recognizes user emotions and adjusts the content and presentation of warning messages based on those emotions.

[1090] Program processing

[1091] 1. Monitoring responses and files

[1092] Server: The server receives answers and files generated by other AI systems, treating the received content as strings or data.

[1093] Server: The analysis module checks the content using predefined keywords (e.g., phishing, malware, suspicious) to see if it contains any suspicious items.

[1094] 2. URL Monitoring

[1095] Server: Parses messages and content containing referral URLs and extracts the URLs.

[1096] Server: Sends an HTTP request to the extracted URL and retrieves the content of the target web page.

[1097] Server: Analyzes the retrieved page content and checks the text against predefined patterns to detect fraudulent or malicious sites, and issues a warning if a problem is detected.

[1098] 3. Program code monitoring

[1099] Server: Obtains program code generated by other AI systems. At this time, the code is obtained as a string and is subject to inspection.

[1100] Server: Regular expressions are used to check whether the program contains suspicious code patterns such as "import os", "import sys", or "exec(". The analysis module checks each line of the code and records any suspicious patterns found.

[1101] 4. Emotion recognition

[1102] Server: The emotion engine receives the user's voice, text, and facial expression data and performs emotion analysis. The emotion analysis module analyzes the user's tone of voice, words used, and facial expression changes to assess the user's emotional state (e.g., stress, anxiety, satisfaction).

[1103] Server: The emotion engine adjusts the display and content of warning messages based on the detected user emotional state. For example, if the user is feeling anxious, the warning message will be presented in a more understandable and less urgent manner.

[1104] 5. Issuing a warning

[1105] Server: A warning module runs and notifies the user of detailed information about detected problems. Warnings are mainly output to the console log, email, or as messages on the GUI.

[1106] Server: The emotion engine optimizes the priority and detail of the warning message based on the user's current emotional state, providing the user with an appropriate warning.

[1107] Specific examples

[1108] For example, when another AI system recommends a link to a user, it must verify whether the link is a fraudulent site. The server analyzes the link and, if it determines that it is likely to be a fraudulent site, warns the user, "This link is not safe." If the emotion engine detects anxiety from the user's tone of voice or facial expression, it provides a softer warning message and adds a help link if necessary. Furthermore, if the program code provided to the user by the other AI system contains fraudulent operations, the server detects this and issues a warning, saying, "This code is dangerous." If the emotion engine detects the user's state of stress, it simplifies the warning message and presents additional support options.

[1109] This allows the server to continuously monitor the outputs of other artificial intelligence systems and provide appropriate warnings based on the user's emotional state, improving both security and user experience.

[1110] The processing flow will be explained below.

[1111] Step 1:

[1112] The server receives answers and files generated by other AI systems, and the data is stored on the server as a string or file.

[1113] Step 2:

[1114] The server analyzes the responses and files it receives and checks their content against predefined keywords (e.g., phishing, malware, suspicious). The analysis module is responsible for this and checks whether any suspicious words are included.

[1115] Step 3:

[1116] The server extracts the reference URLs contained in the response or file, using regular expressions to extract the URLs from the text, and creates a list.

[1117] Step 4:

[1118] The server sends an HTTP request to the extracted URL to retrieve the content of the target web page. The request module issues an HTTP request, and the server receives the response.

[1119] Step 5:

[1120] The server performs text analysis on the retrieved web page content, checking for patterns (e.g., phishing terms, suspicious links) to detect fraudulent or malicious sites. The content is analyzed as text to see if it contains any suspicious patterns.

[1121] Step 6:

[1122] The server acquires program code generated by other AI systems and analyzes the code content. At this time, the code is acquired as a string and is the subject of inspection.

[1123] Step 7:

[1124] The server uses regular expressions to check whether the acquired program code contains suspicious code patterns such as "import os", "import sys", or "exec(". The analysis module checks each line of the code and records any suspicious patterns it finds.

[1125] Step 8:

[1126] The server uses an emotion engine to recognize the user's emotions. It analyzes the voice, text, and facial expression data obtained from the user to evaluate their emotional state (e.g., stress, anxiety, satisfaction). This analysis includes voice tone analysis, emotional keyword analysis in the text, and facial expression analysis.

[1127] Step 9:

[1128] The server adjusts the content and presentation of the warning message based on the emotional state recognized by the emotion engine. For example, if the user expresses anxiety, the server changes the warning message to be more gentle and detailed.

[1129] Step 10:

[1130] The server combines the analysis results from each module and, if any suspicious content is detected, generates a warning message that contains detailed information about the detected problem and is presented in an easy-to-understand format to the user.

[1131] Step 11:

[1132] The server notifies the user of the generated warning message. Notification methods include console log, email, or a message on the user interface. The priority and detail of the message are adjusted according to the analysis results of the emotion engine. This allows the user to quickly understand which parts are suspicious.

[1133] This allows the server to continuously monitor the outputs of other artificial intelligence systems, ensuring user security and providing warnings that take into account the user's emotional state.

[1134] Example 2

[1135] Next, a description will be given of Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal."

[1136] Conventional monitoring systems primarily focus on monitoring answers, files, reference links, and programs generated by other AI systems. However, because they do not take the user's emotional state into consideration, the content and display method of warning messages are uniform, which can lead to poor user experience. In particular, appropriate responses are required for users experiencing stress or anxiety, but these responses are insufficient.

[1137] The identification process by the identification processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means. In this invention, the server includes a means for monitoring answers and files generated by other artificial intelligence systems and checking for suspicious content, a means for monitoring reference links and detecting fraudulent or malicious pages, a means for monitoring generated programs and checking whether they contain malicious software, a means for issuing a warning when malicious content is detected, and a means for recognizing the user's emotional state and adjusting the display method and content of the warning message. This makes it possible to improve security and the user experience by providing appropriate warnings according to the user's emotional state.

[1138] "Suspicious content" is content in network responses or files that indicates security threats such as phishing or malware.

[1139] A "reference link" is a URL or web address contained in a response or file.

[1140] A "scam page" is a malicious website designed to deceive users and obtain their personal information.

[1141] A "malicious page" is a website that contains computer viruses or malware and is designed to harm a user's system.

[1142] "Generated Program" refers to program code created by another artificial intelligence system.

[1143] "Malicious software" refers to programs designed to cause damage to systems or data, including viruses, malware, and spyware.

[1144] A "warning message" is a message intended to notify the user of problems or potential risks detected by the system.

[1145] "Emotional state" refers to a user's psychological state, such as stress, anxiety, or satisfaction, and is recognized through voice, text, and facial expression data.

[1146] An "emotion recognition engine" is a software module for analyzing and evaluating a user's emotional state.

[1147] This invention integrates an emotion engine into a system that monitors answers and files generated by other AI systems, detects suspicious content, and issues warnings to users. This system operates as software installed on a server, and aims to improve user security and experience by adjusting the content and presentation of warnings according to the user's emotional state.

[1148] Basic system configuration

[1149] The system consists of the following components:

[1150] 1. Answer and file monitoring module: This module takes answers and files generated by other AI systems and analyzes them for suspicious content.

[1151] 2. URL Monitoring Module: Monitors referral links in answers and files to ensure they are not fraudulent or malicious.

[1152] 3. Program code monitoring module: Checks the program code generated by other AI systems to ensure it does not contain malicious software.

[1153] 4. Alert module: Issues an alert to the user if a problem is detected as a result of monitoring.

[1154] 5. Emotion Recognition Engine: Recognizes user emotions and adjusts the content and presentation of warning messages based on those emotions.

[1155] System processing flow and specific examples

[1156] Watch for answers and files:

[1157] Server: Periodically retrieves answers and files generated by other AI systems via API calls and database queries.

[1158] Server: Saves the obtained answers and files as string or binary data.

[1159] Server: The analysis module runs and performs text analysis on the stored data, checking for suspicious content using predefined keywords (e.g., phishing, malware, suspicious).

[1160] Extract and validate the URL:

[1161] Server: Analyzes the text data extracted from the responses and files and extracts all the URLs contained within.

[1162] Server: Sends an HTTP request to each extracted URL to retrieve the web page content, including timeout and error handling.

[1163] Server: Analyzes the HTML content of retrieved web pages and filters them to detect fraudulent or malicious sites by checking for specific keywords and patterns.

[1164] Program code monitoring:

[1165] Server: Retrieves program code generated by other AI systems in the form of strings, via API calls or database queries.

[1166] Server: Analyzes the acquired program code line by line and checks for suspicious code patterns using regular expressions. For example, it performs pattern matching on "import os", "import sys", "exec(", etc.

[1167] Emotion Recognition and Analysis:

[1168] Device: Captures the user's voice, text, and facial expression data using input devices such as a microphone, keyboard, and camera.

[1169] Server: Receives data sent from the device and passes it to the emotion engine.

