System

Abstract

A system is provided.SOLUTION: A system comprising: means for learning past harassment examples using a generative model and monitoring speech and behavior in real time; means for detecting a word or sentence that may be harassment; and means for issuing an alert to notify a user when the word or sentence 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] In modern society, psychological harassment, commonly known as moral harassment (hereinafter referred to as "moral harassment"), has become a serious social problem. Many people often engage in moral harassment without realizing that their own words and actions constitute moral harassment, which leads to a deterioration of the workplace environment and increased mental stress. Existing countermeasures are difficult to prevent moral harassment from occurring, so there is a need for a system that can monitor potentially harassing words and actions in real time and issue an alert to the person making the comment. [Means for solving the problem]

[0005] The present invention provides a system that uses a generative model to learn from past harassment cases and monitors users' behavior in real time. Specifically, the system includes the following means:

[0006] The system includes a means for using a generative model to learn from past harassment cases and monitor speech and behavior in real time, a means for detecting potentially harassing words and sentences, and a means for issuing an alert to notify the user when such words or sentences are detected. It also includes a means for recording the user's voice and converting it into text, and a means for analyzing the converted text data. It also includes a means for monitoring text messages entered by the user in real time and a means for sending the collected text data to a server.

[0007] It includes a means to generate feedback based on the analysis results of the generative model and a means to notify the user of the feedback. It also provides a means to present specific improvement measures to the user and a means to save past detection results as a log.

[0008] This will allow users to be aware that their comments may constitute moral harassment and prevent the situation from escalating, thereby promoting healthy communication.

[0009] A "generative model" is an artificial intelligence model that generates and analyzes new data based on past data and learning.

[0010] "Past cases of harassment" refer to specific cases and data of moral harassment or psychological harassment that occurred in the past.

[0011] "Real-time monitoring" means continuously and immediately monitoring a user's statements and actions without delay.

[0012] "Potentially harassing words or sentences" are words or expressions that constitute or are likely to constitute psychological harassment.

[0013] "User notification alerts" are messages or notification functions that notify users of specific information or warnings.

[0014] "Recording voice and converting it to text" refers to the process of recording a user's speech as voice data and converting that voice data into text information.

[0015] "Convert to text" is the process of converting audio data into written information using speech recognition technology.

[0016] "Analyzing text data" refers to the process of analyzing the converted text information to understand its meaning and content.

[0017] "Generating feedback" means generating evaluations and advice for users' actions and comments.

[0018] "Saving as a log" refers to recording past data or events and saving them for later review.

[0019] "Healthy communication" means mutual respect, no psychological harassment, and smooth communication. [Brief explanation of the drawings]

[0020] [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

[0021] 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.

[0022] First, the terms used in the following description will be explained.

[0023] 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).

[0024] 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.

[0025] 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.

[0026] 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.

[0027] 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."

[0028] [First embodiment]

[0029] FIG. 1 shows an example of the configuration of a data processing system 10 according to the first embodiment.

[0030] 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.

[0031] 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).

[0032] 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.

[0033] 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.

[0034] 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.

[0035] 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.

[0036] FIG. 2 shows an example of the main functions of the data processing device 12 and the smart device 14.

[0037] 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.

[0038] 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.

[0039] 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.

[0040] 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."

[0041] System Configuration

[0042] This invention is a system for preventing moral harassment using generative models, which is mainly composed of three parties: a server, a terminal, and a user. The server analyzes voice and text data collected from the terminal and aims to notify the user of an alert in real time.

[0043] Initial Setup

[0044] server

[0045] Integration and training of generative models: The server uses generative models to learn from past harassment cases, enabling analysis based on new data.

[0046] Database preparation: Prepare a database of past harassment cases and provide it as data to the generative model.

[0047] Terminal

[0048] Installing the app: Install the moral harassment prevention tool app on the user's smartphone or PC. The user must grant the necessary permissions (audio recording, text analysis, etc.).

[0049] Server collaboration: Collaborate with the server to download the generative model and necessary data.

[0050] Real-time monitoring and alert notifications

[0051] Terminal

[0052] Voice recording and text conversion: Record the user's voice in real time using the device's microphone and convert it into text data using voice recognition technology.

[0053] Text data collection: The contents of chat messages and emails entered by users are monitored in real time to collect text data.

[0054] Data transmission: The collected text data is sent to the server.

[0055] server

[0056] Data analysis: The server analyzes the received text data using a generative model to detect words and phrases that may be indicative of moral harassment.

[0057] Feedback generation: Based on the detection results, feedback is generated to notify the user.

[0058] Terminal

[0059] Alert Notification: Notify users of generated feedback in real time using pop-up alerts, notification bars, etc.

[0060] Suggesting improvement measures: Providing specific improvement measures to the user, such as "We recommend you take a break to calm down temporarily."

[0061] Saving logs: Past detection results and notification contents are saved as logs so that users can check them at a later date.

[0062] Specific examples

[0063] Example 1: Speaking during a meeting

[0064] 1. A user says during a meeting, "You're the reason this project failed."

[0065] 2. The device records the audio and converts it into text.

[0066] 3. The server analyzes the converted text using a generative model and determines that the phrase "It's your fault" is likely to be moral harassment.

[0067] 4. The server generates feedback and notifies the user that "Comments that blame others increase the mental burden on the other person and may constitute moral harassment."

[0068] 5. The device displays an alert to the user and suggests specific measures to improve the situation.

[0069] Example 2: Message in chat

[0070] 1. A user sends a message in chat saying, "Your mistake is causing trouble for everyone."

[0071] 2. The device monitors and collects chat content in real time.

[0072] 3. The server analyzes the collected text using a generative model and determines that the phrase "Everyone is in trouble because of your mistake" is a possible example of moral harassment.

[0073] 4. The server generates feedback stating that "attacks aimed at identifying and blaming individuals may constitute moral harassment."

[0074] 5. The device displays an alert to the user and a message encouraging them to stay calm.

[0075] This system can detect in advance whether comments or messages may constitute moral harassment, and by encouraging users to be aware of this, it can contribute to preventing moral harassment.

[0076] The processing flow will be explained below.

[0077] Step 1:

[0078] Users install the moral harassment prevention tool app on their smartphone or PC and grant the necessary permissions (audio recording, text analysis, etc.).

[0079] Step 2:

[0080] The device works in conjunction with the server to download the generative model and necessary data.

[0081] Step 3:

[0082] The device uses a microphone to record the user's voice in real time and converts it into text data using voice recognition technology.

[0083] Step 4:

[0084] The terminal monitors the contents of chat messages and emails entered by users in real time and collects text data.

[0085] Step 5:

[0086] The terminal transmits the collected text data (including the text converted from the speech) to the server.

[0087] Step 6:

[0088] The server analyzes the received text data using a generative model to detect words and phrases that may be indicative of moral harassment.

[0089] Step 7:

[0090] The server generates feedback to notify the user based on the detection results.

[0091] Step 8:

[0092] The device notifies the user of the feedback received from the server in real time using methods such as pop-up alerts and a notification bar.

[0093] Step 9:

[0094] The device will then suggest specific measures to improve the situation to the user, such as "We recommend you take a break to calm down temporarily."

[0095] Step 10:

[0096] The device stores past detection results and notification contents as a log so that the user can check them at a later date.

[0097] Example 1

[0098] 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."

[0099] In modern society, moral harassment (morahara) in the workplace and at home has become a serious problem. Morahara not only damages people's mental health and reduces productivity, but in severe cases can even lead to legal issues. Conventional systems and tools have made it difficult to detect and prevent moral harassment, and real-time monitoring and notification have rarely been realized. This has made it difficult to respond quickly and appropriately when moral harassment occurs.

[0100] 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.

[0101] In this invention, the server includes means for learning past cases of harassment using a generative model and monitoring speech and behavior in real time, means for detecting potentially harassing expressions, means for issuing an alert to notify the user when such expressions are detected, means for collecting messages and voice data entered by the user, means for transmitting the collected data to the server, means for analyzing the data received by the server and generating feedback, means for notifying the user of the generated feedback, and means for recording and saving past detection results and notification content. This makes it possible to detect possible moral harassment in real time and promptly notify the user.

[0102] A "generative model" is a type of artificial intelligence that learns from past data and analyzes and generates new data.

[0103] "Means for monitoring speech and behavior" refers to means for detecting user statements and behavior in real time and recording or analyzing the content.

[0104] "Means for detecting expressions" refers to means for detecting specific words or phrases from audio or text data and evaluating whether there is a problem.

[0105] "Means for issuing alerts" are means for notifying or warning users about detected problematic behavior or inappropriate comments.

[0106] "Means for collecting message and voice data" means means for recording text messages and voice input by users in real time and retaining them for subsequent processing.

[0107] The "means for transmitting data to a server" refers to a means for transmitting data collected by a terminal to a server via a network.

[0108] "Means for analyzing data" refers to means for analyzing received data in detail and detecting specific patterns or anomalies.

[0109] The "means for generating feedback" refers to the means for generating and notifying the user of countermeasures or warnings based on the analysis results.

[0110] "Means for notifying the user of feedback" refers to means for displaying or notifying the user of the generated feedback in real time.

[0111] "Means for recording and saving" refers to a means for saving past detection results and notification contents in a database or the like so that they can be referenced at a later date.

[0112] This invention is a moral harassment prevention system that uses a generative model. This system is mainly composed of three entities: a server, a terminal, and a user. A specific embodiment of the system will be described below.

[0113] Hardware and Software Configuration

[0114] server

[0115] The server is constructed using the following hardware and software:

[0116] Hardware: A server machine equipped with a high-performance CPU, GPU, and large memory capacity

[0117] Software: Python programming language, TensorFlow or PyTorch, MySQL or PostgreSQL database

[0118] Terminal

[0119] The terminal is a communication device such as a user's smartphone or PC.