[1170] Server: The emotion engine analyzes the tone of voice, words used, and facial expressions to assess the user's emotional state.

[1171] Sending a warning:

[1172] Server: The alert module is launched to collect detailed information about the problem detected in the previous step.

[1173] Server: Using the evaluation results of the emotion engine, generate a warning message according to the user's emotional state. If the user is feeling anxious, use a less urgent message such as "This link is not safe. Please rest assured, we will take action immediately."

[1174] Server: Notifies the user of a warning message in a variety of formats, including console log, email, or a GUI popup message.

[1175] Examples of concrete examples and prompts

[1176] For example, suppose a user is recommended a link and needs to verify whether it is a fraudulent page. The server analyzes the link and, if it determines that it is likely a fraudulent page, warns the user that "This link is not safe." If the emotion engine detects anxiety in the user's tone of voice or facial expression, it will provide a softer warning message and, if necessary, add a help link.

[1177] Example prompt sentence:

[1178] "Please explain how your system verifies the possibility that other AI-generated links may be fraudulent sites and issues an appropriate warning to the user if it finds them. Also, please explain how you would recognize the user's emotions and adjust the warning message accordingly."

[1179] In this way, the server, terminal, and user work together to carry out the process, creating a system that improves the user experience while ensuring the security of the information provided by the generative AI model.

[1180] The flow of the identification process in the second embodiment will be described with reference to FIG.

[1181] Step 1: Get answers and files

[1182] Input: Answers generated from other artificial intelligence systems or data from files.

[1183] Server: Periodically retrieves generated answers and files from other AI systems using REST API calls and database queries.

[1184] Server: Saves the retrieved data as string data or binary data.

[1185] Output: Saved answers and file data.

[1186] Step 2: Monitor and analyze responses and files

[1187] Input: Saved answers and file data.

[1188] Server: The analysis module runs and performs text analysis on the stored data, checking for suspicious content using predefined keywords (e.g., phishing, malware, suspicious).

[1189] Server: If suspicious content is detected as a result of text analysis, it will record the content in a log file and notify the warning module.

[1190] Output: Analysis results and log file updates.

[1191] Step 3: Extracting the URL

[1192] Input: Saved answers and file data.

[1193] Server: Analyzes the text data extracted from the responses and files and extracts all the URLs contained within.

[1194] Server: Use text analysis tools and libraries to identify and catalog URL patterns in responses and files.

[1195] Output: A list of extracted URLs.

[1196] Step 4: Verify the URL

[1197] Input: A list of extracted URLs.

[1198] Server: Sends HTTP requests to each URL to retrieve the web page content, including timeout and error handling.

[1199] Server: Analyzes the HTML of the retrieved page content and filters it to detect fraudulent or malicious sites, checking for specific keywords and suspicious patterns.

[1200] Server: Records the results of URL validation in a log file and notifies the alert module if a problem is detected.

[1201] Output: URL validation result and log file update.

[1202] Step 5: Monitoring the program code

[1203] Input: Program code generated by another artificial intelligence system.

[1204] Server: Retrieves the generated program code in string form, either via an API call or a database query.

[1205] Server: Analyzes the acquired program code line by line and checks for suspicious code patterns (e.g., import os, import sys, exec()) using regular expressions.

[1206] Server: If pattern matching detects suspicious code, identify the line and log it with a detailed comment.

[1207] Output: Code analysis results and log file updates.

[1208] Step 6: Emotion Recognition and Analysis

[1209] Input: User voice, text, and facial expression data.

[1210] Device: Captures the user's voice, text, and facial expression data using input devices such as a microphone, keyboard, and camera.

[1211] Server: Receives data sent from the device and passes it to the emotion engine.

[1212] Server: The emotion engine analyzes the tone of voice, words used, and facial expressions to assess the user's emotional state.

[1213] Server: Quantifies or categorizes the emotion evaluation results and notifies the warning module.

[1214] Output: Emotion data reflected in the emotion rating results and warnings.

[1215] Step 7: Sending an alert

[1216] Input: Analysis results and emotion evaluation results.

[1217] Server: The warning module is launched to collect detailed information about the problem detected in the previous step. Using the evaluation results of the emotion engine, a warning message is generated based on the user's emotional state.

[1218] Server: Notifies each user of warning messages in a format suitable for them, such as console logs, emails, or GUI popup messages.

[1219] User: Receives a warning message and takes action. Users can also view the provided help links and support options.

[1220] Output: Warning message provided to user and log file updates.

[1221] In this way, by having the server, terminal, and user work together to carry out processing, we have created a system that improves the user experience while ensuring the security of the information provided by the generative AI model.

[1222] (Application example 2)

[1223] Next, a description will be given of Application Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal."

[1224] The monitoring of answers, files, reference URLs, and program code generated by conventional AI systems has limited ability to detect suspicious content and malicious sites, making it difficult to fully guarantee user security. Furthermore, these systems issue warnings without taking into account the user's emotional state, which can lead to a poor user experience.

[1225] The identification processing by the identification processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means. In this invention, the server includes means for monitoring answers and files generated by other artificial intelligence systems and checking for suspicious content, means for monitoring reference URLs and detecting fraudulent or malicious sites, means for monitoring generated program code and checking whether it contains malicious programs, and means for issuing a warning based on the user's emotional state when malicious content is detected. This makes it possible to improve user security and the user experience.

[1226] "Answers and files generated by other artificial intelligence systems" refers to responses and document data automatically created by external artificial intelligence systems.

[1227] "Means for checking for suspicious content" refers to a method or device that has the functionality to analyze and detect fraudulent or malicious content contained in messages or files.

[1228] "Means for monitoring referring URLs to detect fraudulent or malicious sites" refers to a method or device for monitoring URLs contained in responses or files and determining whether the linked site is a fraudulent or malicious site.

[1229] "Means for monitoring generated program code and checking whether it contains malicious code" refers to a method or device for monitoring program code generated by another artificial intelligence system and detecting malicious code contained therein.

[1230] "Means for issuing a warning based on the user's emotional state when malicious content is detected" refers to a method or device for issuing a warning in an optimal manner based on the results of an emotional analysis of the user in response to detected suspicious content.

[1231] "Means for assessing emotional state" means a method or device for analyzing a user's text or other information to determine whether the user is in a positive, negative, or neutral emotional state.

[1232] "Means for adjusting the display method and content of a warning message" refers to a method or device for appropriately adjusting a warning message according to the user's emotional state and displaying it in the most suitable format and content for the user.

[1233] The system of the present invention includes means for monitoring answers and files generated by other artificial intelligence systems and checking for suspicious content, means for monitoring referred URLs and detecting fraudulent or malicious sites, means for monitoring generated program code and checking whether it contains malicious programs, and means for issuing a warning based on the user's emotional state when malicious content is detected.

[1234] 1. Answers and file monitoring methods:

[1235] The server monitors answers and files generated by external AI systems in real time. This includes analyzing the content using predefined keywords (e.g., phishing, malware, suspicious) to check for suspicious content. For example, if the monitored text contains a keyword such as "phishing," the server will determine that the text is suspicious.

[1236] 2. URL monitoring methods:

[1237] The server extracts the URL contained in the response or file and sends an HTTP request to retrieve the content of the web page pointed to by the URL. The retrieved page content is then analyzed and checked against patterns designed to detect fraudulent or malicious sites. For example, a message containing the URL "http: / / example-phishing.com" is analyzed, and if the URL is determined to be a fraudulent site, an alert is issued.

[1238] 3. Program code monitoring methods:

[1239] The server monitors program code generated by external artificial intelligence systems, checking it using regular expressions to see if it contains malicious program patterns such as "import os," "import sys," or "exec(." If suspicious code is detected, it records its contents and issues a warning to the user.

[1240] 4. Emotion recognition means:

[1241] The server receives the user's voice, text, and facial expression data and performs emotion analysis using an emotion engine. This analyzes the user's tone of voice, words used, and facial expression changes to assess the user's emotional state (e.g., stress, anxiety, satisfaction), and adjusts the display method and content of warning messages descriptively based on the user's emotional state.

[1242] 5. Warning Method:

[1243] The server optimizes the content and presentation of warning messages based on the user's emotional state, as detected by the emotion engine. For example, if the user is feeling anxious, the server will present the warning message in a more understandable and less stressful format, and also add a help link with specific actions to take.

[1244] Hardware and software used

[1245] Server: Server hardware that includes a central processing unit (CPU), memory, and storage.

[1246] Emotion Recognition Engine: Uses TextBlob as a model for emotion analysis.

[1247] URL monitoring: Uses the Requests library to send HTTP requests.

[1248] Program code monitoring: Uses the re library for regular expression analysis.

[1249] Sending email: Uses the smtplib library to send alert messages.

[1250] Specific examples

[1251] For example, when a user receives a message, it may contain a link to a fraudulent website. The server analyzes the URL and, if it determines that the site is likely fraudulent, warns the user, "This link is not safe." At the same time, the emotion engine analyzes the user's text, "I have a bad feeling about this," and, if it detects a negative emotional state, it makes the warning message more gentle and adds a help link if necessary.

[1252] Prompt sentence for generative AI model

[1253] "Analyze text entered by users and assess its emotional state (positive, negative, neutral). Develop an application that monitors provided messages, URLs, and program code for suspicious content and issues alerts accordingly."

[1254] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[1255] Step 1:

[1256] The server receives answers or files generated by other AI systems. These answers or files are input as string data. The server analyzes the content using predefined keywords (e.g., phishing, malware, suspicious) to check for suspicious items. The results of this analysis are provided as output.

[1257] Step 2:

[1258] The server automatically extracts the URLs contained in the response or file. The list of URLs is input. The server sends an HTTP request to each URL to retrieve the web page content. This content is output as text data.

[1259] Step 3:

[1260] The server analyzes the content of the retrieved web page and checks the text against predefined patterns to detect fraudulent or malicious sites. If a fraudulent or malicious site is detected, the URL and reason for the detection are printed.

[1261] Step 4:

[1262] The server receives program code generated by other AI systems. The code is input as string data. The server uses regular expressions to check whether the program contains suspicious code patterns such as "import os," "import sys," or "exec(." The server then outputs the analysis results.

[1263] Step 5:

[1264] The server receives the user's voice, text, and facial expression data. These data are used as input. The emotion engine analyzes the user's tone of voice, the words used, and changes in facial expressions to assess the user's emotional state (e.g., stress, anxiety, satisfaction). This emotion assessment is output.

[1265] Step 6:

[1266] The server adjusts the display method and content of the warning message based on the user's emotional state detected by the emotion engine. The warning message template is used as input, and a warning message adapted to the user's emotional state is generated as the final output.

[1267] Step 7:

[1268] The device receives the optimized warning message sent from the server and displays it to the user, specifically, a warning message containing gentle language and additional support information according to the user's emotional state.

[1269] 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.

[1270] The data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of the data generation model 58 is ChatGPT (Internet Search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search <url: https: gemini.google.com ?hl="ja">) and other generation AIs. The data generation model 58 is obtained by performing deep learning on a neural network. A prompt including an instruction is input to the data generation model 58, and inference data such as voice data indicating voice, text data indicating text, and image data indicating an image is also input. The data generation model 58 performs inference on the input inference data in accordance with the instruction indicated by the prompt, and outputs the inference result in a data format such as voice data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[1271] In the above embodiment, an example was given in which the specific processing is performed by the data processing device 12, but the technology of the present disclosure is not limited to this, and the specific processing may be performed by the headset type terminal 314.

[1272] [Fourth embodiment]

[1273] FIG. 7 shows an example of the configuration of a data processing system 410 according to the fourth embodiment.

[1274] 7, a data processing system 410 includes a data processing device 12 and a robot 414. An example of the data processing device 12 is a server.

[1275] The data processing device 12 includes a computer 22, a database 24, and a communication I / F 26. The computer 22 is an example of a "computer" according to the technology of the present disclosure. The computer 22 includes a processor 28, a RAM 30, and a storage 32. The processor 28, the RAM 30, and the storage 32 are connected to a bus 34. The database 24 and the communication I / F 26 are also connected to the bus 34. The communication I / F 26 is connected to a network 54. Examples of the network 54 include a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[1276] 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.

[1277] The microphone 238 receives instructions and the like from the user 20 by receiving voice uttered by the user 20. The microphone 238 captures the voice uttered by the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio in accordance with instructions from the processor 46.

[1278] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an imaging element such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the surroundings of user 20 (for example, an imaging range defined by an angle of view equivalent to the field of vision of a typical healthy person).

[1279] 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.

[1280] The control object 443 includes a display device, LEDs in the eyes, and motors for driving the arms, hands, and feet. The posture and gestures of the robot 414 are controlled by controlling the motors of the arms, hands, and feet. Some of the emotions of the robot 414 can be expressed by controlling these motors. In addition, the facial expressions of the robot 414 can also be expressed by controlling the light emission state of the LEDs in the eyes of the robot 414.

[1281] 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.

[1282] The specific processing program 56 is an example of a "program" according to the technology of the present disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific processing is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific processing program 56 executed on the RAM 30.

[1283] 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.

[1284] In the robot 414, the processor 46 performs the reception output process. A reception output program 60 is stored in the storage 50. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output process is realized by the processor 46 operating as the control unit 46A in accordance with the reception output program 60 executed on the RAM 48.

[1285] Next, a description will be given of the specific processing performed by the specific processing unit 290 of the data processing device 12. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal."

[1286] The present invention is a system that monitors answers and files generated by other AI systems, detects suspicious content, and issues a warning to users. This system operates as software installed on a server, and plays a role in enhancing user security.

[1287] Basic system configuration

[1288] The system consists of the following components:

[1289] 1. Answer and file monitoring module: This module takes answers and files generated by other AI systems and analyzes them for suspicious content.

[1290] 2. URL Monitoring Module: Monitors referenced URLs in answers and files to check whether they are fraudulent or malicious sites.

[1291] 3. Program code monitoring module: Checks the program code generated by other AI systems to monitor whether it contains malicious code.

[1292] 4. Alert module: Issues an alert to the user if a problem is detected as a result of monitoring.

[1293] Program processing

[1294] 1. Monitoring responses and files

[1295] Server: The server receives answers and files generated by other AI systems, treating the received content as strings or data.

[1296] Server: The analysis module checks the content using predefined keywords (e.g., phishing, malware, suspicious) and issues a warning if any suspicious items are included.

[1297] 2. URL Monitoring

[1298] Server: Parses messages and content containing referral URLs and extracts the URLs.

[1299] Server: Sends an HTTP request to the extracted URL and retrieves the page content.

[1300] Server: Analyzes the retrieved page content and checks the text against predefined patterns to detect fraudulent or malicious sites, and issues a warning if a problem is detected.

[1301] 3. Program code monitoring

[1302] Server: Obtains program code generated by other AI systems, and analyzes the instructions in the code as strings.

[1303] Server: Uses regular expressions to check whether a program contains suspicious code patterns such as "import os", "import sys", or "exec(", and issues a warning if they are found.

[1304] 4. Issuing a warning

[1305] Server: A warning module runs and notifies the user of detailed information about detected problems. Warnings are mainly output to the console log, email, or as messages on the GUI.

[1306] Specific examples

[1307] For example, when another AI system recommends a link to a user, the server must verify whether the link is a fraudulent site. If the server analyzes the link and determines that it is likely to be a fraudulent site, it warns the user that "this link is not safe." Also, if the program code provided to the user by the other AI system contains fraudulent operations, the server detects this and issues a warning that "this code is dangerous."

[1308] This will allow users to minimize the risk from malicious AI technology and use the system with peace of mind, and this system will play an important role in improving user security.

[1309] The processing flow will be explained below.

[1310] Step 1:

[1311] The server receives answers and files generated by other AI systems, and the data is stored on the server as a string or file.

[1312] Step 2:

[1313] The server analyzes the responses and files it receives and checks their content against predefined keywords (e.g., phishing, malware, suspicious). The analysis module is responsible for this and checks whether any suspicious words are included.

[1314] Step 3:

[1315] The server extracts the reference URLs contained in the response or file, using regular expressions to extract the URLs from the text, and creates a list.

[1316] Step 4:

[1317] The server sends an HTTP request to the extracted URL to retrieve the content of the target web page. The request module issues an HTTP request and receives the response from the server.

[1318] Step 5:

[1319] The server performs text analysis on the retrieved web page content, checking for patterns (e.g., phishing terms, suspicious links) to detect fraudulent or malicious sites. The content is analyzed as text to see if it contains any suspicious patterns.

[1320] Step 6:

[1321] The server acquires program code generated by other AI systems and analyzes the code content. At this time, the code is acquired as a string and is the subject of inspection.

[1322] Step 7:

[1323] The server uses regular expressions to check whether the acquired program code contains suspicious code patterns such as "import os", "import sys", or "exec(". The analysis module checks each line of the code and records any suspicious patterns it finds.

[1324] Step 8:

[1325] The server combines the analysis results from each module and, if any suspicious content is detected, generates a warning message that contains detailed information about the detected problem and is presented in a user-friendly format.

[1326] Step 9:

[1327] The server notifies the user of any generated warning messages via console logs, email, or messages on the user interface, allowing the user to quickly identify suspicious areas.

[1328] This allows the server to continuously monitor the outputs of other artificial intelligence systems, creating a system that ensures user security.

[1329] Example 1

[1330] Next, a description will be given of Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal."