[0120] Hardware: Smartphones (iOS, Android), PCs (Windows, macOS)

[0121] Software: Moral harassment prevention tool app, Google Speech-to-Text API, REST API client

[0122] System configuration

[0123] The system is realized by combining the following means:

[0124] 1. Generative Model Integration and Training

[0125] The server uses a generative model (e.g., GPT-4) to learn from past harassment cases, enabling real-time analysis based on new data.

[0126] 2. Prepare the database

[0127] The server prepares a database of past harassment cases and supplies it to the generative model as data. The database is MySQL or PostgreSQL.

[0128] 3. Install the app

[0129] Install a moral harassment prevention tool app on the user's smartphone or PC. The app is compatible with iOS and Android, and the user must grant the necessary permissions (voice recording, text analysis, etc.).

[0130] 4. Server Integration

[0131] The device connects to the server and downloads the generative model and necessary data. The data is obtained using a REST API.

[0132] Examples of concrete examples and prompts

[0133] Example 1: Statements made during a meeting

[0134] 1. A user says during a meeting, "You're the reason this project failed."

[0135] 2. The device records the audio and converts it to text using the Google Speech-to-Text API.

[0136] 3. The terminal sends the converted text to the server.

[0137] 4. The server analyzes the received text using the GPT-4 generative model and determines that the phrase "It's your fault" is likely to be moral harassment.

[0138] 5. The server generates feedback and notifies the user that the comment may be emotionally abusive.

[0139] 6. The device displays an alert to the user saying, "A statement that may be moral harassment has been detected."

[0140] Example 2: Chat messages

[0141] 1. A user sends a message in chat saying, "Your mistake is causing trouble for everyone."

[0142] 2. The device monitors and collects chat content in real time.

[0143] 3. The device sends the collected text to the server.

[0144] 4. The server analyzes the received text using the BERT generative model and determines that the phrase "Everyone is suffering because of your mistake" is a possible example of moral harassment.

[0145] 5. The server generates feedback stating that "attacks aimed at identifying and blaming individuals may constitute moral harassment."

[0146] 6. The device displays an alert to the user saying, "A statement that may be moral harassment has been detected."

[0147] Examples of prompt statements

[0148] "Respond to harassing phrases"

[0149] "Please determine whether this sentence constitutes moral harassment."

[0150] This invention can contribute to preventing moral harassment by detecting in real time whether a user's comments or messages may constitute moral harassment and quickly notifying the user.

[0151] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0152] Program processing flow

[0153] Step 1: Collecting audio data

[0154] Input: What the user says

[0155] Output: Collected audio data

[0156] Specific description:

[0157] The device collects the user's voice in real time through a microphone, specifically the built-in microphone of a smartphone or PC.

[0158] The app will launch the audio collection module and notify the user that "Audio recording has started."

[0159] Step 2: Convert audio data to text

[0160] Input: Collected audio data

[0161] Output: Converted text data

[0162] Specific description:

[0163] The device sends the collected voice data to the Google Speech-to-Text API, which converts the voice data into text.

[0164] Once the conversion is complete, the user is notified that "the audio data has been converted to text."

[0165] Step 3: Monitor and collect chat messages

[0166] Input: The chat message typed by the user

[0167] Output: Collected text messages

[0168] Specific description:

[0169] The device monitors the user's chat app (e.g., Slack or WhatsApp) and collects messages sent in real time.

[0170] Specifically, messages are retrieved using the chat app's API, and are temporarily stored on the device.

[0171] Step 4: Sending text data

[0172] Input: Collected text data

[0173] Output: Data sent to the server

[0174] Specific description:

[0175] The terminal transmits the collected text data to a server.

[0176] When sending, the data is AES encrypted to make it secure, and the user is notified that the data has been sent to the server.

[0177] Step 5: Data analysis

[0178] Input: Transmitted text data

[0179] Output: Analysis results

[0180] Specific description:

[0181] The text data received by the server is analyzed by a generative model (GPT-4 or BERT).

[0182] Specifically, to detect inappropriate words and phrases, prompts such as "Please respond to harassing phrases" are used.

[0183] When the processing is completed on the server, it records "Data analysis completed" in the log.

[0184] Step 6: Feedback generation

[0185] Input: Analysis results

[0186] Output: Generated feedback

[0187] Specific description:

[0188] The server generates feedback based on the analysis results of the generative model.

[0189] Specifically, it generates a notification message such as, "That comment may be emotional harassment. We recommend that you take a break to calm down."

[0190] Log "Feedback generated."

[0191] Step 7: Alert Notifications

[0192] Input: Generated feedback

[0193] Output: Alert notification to the user

[0194] Specific description:

[0195] The terminal notifies the user of the feedback received from the server in real time.

[0196] A pop-up alert or notification bar will be displayed to the user, informing them that "statements that may be emotionally abusive have been detected."

[0197] When a user acknowledges a notification, "User acknowledged notification" is logged.

[0198] Step 8: Propose improvement measures

[0199] Input: Generated feedback

[0200] Output: Present specific improvement measures to the user

[0201] Specific description:

[0202] The device will present specific improvement measures to the user.

[0203] For example, it will display advice such as, "We recommend you take a short walk to calm down."

[0204] When a user confirms a remediation, "User confirmed remediation" is logged.

[0205] Step 9: Save the logs

[0206] Input: Detection results and notification content

[0207] Output: Saved log data

[0208] Specific description:

[0209] The device stores past detection results and notification contents in an SQLite database.

[0210] After saving, you will be notified that "The log has been saved to the database."

[0211] Users can check this log at any time.

[0212] (Application example 1)

[0213] 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."

[0214] There is a need to prevent harassment that occurs in communication between employees in commercial facilities and improve the working environment. In particular, a system that allows employees to review their own behavior and take appropriate action through real-time monitoring and immediate feedback is needed, but this type of system has not yet been fully developed.

[0215] 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.

[0216] In this invention, the server includes means for learning from past cases of harassment using a generative model and monitoring speech and behavior in real time, means for detecting potentially harassing words and sentences, means for issuing an alert to notify users when such words or sentences are detected, means for recording the voices of employees at commercial facilities and converting them into text data, means for analyzing the converted text data, and means for generating feedback and notifying alerts and remedial measures. This makes it possible to immediately detect harassing behavior in real-time communication between employees and propose remedial measures, thereby ensuring a safe and comfortable working environment.

[0217] A "generative model" is a machine learning model that can generate new data based on past data.

[0218] "Harassment cases" refer to cases of conduct that cause psychological or physical distress to another person.

[0219] "Real-time monitoring means" are methods for instantly observing and recording current data and behavior.

[0220] An "alert mechanism" is a mechanism that sends a warning or notification to the user when certain conditions are met.

[0221] A "commercial facility" is a business establishment that has the facilities and space to offer goods and services.

[0222] "Means for recording employee voice and converting it into text data" refers to technology that records employee speech using a microphone or other device and converts that voice data into text information.

[0223] "Means for analyzing text data" refers to techniques for evaluating and analyzing textual information and extracting specific patterns and meanings.

[0224] "Means for generating feedback and notifying alerts and improvement measures" refers to methods for generating warnings and advice to users based on the analysis results and immediately communicating them.

[0225] This invention is a system for improving communication between employees in commercial facilities and preventing harassment. This system is composed of three entities: a server, a terminal, and a user.

[0226] server

[0227] The server uses a generative model to learn from past harassment cases and analyze new data. This allows it to analyze incoming data in real time and detect potentially abusive words and phrases. The server also generates feedback based on the analysis results and sends alerts and recommendations to the device.

[0228] Specifically, the server uses the following software and hardware:

[0229] Hardware: High-performance processor, memory, and large storage capacity

[0230] Software: Python, TensorFlow, Keras, SpeechRecognition library, Requests library

[0231] Terminal

[0232] The devices include the employee's smartphone or smart glasses. The devices have the ability to record voice data and convert it into text. The converted text data is then sent to the server in real time, and the received feedback is notified to the user.

[0233] The specific operation method is as follows:

[0234] 1. Record your voice using a microphone

[0235] 2. Convert the recorded audio into text using the SpeechRecognition library

[0236] 3. Send the converted text data to the server using the Requests library

[0237] 4. Receive feedback from the server and display it to the user

[0238] User

[0239] Users are primarily employees of commercial facilities, and their role is to record voice in real time, convert it into text data, and send it to a server. They also receive feedback from the server and correct their behavior.

[0240] Specific examples

[0241] Example of audio data

[0242] If an employee says something like "Did you make another mistake?", it will be handled as follows:

[0243] 1. Record audio using the device's microphone

[0244] 2. Convert recorded audio to text

[0245] 3. Send text data to the server

[0246] 4. The server analyzes and detects keywords such as "mistake."

[0247] 5. The server generates feedback and sends it to the device.

[0248] 6. The device will send an alert saying, "Comments that emphasize personal mistakes may lead to moral harassment."

[0249] 7. As an improvement measure, the message reads, "Work with other staff to identify problems and come up with solutions together."

[0250] Prompt Sentence Examples

[0251] "User sent text: 'Did you make another mistake?' Evaluate potential abuse and provide appropriate feedback and remediation."

[0252] In this way, this invention can prevent harassment between employees in commercial facilities. By providing real-time feedback, employees can immediately correct their behavior and promote healthy communication.

[0253] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0254] Step 1:

[0255] The user records their voice using the microphone on their device. The input is the user's speech, and the output is voice data. Specifically, the user's words are recorded while they are talking to other employees in a commercial facility.

[0256] Step 2:

[0257] The device converts the recorded voice data into text. The input is voice data, and the output is text data that converts the voice into text information. Specifically, the voice-to-text conversion process is performed using the SpeechRecognition library.