[1331] Answers and files generated by conventional AI systems sometimes contain malicious content, fraudulent websites, or malicious code, posing a threat to user security. Furthermore, there is a lack of systems that can efficiently detect such malicious content and quickly warn users. Therefore, there is a need for a reliable monitoring system that allows users of AI systems to use them with peace of mind.

[1332] The specific processing by the specific processing unit 290 of the data processing device 12 in the first embodiment is realized by the following means.

[1333] In this invention, the server includes means for monitoring answers and files generated by other artificial intelligence systems and analyzing suspicious content using predefined keywords, means for monitoring referred URLs and analyzing page content obtained by sending HTTP requests to detect fraudulent or malicious sites, means for monitoring generated program code and checking specific code patterns with regular expressions to detect whether malicious programs are included, and means for notifying users of detailed information about detected problems via console logs, email, or GUI messages. This makes it possible to effectively detect security risks contained in content generated by artificial intelligence systems and promptly notify users.

[1334] "Answers and files" refers to text data and document data generated by other artificial intelligence systems.

[1335] "Monitoring" means the act of continuously or periodically checking or auditing a particular object.

[1336] "Keywords" refers to predefined important phrases or words for a specific purpose.

[1337] "Referring URL" refers to a string of characters that indicates a link to a web page or resource.

[1338] "HTTP request" refers to a request sent by a web browser or client application to retrieve information from a web server.

[1339] "Page Content" refers to the text, images, scripts, and other information displayed on a web page.

[1340] "Fraudulent Site" refers to a fraudulent website designed to deceive users.

[1341] "Malicious Site" means a website whose purpose is to steal user information, cause damage to systems, or perform other fraudulent activities.

[1342] "Program code" refers to the source code for writing a computer program.

[1343] A "code pattern" refers to a set of notations or instructions that have a specific purpose or function.

[1344] "Regular expression" refers to a formal language for describing specific patterns.

[1345] "Console log" refers to character-based output that displays system or application status and error messages.

[1346] "Mail" means electronic mail, a means of sending and receiving messages in digital form.

[1347] "GUI Message" means a warning or informational message displayed on a graphical user interface.

[1348] The present invention is a system that monitors answers and files generated by other AI systems, detects suspicious content, and issues a warning to the user. Specific embodiments for implementing this system are described below.

[1349] Basic system configuration

[1350] The system runs as software installed on a server and consists of the following components:

[1351] 1. Answer and file monitoring module

[1352] 2. URL Monitoring Module

[1353] 3. Program Code Monitoring Module

[1354] 4. Warning Module

[1355] Monitor responses and files

[1356] server

[1357] The server retrieves answers and files generated by other AI systems, using API requests and file system monitoring.

[1358] The retrieved answers and file contents are analyzed as text data, and the server searches the text using predefined keywords (e.g., phishing, malware, suspicious) to extract suspicious content.

[1359] URL monitoring

[1360] server

[1361] The server extracts URLs from the responses and files and generates a list of URLs using regular expressions.

[1362] Send an HTTP request to the extracted URL and get the response.

[1363] The text content of the response is analyzed and matched against patterns (e.g., characteristics of phishing pages) to detect fraudulent or malicious websites.

[1364] Program code monitoring

[1365] server

[1366] The server retrieves program code generated by other artificial intelligence systems from specific folders or APIs.

[1367] The commands in the acquired program code are analyzed as strings and checked for specific patterns (e.g., "import os", "import sys", "exec()") using regular expressions.

[1368] Sending a warning

[1369] server

[1370] The server generates detailed warning messages about any problems it detects, including the type, location, and detailed description of the problem detected.

[1371] The generated warning messages are not only output to the console log, but also notified to the user by email and displayed as a popup message on the GUI.

[1372] Specific examples

[1373] A specific example of operation is shown below.

[1374] Example 1: If the server retrieves a new answer and it contains the keyword "phishing", the server will notify the user "Warning: This answer contains suspicious content."

[1375] Example 2: If the server sends an HTTP request to a URL extracted from a file and the result has characteristics of a phishing site, the server notifies the user with "Warning: This URL is not safe."

[1376] Example 3: When the server parses a new code file and it contains the instruction "exec(", the server will notify the user "Warning: this code is dangerous".

[1377] Prompt Sentence Examples

[1378] Check if this link is safe

[1379] "Check this code for illegal instructions"

[1380] By inputting such prompt sentences into the generative AI model, the server will initiate a specific analysis process, allowing users to obtain reliable information and minimizing security risks.

[1381] The flow of the identification process in the first embodiment will be described with reference to FIG.

[1382] Step 1:

[1383] Get answers and files

[1384] Server: Retrieves answers and files generated by other AI systems, using API requests and file system monitoring.

[1385] Input: Answers or files generated by other AI systems.

[1386] Output: Text data of the obtained answers and files.

[1387] Step 2:

[1388] Check the contents

[1389] Server: Analyzes the obtained answers and file contents as text data. Analyzes suspicious content using predefined keywords (e.g., phishing, malware, suspicious).

[1390] Input: Text data from the answers or files obtained.

[1391] Output: Result on whether or not the content is suspicious.

[1392] What happens: The server searches for text against a list of keywords and flags it as suspicious if it contains a matching keyword.

[1393] Step 3:

[1394] URL extraction

[1395] Server: Extract URLs from responses and files using regular expressions.

[1396] Input: The text data from the parsed answer or file.

[1397] Output: The extracted URL list.

[1398] What it does: The server searches for strings that start with "http: / / " or "https: / / " and lists all URLs.

[1399] Step 4:

[1400] URL validation

[1401] Server: Sends an HTTP request to the extracted URL and obtains the response data.

[1402] Input: The extracted URL list.

[1403] Output: Response data for the URL.

[1404] Specific operation: The server sends an HTTP request for each URL and saves the page content of the response it receives.

[1405] Step 5:

[1406] Analyzing page content

[1407] Server: Analyzes the response data and matches it with patterns (e.g., characteristics of phishing sites) to detect fraudulent or malicious sites.

[1408] Input: Response data for the URL.

[1409] Output: Detected fraudulent and malicious sites.

[1410] What it does: The server performs pattern matching on the text content of the response and flags any fraudulent or malicious features detected.

[1411] Step 6:

[1412] Obtaining the program code

[1413] Server: Retrieves program code generated by other AI systems from specific folders or APIs.

[1414] Input: Program code generated by another artificial intelligence system.

[1415] Output: Text data of the retrieved program code.

[1416] What it does: The server monitors newly added code files and incorporates them one by one.

[1417] Step 7:

[1418] Code content analysis

[1419] Server: Analyzes the program code and checks for specific code patterns (e.g., "import os", "import sys", "exec()") using regular expressions.

[1420] Input: Text data of the retrieved program code.

[1421] Output: Suspicious code pattern detection results.

[1422] What it does: The server parses instructions in your code using regular expressions and flags when certain patterns are detected.

[1423] Step 8:

[1424] Generate a warning message

[1425] Server: Generates detailed warning messages about detected problems.

[1426] Input: Suspicious content, URL, or code pattern detection results.

[1427] Output: A detailed warning message.

[1428] What happens: The server generates a message containing the type, location, and detailed description of the problem detected.

[1429] Step 9:

[1430] User Notification

[1431] Server: Generated warning messages are output to the console log and notified to the user via email or GUI message.

[1432] Input: A detailed warning message.

[1433] Output: The warning message that was sent to the user.

[1434] Specific behavior: The server not only logs the warning to the console, but also sends an email to the user and displays a warning as a popup message in the GUI.

[1435] (Application example 1)

[1436] Next, a description will be given of Application Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal."

[1437] With the development of modern information and communication technology, security risks are increasing for answers, files, and program code generated by other artificial intelligence systems. Furthermore, the potential for suspicious URLs and malicious programs contained in this content to cause harm to users is also increasing. In particular, with the widespread use of smart devices, real-time monitoring and alerts are becoming increasingly important. However, current systems lack the means to effectively address these multifaceted threats.

[1438] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 1 is realized by the following means.

[1439] In this invention, the server includes means for monitoring answers and files generated by other artificial intelligence systems and checking for suspicious content, means for monitoring referred URLs and detecting fraudulent or malicious sites, means for monitoring generated program code and checking whether it contains malicious programs, means for issuing a warning when malicious content is detected, and means for operating on the smart device and monitoring the user's digital content in real time, thereby enabling real-time monitoring and warning of the digital content used by the user.

[1440] "Other artificial intelligence systems" refers to artificial intelligence technologies operated or utilized by users to analyze and generate data.

[1441] "Answer or file" refers to text, documents, images, audio, or other forms of data generated by other artificial intelligence systems.

[1442] "Suspicious Content" refers to elements within a file or message that may indicate phishing, malware, spam, or other malicious content.

[1443] "Referring URL" refers to a website address or link contained in a response or file.