[0258] Step 3:

[0259] The terminal sends text data to the server. The input is text data, and the output is an HTTP request to the server. Specifically, the Requests library is used to send text data to the server via a POST request.

[0260] Step 4:

[0261] The server analyzes the received text data. The input is the text data, and the output is the analysis result. Specifically, it uses a generative AI model to analyze the text data and detect words and phrases that may be abusive or harassing.

[0262] Step 5:

[0263] The server generates feedback based on the analysis results. The input is the analysis results, and the output is a feedback message. Specifically, based on the analysis results provided by the generative AI model, it creates feedback for the user that includes appropriate alerts and improvement measures.

[0264] Step 6:

[0265] The server sends the generated feedback to the terminal. The input is a feedback message, and the output is an HTTP response to the terminal. Specifically, data including the feedback message is sent to the terminal as an HTTP response.

[0266] Step 7:

[0267] The terminal notifies the user of the received feedback. The input is a feedback message, and the output is an alert or notification to the user. Specifically, the terminal displays the feedback message as a pop-up on the display or in the notification bar.

[0268] Step 8:

[0269] The user checks the notification and corrects their behavior. The input is the feedback message, and the output is the user's improved behavior. Specifically, the user understands the alert and the improvement measures notified to them, and corrects their statements and behavior based on them.

[0270] This allows real-time communication between employees within a commercial facility to be monitored, preventing harassment before it occurs.

[0271] 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.

[0272] System Configuration

[0273] This invention is a system for preventing moral harassment that uses a generative model and an emotion engine, and is primarily composed of three parties: a server, a device, and a user. The server analyzes voice and text data collected from the device and aims to notify the user of an alert in real time. In addition, by combining this with an emotion engine, the system can analyze the user's emotions and provide more appropriate alerts and feedback.

[0274] Initial Setup

[0275] server

[0276] Integration and training of generative models: The server uses generative models to learn from past harassment cases, enabling analysis based on new data.

[0277] Emotion engine integration: An emotion engine is integrated into the server to recognize emotions from user voice and text data.

[0278] Database preparation: Prepare a database of past harassment cases and provide it as data to the generative model.

[0279] Terminal

[0280] Installing the app: Install the moral harassment prevention tool app on the user's smartphone or PC. The user must grant the necessary permissions (audio recording, text analysis, etc.).

[0281] Server collaboration: Collaborate with the server to download the generative model, emotion engine, and necessary data.

[0282] Real-time monitoring and alert notifications

[0283] Terminal

[0284] Voice recording and text conversion: Record the user's voice in real time using the device's microphone and convert it into text data using voice recognition technology.

[0285] Text data collection: The contents of chat messages and emails entered by users are monitored in real time to collect text data.

[0286] Data transmission: The collected text data (including the text converted from the speech) is sent to the server.

[0287] server

[0288] Data analysis: The server analyzes the received text data using a generative model to detect words and phrases that may be indicative of moral harassment.

[0289] Emotion analysis: Using an emotion engine, we recognize emotions from the user's voice and text data, such as anger and sadness.

[0290] Feedback generation: Based on the detection results and sentiment analysis results, feedback is generated to notify the user.

[0291] Terminal

[0292] Alert notification: The generated feedback is notified to the user in real time using methods such as pop-up alerts and notification bars.

[0293] Suggesting improvement measures: Presents specific improvement measures to the user. For example, based on the results of sentiment analysis, the system displays advice such as, "What you just said may be perceived as offensive to the other person. We recommend that you take some time to calm down."

[0294] Saving logs: Past detection results and notification contents are saved as logs so that users can check them at a later date.

[0295] Specific examples

[0296] Example 1: Speaking during a meeting

[0297] 1. A user says during a meeting, "You're the reason this project failed."

[0298] 2. The device records the audio and converts it into text.

[0299] 3. The server analyzes the converted text using a generative model and determines that the phrase "It's your fault" is likely to be moral harassment.

[0300] 4. The server uses the emotion engine to recognize the user's emotion, for example, recognizing that the user has an emotion of anger.

[0301] 5. The server generates feedback and informs the user, "Speaking in anger may come across as aggressive towards the other person. We recommend taking some time to calm down."

[0302] 6. The device displays an alert to the user and suggests specific measures to improve the situation.

[0303] Example 2: Message in chat

[0304] 1. A user sends a message in chat saying, "Your mistake is causing trouble for everyone."

[0305] 2. The device monitors and collects chat content in real time.

[0306] 3. The server analyzes the collected text using a generative model and determines that the phrase "Everyone is in trouble because of your mistake" is a possible example of moral harassment.

[0307] 4. The server uses the emotion engine to recognize the user's emotion, for example, recognizing that the user is dissatisfied.

[0308] 5. The server generates feedback and informs the user, "Sending this message out of frustration may come across as blaming. We recommend that you send the message again after you have calmed down a bit."

[0309] 6. The device will alert the user and provide specific advice to encourage them to stay calm.

[0310] This system can further contribute to preventing moral harassment by detecting in advance comments or messages that may constitute moral harassment and providing feedback and improvement measures that take into account the user's emotional state.

[0311] The processing flow will be explained below.

[0312] System Configuration

[0313] This invention is a system for preventing moral harassment that uses a generative model and an emotion engine, and is primarily composed of three parties: a server, a device, and a user. The server analyzes voice and text data collected from the device and aims to notify the user of an alert in real time. In addition, by combining this with an emotion engine, the system can recognize the user's emotions and provide more appropriate alerts and feedback.

[0314] Initial Setup

[0315] Step 1:

[0316] The server uses a generative model to learn from past harassment cases, allowing it to analyze new data.

[0317] Step 2:

[0318] The server integrates an emotion engine to recognize emotions from the user's voice and text data.

[0319] Step 3:

[0320] The server prepares a database of past harassment cases and supplies it as data to the generative model.

[0321] Step 4:

[0322] The device involves installing a moral harassment prevention tool app on the user's smartphone or PC, and the user granting the necessary permissions (voice recording, text analysis, etc.).

[0323] Step 5:

[0324] The device connects to the server and downloads the generative model, emotion engine, and necessary data.

[0325] Real-time monitoring and alert notifications

[0326] Step 6:

[0327] The device uses a microphone to record the user's voice in real time and converts it into text data using voice recognition technology.

[0328] Step 7:

[0329] The terminal monitors the contents of chat messages and emails entered by users in real time and collects text data.

[0330] Step 8:

[0331] The terminal transmits the collected text data (including the text converted from the speech) to the server.

[0332] Step 9:

[0333] The server analyzes the received text data using a generative model to detect words and phrases that may be indicative of moral harassment.

[0334] Step 10:

[0335] The server uses an emotion engine to recognize emotions from the user's voice and text data, such as anger and sadness.

[0336] Step 11:

[0337] The server generates feedback to notify the user based on the analysis results of the generative model and the emotion analysis results.

[0338] Step 12:

[0339] The device notifies the user of the feedback received from the server in real time using methods such as pop-up alerts and a notification bar.

[0340] Step 13:

[0341] The device will then present the user with specific advice on how to improve the situation, such as "What you just said may be perceived as offensive to the other person. We recommend that you take a moment to calm down."

[0342] Step 14:

[0343] The device stores past detection results and notification contents as a log so that the user can check them at a later date.

[0344] Specific examples

[0345] Example 1: Speaking during a meeting

[0346] Step 1:

[0347] A user says during a meeting, "You are the reason this project failed."

[0348] Step 2:

[0349] The device will record the audio and convert it into text.

[0350] Step 3:

[0351] The server analyzes the converted text using a generative model and determines that the phrase "It's your fault" is likely to be moral harassment.

[0352] Step 4:

[0353] The server uses an emotion engine to recognize the user's emotion, for example, recognizing that the user has an emotion of anger.

[0354] Step 5:

[0355] The server generates feedback, informing the user that "Speaking in anger may come across as aggressive towards the other person. We recommend that you take some time to calm down."

[0356] Step 6:

[0357] The device will display an alert to the user and suggest specific remedial measures.

[0358] Example 2: Message in chat

[0359] Step 1:

[0360] A user sends a message in chat saying, "Your mistake is causing trouble for everyone."

[0361] Step 2:

[0362] The device monitors and collects chat content in real time.

[0363] Step 3:

[0364] The server analyzes the collected text using a generative model and determines that the phrase "Everyone is in trouble because of your mistake" may be an example of moral harassment.

[0365] Step 4:

[0366] The server uses an emotion engine to recognize the user's emotion, for example, recognizing that the user is dissatisfied.

[0367] Step 5:

[0368] The server generates feedback, informing the user, "Sending this message out of frustration may come across as blaming. We recommend that you send the message again after you have calmed down a bit."

[0369] Step 6:

[0370] The device will display an alert to the user and provide specific advice to encourage them to stay calm.

[0371] In this way, a system that combines a generative model and an emotion engine can detect in advance whether a statement or message may constitute moral harassment, encourage the user to be aware of it, and provide appropriate feedback and improvement measures tailored to the user's emotional state.

[0372] Example 2

[0373] 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."

[0374] In modern society, moral harassment (moral harassment) is becoming a serious problem in various settings, including the workplace and school. This not only causes victims psychological damage, stress, and anxiety, but also reduces their productivity and quality of life. While existing moral harassment prevention systems can address certain cases, they are not considered effective preventative measures due to insufficient real-time monitoring, analysis, and feedback. Therefore, there is a need for a more advanced and comprehensive moral harassment prevention system that can reduce stress in users' daily lives and work and improve their mental health.

[0375] The specific processing by the specific processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means.

[0376] In this invention, the server includes means for learning from past harassment cases using a generative model and monitoring speech and behavior in real time, means for detecting potentially harassing words and sentences, means for issuing an alert to notify the user when such words or sentences are detected, means for integrating an emotion engine that analyzes the user's emotions, and means for generating feedback based on the emotion analysis results. This makes it possible to detect in real time whether a statement or message may constitute moral harassment and provide appropriate feedback and improvement measures.