[1444] A "fraudulent site" refers to a malicious website that attempts to fraudulently obtain users' personal information or money.

[1445] "Malicious Sites" refers to websites that install malware on users' devices or perform other malicious activities.

[1446] "Program Code" refers to the source code of software generated by other artificial intelligence systems.

[1447] "Malicious Program" refers to source code or a program designed to perform unauthorized operations or undesirable behavior.

[1448] "Warning" refers to a notification or alert issued to a user when the system detects suspicious or malicious content.

[1449] "Smart devices" refers to electronic devices such as mobile phones, tablets, and smart glasses that have internet connectivity.

[1450] "Digital content" refers to electronically stored information or data, such as web pages, messages, files, and program code.

[1451] "Real-time" refers to data being processed immediately at the moment it is generated or acquired.

[1452] This invention provides a security system called "AI Guard," which monitors answers and files generated by other artificial intelligence systems and analyzes whether they contain suspicious content. It also monitors referred URLs to determine whether they are fraudulent or malicious sites, and checks generated program code to monitor whether they contain malicious programs. Based on these monitoring results, it issues a warning to the user when malicious content is detected. Furthermore, this system runs on smart devices and can monitor users' digital content in real time.

[1453] System Configuration

[1454] The system consists of the following elements:

[1455] 1. Answer and file monitoring module:

[1456] The server receives answers and files generated by other AI systems, whose contents are processed as strings or data.

[1457] The server analyzes the content using predefined keywords (e.g., phishing, malware, suspicious) and issues a warning if it contains any suspicious items.

[1458] 2. URL Monitoring Module:

[1459] Parse messages and content that contain URLs and extract them.

[1460] An HTTP request is sent to the extracted URL to retrieve the page content.

[1461] It analyzes the retrieved page content, checks the text against defined patterns to detect fraudulent or malicious sites, and issues a warning if a problem is detected.

[1462] 3. Program code monitoring module:

[1463] It acquires program code generated by other AI systems and analyzes the instructions in the code as strings of characters.

[1464] It uses regular expressions to check whether a program contains suspicious code patterns such as "import os", "import sys", or "exec(", and issues a warning if they are found.

[1465] 4. Alert sending module:

[1466] The server includes a module to notify the user with detailed information about detected problems. Warnings are typically output to the console log, email, or as messages in the graphical user interface (GUI).

[1467] Hardware and software used

[1468] Hardware: Smartphones (e.g., iPhones and Android devices), smart glasses (e.g., Google Glass).

[1469] Software: Python programming language, re module (regular expression processing), requests module (HTTP request processing).

[1470] Example of a system

[1471] The "AI Guard" app creates warning messages for users by inputting prompt sentences such as the following into a generative AI model:

[1472] Prompt Sentence Examples

[1473] "The site you are trying to visit, URL: {url}, may be fraudulent or malicious. We do not recommend visiting this site. More information: {details}"

[1474] Specific examples

[1475] For example, if a user receives a link via SMS to "http: / / phishing-site.com," the following steps will occur:

[1476] The server parses the URL and checks it against a blacklist.

[1477] Get the page content and detect the keyword "phishing".

[1478] It warns users that "this link is not safe."

[1479] In this way, the "AI Guard" system of the present invention plays a role in keeping users' digital lives safe by monitoring their digital content in real time and early detecting and warning of potential security risks.

[1480] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[1481] Step 1:

[1482] The server receives the answers and files generated by other AI systems. The input is the answers and files generated by other AI systems, which are provided in the form of strings and data. The server receives this and analyzes it in the next step.

[1483] Step 2:

[1484] The server analyzes the retrieved answers and file contents using keyword filtering technology to detect suspicious keywords (e.g., phishing, malware, suspicious). The input is text data, and the output is the analysis result, indicating whether or not suspicious keywords are present. Through this process, the server makes an initial detection of suspicious content.

[1485] Step 3:

[1486] The server extracts the referrer URLs contained in the parsed answers and files. The input is the text data of the answers and files, and the output is the extracted URLs. This is then passed to the URL monitoring module.

[1487] Step 4:

[1488] The server sends an HTTP request to the URL extracted using the URL monitoring module to obtain the web page content. The input is the extracted URL, and the output is the HTML content of the web page corresponding to the URL. In this process, the server communicates over the network to obtain the page content of the target URL.

[1489] Step 5:

[1490] The server analyzes the retrieved HTML content and checks the text against predefined patterns (e.g., phishing keywords, malware download links) to detect fraudulent or malicious sites. The input is the HTML data of the web page, and the output is a judgment result on whether the page is suspicious or not. This allows the server to evaluate the safety of the URL.

[1491] Step 6:

[1492] The server retrieves program code generated by other AI systems and analyzes its contents as strings. The input is the program code, and the output is the analysis result, indicating whether or not there are suspicious code patterns. In this process, the server scans the strings of the program code with regular expressions to detect malicious code.

[1493] Step 7:

[1494] The server issues a warning to the user based on the detection results. The input is the analysis results and judgment results, and the output is a warning message. The warning is notified to the user via email, log, or GUI message. This process allows the user to receive warnings about security risks in real time.

[1495] Through the above processing steps, the server can monitor the user's digital content in real time and issue early warnings of potential security risks.

[1496] Furthermore, an emotion engine that estimates the user's emotion may be further combined. That is, the identification processing unit 290 may estimate the user's emotion using the emotion identification model 59, and perform identification processing using the user's emotion.

[1497] This invention integrates an emotion engine into a system that monitors answers and files generated by other AI systems, detects suspicious content, and issues warnings to users. This system operates as software installed on a server, and aims to improve user security and experience by adjusting the content and presentation of warnings according to the user's emotional state.

[1498] Basic system configuration

[1499] The system consists of the following components:

[1500] 1. Answer and file monitoring module: This module takes answers and files generated by other AI systems and analyzes them for suspicious content.

[1501] 2. URL Monitoring Module: Monitors referenced URLs in answers and files to check whether they are fraudulent or malicious sites.

[1502] 3. Program code monitoring module: Checks the program code generated by other AI systems to monitor whether it contains malicious code.

[1503] 4. Alert module: Issues an alert to the user if a problem is detected as a result of monitoring.

[1504] 5. Emotion Engine: Recognizes user emotions and adjusts the content and presentation of warning messages based on those emotions.

[1505] Program processing

[1506] 1. Monitoring responses and files

[1507] Server: The server receives answers and files generated by other AI systems, treating the received content as strings or data.

[1508] Server: The analysis module checks the content using predefined keywords (e.g., phishing, malware, suspicious) to see if it contains any suspicious items.

[1509] 2. URL Monitoring

[1510] Server: Parses messages and content containing referral URLs and extracts the URLs.

[1511] Server: Sends an HTTP request to the extracted URL and retrieves the content of the target web page.

[1512] Server: Analyzes the retrieved page content and checks the text against predefined patterns to detect fraudulent or malicious sites, and issues a warning if a problem is detected.

[1513] 3. Program code monitoring

[1514] Server: Obtains program code generated by other AI systems. At this time, the code is obtained as a string and is subject to inspection.

[1515] Server: Regular expressions are used to check whether the program contains suspicious code patterns such as "import os", "import sys", or "exec(". The analysis module checks each line of the code and records any suspicious patterns found.

[1516] 4. Emotion recognition

[1517] Server: The emotion engine receives the user's voice, text, and facial expression data and performs emotion analysis. The emotion analysis module analyzes the user's tone of voice, words used, and facial expression changes to assess the user's emotional state (e.g., stress, anxiety, satisfaction).

[1518] Server: The emotion engine adjusts the display and content of warning messages based on the detected user emotional state. For example, if the user is feeling anxious, the warning message will be presented in a more understandable and less urgent manner.

[1519] 5. Issuing a warning

[1520] Server: A warning module runs and notifies the user of detailed information about detected problems. Warnings are mainly output to the console log, email, or as messages on the GUI.

[1521] Server: The emotion engine optimizes the priority and detail of the warning message based on the user's current emotional state, providing the user with an appropriate warning.

[1522] Specific examples

[1523] For example, when another AI system recommends a link to a user, it must verify whether the link is a fraudulent site. The server analyzes the link and, if it determines that it is likely to be a fraudulent site, warns the user, "This link is not safe." If the emotion engine detects anxiety from the user's tone of voice or facial expression, it provides a softer warning message and adds a help link if necessary. Furthermore, if the program code provided to the user by the other AI system contains fraudulent operations, the server detects this and issues a warning, saying, "This code is dangerous." If the emotion engine detects the user's state of stress, it simplifies the warning message and presents additional support options.

[1524] This allows the server to continuously monitor the outputs of other artificial intelligence systems and provide appropriate warnings based on the user's emotional state, improving both security and user experience.

[1525] The processing flow will be explained below.