[0377] A "generative model" is a type of artificial intelligence that learns from past data and generates appropriate outputs for new data.

[0378] "Harassment cases" are specific cases of moral harassment or psychological aggression that have occurred in the past.

[0379] "Means of real-time monitoring" refers to technology that allows for immediate monitoring of users' behavior and input.

[0380] "Word and phrase detection methods" are technologies that identify specific signs of harassment in users' statements and messages.

[0381] An "alert mechanism" is a mechanism for notifying users of detected potential harassment.

[0382] An "emotion engine" is a system for analyzing user emotions from voice and text data.

[0383] "Means for generating feedback" refers to technology that creates advice for users based on the results of detected harassment and sentiment analysis.

[0384] This invention is a system for preventing moral harassment that uses a generative model and an emotion engine to monitor users' behavior in real time and provide appropriate feedback. This system is mainly composed of three parties: a server, a terminal, and a user.

[0385] System Configuration

[0386] server

[0387] 1. Generative model integration and training:

[0388] The server uses a generative model to learn from past harassment cases. The specific generative model used is a "generative AI model," a general language model. This model learns patterns from past case data and can be applied to new situations.

[0389] 2. Emotion engine integration:

[0390] The server will be integrated with Emotion Engine, a general-purpose emotion analysis software, which will enable real-time analysis of emotions from user voice and text data, enabling more appropriate alerts and feedback to be provided.

[0391] 3. Prepare the database:

[0392] The server prepares a database that accumulates past cases of harassment. This database contains text data and emotion tags, and is used to analyze cases of moral harassment. The specific database management system used is MongoDB or MySQL.

[0393] Terminal

[0394] 1. Install the app:

[0395] Users install the anti-moral harassment app on their smartphones or PCs. The app asks the user for permission to record voice and monitor chat logs, and once permission is granted, the system will function normally.

[0396] 2. Server integration:

[0397] The device connects to the server to download data for the generative model and emotion engine, using secure protocols such as HTTPS for communication.

[0398] Specific examples

[0399] Remarks made during the meeting

[0400] 1. A user says during a meeting, "You're the reason this project failed."

[0401] 2. The device records the audio and converts it into text. The speech recognition technology uses the Google Cloud Speech-to-Text API.

[0402] 3. The server analyzes the converted text using a generative model and determines that the phrase "It's your fault" is likely to be moral harassment.

[0403] 4. The server uses an emotion engine to recognize the user's emotion. For example, it determines that the user is angry.

[0404] 5. The server generates feedback and displays an alert saying, "Speaking in anger may come across as aggressive towards the other person. We recommend taking some time to calm down."

[0405] 6. The device provides generated feedback to the user as an alert notification, either via a pop-up alert or a notification bar.

[0406] Messages in chat

[0407] 1. A user sends a message in chat saying, "Your mistake is causing trouble for everyone."

[0408] 2. The device monitors chat content in real time and collects text data.

[0409] 3. The server analyzes the collected text using a generative model and determines that the phrase "Everyone is in trouble because of your mistake" is a possible example of moral harassment.

[0410] 4. The server uses an emotion engine to recognize the user's emotion. For example, it determines that the user is dissatisfied.

[0411] 5. The server generates feedback and displays an alert saying, "Sending this message out of frustration may come across as blaming. We recommend you send the message again after you have calmed down a bit."

[0412] 6. The device provides the generated feedback to the user as an alert notification, using the chat app's notification function.

[0413] This system not only prevents moral harassment in real time, but also contributes to reducing mental stress by providing appropriate feedback that takes into account the user's emotional state.The system also monitors users' comments and messages and provides immediate feedback, enabling early detection and resolution of problems.

[0414] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0415] Step 1:

[0416] The server integrates a generative model and learns from past harassment cases.

[0417] As input, the generative model is fed with historical harassment case data, which is in text format and contains a large number of examples.

[0418] As part of data processing and data calculation, a generative model (e.g., a generative AI model) analyzes this text data and learns patterns of harassment. Specifically, the model processes the text data and performs a learning process to adjust weights and biases.

[0419] The output is a trained generative model.

[0420] Step 2:

[0421] The server integrates the emotion engine.

[0422] As input, you install the software for sentiment analysis on the server and enter the necessary API keys and settings.

[0423] As part of data processing and data calculation, the emotion engine (e.g., emotion analysis software) performs the initial setup and integration for analyzing emotions from voice and text data. Specific operations include configuring APIs and installing emotion analysis models.

[0424] As an output, we obtain a server with an integrated emotion engine.

[0425] Step 3:

[0426] The device installs the application and performs initial setup.

[0427] The input is to download and install the moral harassment prevention tool application onto your smartphone or PC.

[0428] For data processing and data calculation, the application obtains permission from the user to access voice recordings and chat logs. Specifically, the user operates the application installation screen and grants the necessary permissions.