[1526] Step 1:

[1527] The server receives answers and files generated by other AI systems, and the data is stored on the server as a string or file.

[1528] Step 2:

[1529] The server analyzes the responses and files it receives and checks their content against predefined keywords (e.g., phishing, malware, suspicious). The analysis module is responsible for this and checks whether any suspicious words are included.

[1530] Step 3:

[1531] The server extracts the reference URLs contained in the response or file, using regular expressions to extract the URLs from the text, and creates a list.

[1532] Step 4:

[1533] The server sends an HTTP request to the extracted URL to retrieve the content of the target web page. The request module issues an HTTP request, and the server receives the response.

[1534] Step 5:

[1535] The server performs text analysis on the retrieved web page content, checking for patterns (e.g., phishing terms, suspicious links) to detect fraudulent or malicious sites. The content is analyzed as text to see if it contains any suspicious patterns.

[1536] Step 6:

[1537] The server acquires program code generated by other AI systems and analyzes the code content. At this time, the code is acquired as a string and is the subject of inspection.

[1538] Step 7:

[1539] The server uses regular expressions to check whether the acquired program code contains suspicious code patterns such as "import os", "import sys", or "exec(". The analysis module checks each line of the code and records any suspicious patterns it finds.

[1540] Step 8:

[1541] The server uses an emotion engine to recognize the user's emotions. It analyzes the voice, text, and facial expression data obtained from the user to evaluate their emotional state (e.g., stress, anxiety, satisfaction). This analysis includes voice tone analysis, emotional keyword analysis in the text, and facial expression analysis.

[1542] Step 9:

[1543] The server adjusts the content and presentation of the warning message based on the emotional state recognized by the emotion engine. For example, if the user expresses anxiety, the server changes the warning message to be more gentle and detailed.

[1544] Step 10:

[1545] The server combines the analysis results from each module and, if any suspicious content is detected, generates a warning message that contains detailed information about the detected problem and is presented in an easy-to-understand format to the user.

[1546] Step 11:

[1547] The server notifies the user of the generated warning message. Notification methods include console log, email, or a message on the user interface. The priority and detail of the message are adjusted according to the analysis results of the emotion engine. This allows the user to quickly understand which parts are suspicious.

[1548] This allows the server to continuously monitor the outputs of other artificial intelligence systems, ensuring user security and providing warnings that take into account the user's emotional state.

[1549] Example 2

[1550] Next, a description will be given of Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal."

[1551] Conventional monitoring systems primarily focus on monitoring answers, files, reference links, and programs generated by other AI systems. However, because they do not take the user's emotional state into consideration, the content and display method of warning messages are uniform, which can lead to poor user experience. In particular, appropriate responses are required for users experiencing stress or anxiety, but these responses are insufficient.

[1552] The identification process by the identification processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means. In this invention, the server includes a means for monitoring answers and files generated by other artificial intelligence systems and checking for suspicious content, a means for monitoring reference links and detecting fraudulent or malicious pages, a means for monitoring generated programs and checking whether they contain malicious software, a means for issuing a warning when malicious content is detected, and a means for recognizing the user's emotional state and adjusting the display method and content of the warning message. This makes it possible to improve security and the user experience by providing appropriate warnings according to the user's emotional state.

[1553] "Suspicious content" is content in network responses or files that indicates security threats such as phishing or malware.

[1554] A "reference link" is a URL or web address contained in a response or file.

[1555] A "scam page" is a malicious website designed to deceive users and obtain their personal information.

[1556] A "malicious page" is a website that contains computer viruses or malware and is designed to harm a user's system.

[1557] "Generated Program" refers to program code created by another artificial intelligence system.

[1558] "Malicious software" refers to programs designed to cause damage to systems or data, including viruses, malware, and spyware.

[1559] A "warning message" is a message intended to notify the user of problems or potential risks detected by the system.

[1560] "Emotional state" refers to a user's psychological state, such as stress, anxiety, or satisfaction, and is recognized through voice, text, and facial expression data.

[1561] An "emotion recognition engine" is a software module for analyzing and evaluating a user's emotional state.

[1562] This invention integrates an emotion engine into a system that monitors answers and files generated by other AI systems, detects suspicious content, and issues warnings to users. This system operates as software installed on a server, and aims to improve user security and experience by adjusting the content and presentation of warnings according to the user's emotional state.

[1563] Basic system configuration

[1564] The system consists of the following components:

[1565] 1. Answer and file monitoring module: This module takes answers and files generated by other AI systems and analyzes them for suspicious content.

[1566] 2. URL Monitoring Module: Monitors referral links in answers and files to ensure they are not fraudulent or malicious.

[1567] 3. Program code monitoring module: Checks the program code generated by other AI systems to ensure it does not contain malicious software.

[1568] 4. Alert module: Issues an alert to the user if a problem is detected as a result of monitoring.

[1569] 5. Emotion Recognition Engine: Recognizes user emotions and adjusts the content and presentation of warning messages based on those emotions.

[1570] System processing flow and specific examples

[1571] Watch for answers and files:

[1572] Server: Periodically retrieves answers and files generated by other AI systems via API calls and database queries.

[1573] Server: Saves the obtained answers and files as string or binary data.

[1574] Server: The analysis module runs and performs text analysis on the stored data, checking for suspicious content using predefined keywords (e.g., phishing, malware, suspicious).

[1575] Extract and validate the URL:

[1576] Server: Analyzes the text data extracted from the responses and files and extracts all the URLs contained within.

[1577] Server: Sends an HTTP request to each extracted URL to retrieve the web page content, including timeout and error handling.

[1578] Server: Analyzes the HTML content of retrieved web pages and filters them to detect fraudulent or malicious sites by checking for specific keywords and patterns.

[1579] Program code monitoring:

[1580] Server: Retrieves program code generated by other AI systems in the form of strings, via API calls or database queries.

[1581] Server: Analyzes the acquired program code line by line and checks for suspicious code patterns using regular expressions. For example, it performs pattern matching on "import os", "import sys", "exec(", etc.

[1582] Emotion Recognition and Analysis:

[1583] Device: Captures the user's voice, text, and facial expression data using input devices such as a microphone, keyboard, and camera.

[1584] Server: Receives data sent from the device and passes it to the emotion engine.

[1585] Server: The emotion engine analyzes the tone of voice, words used, and facial expressions to assess the user's emotional state.

[1586] Sending a warning:

[1587] Server: The alert module is launched to collect detailed information about the problem detected in the previous step.

[1588] Server: Using the evaluation results of the emotion engine, generate a warning message according to the user's emotional state. If the user is feeling anxious, use a less urgent message such as "This link is not safe. Please rest assured, we will take action immediately."

[1589] Server: Notifies the user of a warning message in a variety of formats, including console log, email, or a GUI popup message.

[1590] Examples of concrete examples and prompts

[1591] For example, suppose a user is recommended a link and needs to verify whether it is a fraudulent page. The server analyzes the link and, if it determines that it is likely a fraudulent page, warns the user that "This link is not safe." If the emotion engine detects anxiety in the user's tone of voice or facial expression, it will provide a softer warning message and, if necessary, add a help link.

[1592] Example prompt sentence:

[1593] "Please explain how your system verifies the possibility that other AI-generated links may be fraudulent sites and issues an appropriate warning to the user if it finds them. Also, please explain how you would recognize the user's emotions and adjust the warning message accordingly."

[1594] In this way, the server, terminal, and user work together to carry out the process, creating a system that improves the user experience while ensuring the security of the information provided by the generative AI model.

[1595] The flow of the identification process in the second embodiment will be described with reference to FIG.

[1596] Step 1: Get answers and files

[1597] Input: Answers generated from other artificial intelligence systems or data from files.

[1598] Server: Periodically retrieves generated answers and files from other AI systems using REST API calls and database queries.

[1599] Server: Saves the retrieved data as string data or binary data.

[1600] Output: Saved answers and file data.

[1601] Step 2: Monitor and analyze responses and files

[1602] Input: Saved answers and file data.

[1603] Server: The analysis module runs and performs text analysis on the stored data, checking for suspicious content using predefined keywords (e.g., phishing, malware, suspicious).

[1604] Server: If suspicious content is detected as a result of text analysis, it will record the content in a log file and notify the warning module.

[1605] Output: Analysis results and log file updates.

[1606] Step 3: Extracting the URL

[1607] Input: Saved answers and file data.

[1608] Server: Analyzes the text data extracted from the responses and files and extracts all the URLs contained within.

[1609] Server: Use text analysis tools and libraries to identify and catalog URL patterns in responses and files.

[1610] Output: A list of extracted URLs.

[1611] Step 4: Verify the URL

[1612] Input: A list of extracted URLs.

[1613] Server: Sends HTTP requests to each URL to retrieve the web page content, including timeout and error handling.

[1614] Server: Analyzes the HTML of the retrieved page content and filters it to detect fraudulent or malicious sites, checking for specific keywords and suspicious patterns.

[1615] Server: Records the results of URL validation in a log file and notifies the alert module if a problem is detected.

[1616] Output: URL validation result and log file update.

[1617] Step 5: Monitoring the program code

[1618] Input: Program code generated by another artificial intelligence system.

[1619] Server: Retrieves the generated program code in string form, either via an API call or a database query.

[1620] Server: Analyzes the acquired program code line by line and checks for suspicious code patterns (e.g., import os, import sys, exec()) using regular expressions.

[1621] Server: If pattern matching detects suspicious code, identify the line and log it with a detailed comment.

[1622] Output: Code analysis results and log file updates.

[1623] Step 6: Emotion Recognition and Analysis

[1624] Input: User voice, text, and facial expression data.

[1625] Device: Captures the user's voice, text, and facial expression data using input devices such as a microphone, keyboard, and camera.

[1626] Server: Receives data sent from the device and passes it to the emotion engine.

[1627] Server: The emotion engine analyzes the tone of voice, words used, and facial expressions to assess the user's emotional state.

[1628] Server: Quantifies or categorizes the emotion evaluation results and notifies the warning module.

[1629] Output: Emotion data reflected in the emotion rating results and warnings.

[1630] Step 7: Sending an alert

[1631] Input: Analysis results and emotion evaluation results.

[1632] Server: The warning module is launched to collect detailed information about the problem detected in the previous step. Using the evaluation results of the emotion engine, a warning message is generated based on the user's emotional state.

[1633] Server: Notifies each user of warning messages in a format suitable for them, such as console logs, emails, or GUI popup messages.

[1634] User: Receives a warning message and takes action. Users can also view the provided help links and support options.

[1635] Output: Warning message provided to user and log file updates.

[1636] In this way, by having the server, terminal, and user work together to carry out processing, we have created a system that improves the user experience while ensuring the security of the information provided by the generative AI model.

[1637] (Application example 2)

[1638] Next, a description will be given of Application Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal."

[1639] The monitoring of answers, files, reference URLs, and program code generated by conventional AI systems has limited ability to detect suspicious content and malicious sites, making it difficult to fully guarantee user security. Furthermore, these systems issue warnings without taking into account the user's emotional state, which can lead to a poor user experience.

[1640] The identification processing by the identification processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means. In this invention, the server includes means for monitoring answers and files generated by other artificial intelligence systems and checking for suspicious content, means for monitoring reference URLs and detecting fraudulent or malicious sites, means for monitoring generated program code and checking whether it contains malicious programs, and means for issuing a warning based on the user's emotional state when malicious content is detected. This makes it possible to improve user security and the user experience.

[1641] "Answers and files generated by other artificial intelligence systems" refers to responses and document data automatically created by external artificial intelligence systems.

[1642] "Means for checking for suspicious content" refers to a method or device that has the functionality to analyze and detect fraudulent or malicious content contained in messages or files.

[1643] "Means for monitoring referring URLs to detect fraudulent or malicious sites" refers to a method or device for monitoring URLs contained in responses or files and determining whether the linked site is a fraudulent or malicious site.

[1644] "Means for monitoring generated program code and checking whether it contains malicious code" refers to a method or device for monitoring program code generated by another artificial intelligence system and detecting malicious code contained therein.

[1645] "Means for issuing a warning based on the user's emotional state when malicious content is detected" refers to a method or device for issuing a warning in an optimal manner based on the results of an emotional analysis of the user in response to detected suspicious content.

[1646] "Means for assessing emotional state" means a method or device for analyzing a user's text or other information to determine whether the user is in a positive, negative, or neutral emotional state.

[1647] "Means for adjusting the display method and content of a warning message" refers to a method or device for appropriately adjusting a warning message according to the user's emotional state and displaying it in the most suitable format and content for the user.

[1648] The system of the present invention includes means for monitoring answers and files generated by other artificial intelligence systems and checking for suspicious content, means for monitoring referred URLs and detecting fraudulent or malicious sites, means for monitoring generated program code and checking whether it contains malicious programs, and means for issuing a warning based on the user's emotional state when malicious content is detected.

[1649] 1. Answers and file monitoring methods:

[1650] The server monitors answers and files generated by external AI systems in real time. This includes analyzing the content using predefined keywords (e.g., phishing, malware, suspicious) to check for suspicious content. For example, if the monitored text contains a keyword such as "phishing," the server will determine that the text is suspicious.

[1651] 2. URL monitoring methods:

[1652] The server extracts the URL contained in the response or file and sends an HTTP request to retrieve the content of the web page pointed to by the URL. The retrieved page content is then analyzed and checked against patterns designed to detect fraudulent or malicious sites. For example, a message containing the URL "http: / / example-phishing.com" is analyzed, and if the URL is determined to be a fraudulent site, an alert is issued.

[1653] 3. Program code monitoring methods:

[1654] The server monitors program code generated by external artificial intelligence systems, checking it using regular expressions to see if it contains malicious program patterns such as "import os," "import sys," or "exec(." If suspicious code is detected, it records its contents and issues a warning to the user.

[1655] 4. Emotion recognition means:

[1656] The server receives the user's voice, text, and facial expression data and performs emotion analysis using an emotion engine. This analyzes the user's tone of voice, words used, and facial expression changes to assess the user's emotional state (e.g., stress, anxiety, satisfaction), and adjusts the display method and content of warning messages descriptively based on the user's emotional state.

[1657] 5. Warning Method:

[1658] The server optimizes the content and presentation of warning messages based on the user's emotional state, as detected by the emotion engine. For example, if the user is feeling anxious, the server will present the warning message in a more understandable and less stressful format, and also add a help link with specific actions to take.

[1659] Hardware and software used

[1660] Server: Server hardware that includes a central processing unit (CPU), memory, and storage.

[1661] Emotion Recognition Engine: Uses TextBlob as a model for emotion analysis.

[1662] URL monitoring: Uses the Requests library to send HTTP requests.

[1663] Program code monitoring: Uses the re library for regular expression analysis.

[1664] Sending email: Uses the smtplib library to send alert messages.

[1665] Specific examples

[1666] For example, when a user receives a message, it may contain a link to a fraudulent website. The server analyzes the URL and, if it determines that the site is likely fraudulent, warns the user, "This link is not safe." At the same time, the emotion engine analyzes the user's text, "I have a bad feeling about this," and, if it detects a negative emotional state, it makes the warning message more gentle and adds a help link if necessary.

[1667] Prompt sentence for generative AI model

[1668] "Analyze text entered by users and assess its emotional state (positive, negative, neutral). Develop an application that monitors provided messages, URLs, and program code for suspicious content and issues alerts accordingly."

[1669] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[1670] Step 1:

[1671] The server receives answers or files generated by other AI systems. These answers or files are input as string data. The server analyzes the content using predefined keywords (e.g., phishing, malware, suspicious) to check for suspicious items. The results of this analysis are provided as output.

[1672] Step 2:

[1673] The server automatically extracts the URLs contained in the response or file. The list of URLs is input. The server sends an HTTP request to each URL to retrieve the web page content. This content is output as text data.

[1674] Step 3:

[1675] The server analyzes the content of the retrieved web page and checks the text against predefined patterns to detect fraudulent or malicious sites. If a fraudulent or malicious site is detected, the URL and reason for the detection are printed.

[1676] Step 4:

[1677] The server receives program code generated by other AI systems. The code is input as string data. The server uses regular expressions to check whether the program contains suspicious code patterns such as "import os," "import sys," or "exec(." The server then outputs the analysis results.

[1678] Step 5:

[1679] The server receives the user's voice, text, and facial expression data. These data are used as input. The emotion engine analyzes the user's tone of voice, the words used, and changes in facial expressions to assess the user's emotional state (e.g., stress, anxiety, satisfaction). This emotion assessment is output.

[1680] Step 6:

[1681] The server adjusts the display method and content of the warning message based on the user's emotional state detected by the emotion engine. The warning message template is used as input, and a warning message adapted to the user's emotional state is generated as the final output.

[1682] Step 7:

[1683] The device receives the optimized warning message sent from the server and displays it to the user, specifically, a warning message containing gentle language and additional support information according to the user's emotional state.

[1684] 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.

[1685] The data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of the data generation model 58 is ChatGPT (Internet Search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search <url: https: gemini.google.com ?hl="ja">) and other generation AIs. The data generation model 58 is obtained by performing deep learning on a neural network. A prompt including an instruction is input to the data generation model 58, and inference data such as voice data indicating voice, text data indicating text, and image data indicating an image is also input. The data generation model 58 performs inference on the input inference data in accordance with the instruction indicated by the prompt, and outputs the inference result in a data format such as voice data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[1686] In the above embodiment, an example was given in which the specific processing is performed by the data processing device 12, but the technology of the present disclosure is not limited to this, and the specific processing may be performed by the robot 414.

[1687] 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.

[1688] FIG. 9 is a diagram illustrating an emotion map 400 on which multiple emotions are mapped. In the emotion map 400, emotions are arranged in concentric circles radiating from the center. Emotions closer to the center of the concentric circles are more primitive. Emotions representing states and actions arising from a state of mind are arranged on the outer edges of the concentric circles. The concept of emotion includes both affect and mental states. Emotions generally generated from reactions occurring in the brain are arranged on the left side of the concentric circles. Emotions generally induced by situational judgment are arranged on the right side of the concentric circles. Emotions generally generated from reactions occurring in the brain and induced by situational judgment are arranged on the upper and lower sides of the concentric circles. Furthermore, the emotion of "pleasure" is arranged on the upper side of the concentric circles, and the emotion of "discomfort" is arranged on the lower side. In this way, in the emotion map 400, multiple emotions are mapped based on the structure by which emotions are generated, and emotions that tend to occur simultaneously are mapped close to each other.

[1689] 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.

[1690] 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).

[1691] Human emotions are based on various balances, such as posture and blood sugar levels. When these balances deviate from the ideal, a state of discomfort is indicated, and when they approach the ideal, a state of pleasure is indicated. Emotions can also be created for robots, automobiles, and motorcycles, based on various balances, such as posture and remaining battery life. When these balances deviate from the ideal, a state of discomfort is indicated, and when they approach the ideal, a state of pleasure is indicated. An emotion map can be generated, for example, based on Dr. Mitsuyoshi's emotion map (Research on Voice Emotion Recognition and Emotional Brain Physiological Signal Analysis Systems, Tokushima University, Doctoral Dissertation: https: / / ci.nii.ac.jp / naid / 500000375379). The left half of the emotion map lists emotions belonging to the "reaction" domain, where sensation is dominant. The right half of the emotion map lists emotions belonging to the "situation" domain, where situational awareness is dominant.

[1692] 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."

[1693] 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.

[1694] The system according to the present disclosure has been described above mainly with respect to the functions of the data processing device 12, but the system according to the present disclosure is not necessarily implemented on a server. The system according to the present disclosure may be implemented as a general information processing system. The present disclosure may be implemented, for example, as a software program running on a personal computer or an application running on a smartphone, etc. The method according to the present disclosure may be provided to users in the form of SaaS (Software as a Service).

[1695] In the above embodiment, an example was given in which the specific processing is performed by one computer 22, but the technology of the present disclosure is not limited to this, and the specific processing may be distributed and performed by a plurality of computers including the computer 22. For example, the data generation model 58 may be provided in an external device of the data processing device 12, and data may be generated in the external device in accordance with input data.

[1696] 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.

[1697] 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.

[1698] 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.

[1699] The hardware resource for executing a specific process can be any of the following processors: An example of a processor is a CPU, which is a general-purpose processor that functions as a hardware resource for executing a specific process by executing software, i.e., a program. Another example of a processor is a dedicated electrical circuit, such as an FPGA (Field-Programmable Gate Array), a PLD (Programmable Logic Device), or an ASIC (Application Specific Integrated Circuit), which is a processor with a circuit configuration designed specifically for executing a specific process. Each processor has built-in or connected memory, and each processor uses the memory to execute the specific process.

[1700] The hardware resource that executes the specific processing may be configured with one of these various processors, or may be configured with a combination of two or more processors of the same or different types (for example, a combination of multiple FPGAs, or a combination of a CPU and an FPGA). Also, the hardware resource that executes the specific processing may be a single processor.

[1701] 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.

[1702] 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.

[1703] 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.

[1704] 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.

[1705] The following is further disclosed regarding the above embodiment.

[1706] (Claim 1)

[1707] A means of monitoring answers and files generated by other AI systems to check for suspicious content; and

[1708] Monitoring referring URLs to detect fraudulent or malicious sites;

[1709] A means for monitoring the generated program code and checking whether it contains malicious programs;

[1710] a means of issuing a warning when malicious content is detected; and

[1711] A system including:

[1712] (Claim 2)

[1713] 10. The system of claim 1, wherein the means for checking for suspicious content analyzes answers or files generated by other artificial intelligence systems using predefined keywords.

[1714] (Claim 3)

[1715] 2. The system of claim 1, wherein the means for monitoring the referenced URL sends an HTTP request to the URL and analyzes the retrieved page content to detect suspicious patterns.

[1716] "Example 1"

[1717] (Claim 1)

[1718] A means of monitoring responses and files generated by other AI systems and analyzing them for suspicious content using predefined keywords; and

[1719] It monitors the referring URL, sends HTTP requests, analyzes the retrieved page content, and detects fraudulent or malicious sites.

[1720] A method for monitoring the generated program code and detecting whether it contains malicious code by checking specific code patterns using regular expressions;

[1721] A means to notify the user of detailed information about detected problems via console logs, email, or GUI messages, and

[1722] A system including:

[1723] (Claim 2)

[1724] 2. The system of claim 1, wherein the suspicious content analysis means analyzes responses or files generated by other artificial intelligence systems using a predefined list of keywords.

[1725] (Claim 3)

[1726] The system of claim 1, wherein the reference URL monitoring means sends an HTTP request to the extracted URL and analyzes the acquired page content by comparing it with patterns for detecting fraudulent or malicious sites.

[1727] "Application Example 1"

[1728] (Claim 1)

[1729] A means of monitoring answers and files generated by other AI systems to check for suspicious content; and

[1730] Monitoring referring URLs to detect fraudulent or malicious sites;

[1731] A means for monitoring the generated program code and checking whether it contains malicious programs;

[1732] a means of issuing a warning when malicious content is detected; and

[1733] A means of monitoring users' digital content in real time, running on smart devices;

[1734] A system including:

[1735] (Claim 2)

[1736] 10. The system of claim 1, wherein the means for checking for suspicious content analyzes answers or files generated by other artificial intelligence systems using predefined keywords.

[1737] (Claim 3)

[1738] 2. The system of claim 1, wherein the means for monitoring the referenced URL sends an HTTP request to the URL and analyzes the retrieved page content to detect suspicious patterns.

[1739] "Example 2: Combining Emotion Engines"

[1740] (Claim 1)

[1741] A means of monitoring answers and files generated by other AI systems to check for suspicious content; and

[1742] A means of monitoring referring links to detect fraudulent or malicious pages;

[1743] A means of monitoring the generated programs to check whether they contain malicious software;

[1744] a means of issuing a warning when malicious content is detected; and

[1745] A means for recognizing the user's emotional state and adjusting the display and content of warning messages;

[1746] A system including:

[1747] (Claim 2)

[1748] 10. The system of claim 1, wherein the means for checking for suspicious content analyzes answers or files generated by other artificial intelligence systems using predefined keywords.

[1749] (Claim 3)

[1750] 10. The system of claim 1, wherein the means for monitoring the referred links sends communication requests to the links and analyzes the retrieved page content to detect suspicious patterns.

[1751] "Application example 2 when combining emotion engines"

[1752] (Claim 1)

[1753] A means of monitoring answers and files generated by other AI systems to check for suspicious content; and

[1754] Monitoring referring URLs to detect fraudulent or malicious sites;

[1755] A means for monitoring the generated program code and checking whether it contains malicious programs;

[1756] a means for issuing a warning based on a user's emotional state when malicious content is detected;

[1757] A system including:

[1758] (Claim 2)

[1759] 10. The system of claim 1, wherein the means for checking for suspicious content analyzes answers or files generated by other artificial intelligence systems using predefined keywords.

[1760] (Claim 3)

[1761] 2. The system of claim 1, wherein the means for monitoring the referenced URL sends an HTTP request to the URL and analyzes the retrieved page content to detect suspicious patterns.

[1762] (Claim 4)

[1763] 10. The system of claim 1, further comprising means for analyzing the user's text to assess their emotional state and means for adjusting the display and content of the warning message based thereon. [Explanation of symbols]

[1764] 10, 210, 310, 410 Data Processing Systems 12 Data Processing Device 14 Smart Devices 214 Smart Glasses 314 Headset-type terminal 414 Robot< / url:> < / url:> < / url:> < / url:>

Claims

1. A means of monitoring answers and files generated by other AI systems to check for suspicious content; and Monitoring referring URLs to detect fraudulent or malicious sites; A means for monitoring the generated program code and checking whether it contains malicious programs; a means of issuing a warning when malicious content is detected; and A system including:

2. 10. The system of claim 1, wherein the means for checking for suspicious content analyzes responses and files generated by other artificial intelligence systems using predefined keywords.

3. 2. The system of claim 1, wherein the means for monitoring the referenced URL sends an HTTP request to the URL and analyzes the retrieved page content to detect suspicious patterns.

Images