[0429] The output will be an application that has been successfully installed and has the necessary permissions.

[0430] Step 4:

[0431] The device collects the voice data and converts it into text.

[0432] As input, the user's speech is recorded in real time through the device's microphone.

[0433] Data processing and calculation involves converting recorded voice data into text data using the Google Cloud Speech-to-Text API, etc. Specific operations include sending the voice data to a cloud server and converting it into text.

[0434] The output is the text data converted from the speech.

[0435] Step 5:

[0436] The device collects the text data entered and sends it to the server.

[0437] As input, the user enters a chat message or email.

[0438] For data processing and data calculation, the device monitors the user's input and collects text data in real time, which is temporarily stored in the device's memory.

[0439] As an output, we get the collected text data, which is then sent to a server using the HTTPS protocol.

[0440] Step 6:

[0441] The server analyzes the received text data.

[0442] As input, it receives text data sent from the terminal.

[0443] As part of the data processing and data calculation process, a generative model is used to analyze text data and detect words and phrases that may be related to moral harassment. Specifically, the model analyzes the text data and detects specific patterns.

[0444] The output is a list of phrases and words that are judged to be likely to be moral harassment.

[0445] Step 7:

[0446] The server performs sentiment analysis.

[0447] It takes parsed text data and audio data as input.

[0448] For data processing and data calculation, an emotion engine is used to analyze user emotions, including emotions such as anger and sadness. Specifically, an emotion analysis model is executed to extract emotions from text and voice data.

[0449] As an output, information about the user's emotional state is obtained.

[0450] Step 8:

[0451] The server generates the feedback.

[0452] As input, it receives the results of moral harassment detection and emotion analysis.

[0453] Based on the results of data processing and calculation, feedback is generated to notify the user. Specifically, appropriate advice or warning messages are created using specific algorithms and templates.

[0454] As an output, we get the generated feedback message.

[0455] Step 9:

[0456] The device will alert the user and suggest remedial measures.

[0457] As input, it receives the feedback message sent by the server.

[0458] As data processing and calculations are performed, alert notifications are displayed to the user. Notification methods include pop-up alerts and notification bars. Specific operations include using the device's notification function to provide real-time feedback to the user.

[0459] The output is an alert notification and remediation provided to the user.

[0460] Step 10:

[0461] The device stores the log.

[0462] Past detection results and notification contents are collected as input.

[0463] This data is then processed and calculated and stored in a database (e.g., SQLite) within the device. Specifically, the data is added to the database and made accessible at a later date.

[0464] As an output, you get the saved log data.

[0465] (Application example 2)

[0466] 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."

[0467] In physical stores, there is a need to prevent problems caused by inappropriate comments or emotional interactions during communication between staff and customers. However, conventional methods have difficulty analyzing comments and emotions in real time, making it difficult to provide appropriate feedback quickly. To solve this problem, a system is needed that can monitor and analyze the behavior of store staff in real time and provide appropriate feedback instantly.

[0468] The identification process 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 learning past harassment cases using a generative model and monitoring speech and behavior in real time, means for detecting words and sentences that may be harassing, means for issuing an alert to notify the user when such words or sentences are detected, means for analyzing emotions from voice and text data using an emotion engine, and means for providing appropriate remedial measures based on the analysis results. This makes it possible to improve the quality of communication between staff and customers in physical stores and prevent problems from occurring.

[0469] A "generative model" is an artificial intelligence technology that learns from past cases of harassment and analyzes the possibility of harassment in new data.

[0470] A "harassment case" is one in which inappropriate words, actions or behaviour cause mental distress or discomfort to a specific individual.

[0471] The "emotion engine" is a system for analyzing users' emotions from voice and text data.

[0472] "Real-time monitoring measures" are technologies for instantly collecting and analyzing ongoing audio and text data.

[0473] "Means of issuing an alert" refers to a method of notifying users when potential harassment is detected.

[0474] "Means for recording voice and converting it into text" refers to technology for recording speech and converting it into text data.

[0475] "Means of analysis" are methods for analyzing collected data to detect specific patterns or risks.

[0476] "Means for providing improvements" refers to methods for providing appropriate advice and action suggestions to users based on the analysis results.

[0477] System Configuration

[0478] This invention is a moral harassment prevention system that uses a generative model and an emotion engine, and is primarily composed of three parties: a server, a terminal, and a user. The server analyzes voice and text data collected from the terminal and aims to notify the user of an alert in real time. In addition, by combining it with an emotion engine, the system can analyze the user's emotions and provide more appropriate alerts and feedback.

[0479] Initial Setup

[0480] server

[0481] Generative model integration and training: The server uses a generative model to learn from past harassment cases, enabling analysis based on new data.

[0482] Emotion engine integration: Integrate an emotion engine into the server to recognize emotions from user voice and text data.

[0483] Database preparation: Prepare a database of past harassment cases and feed it to the generative model as data.

[0484] Terminal

[0485] Installing the app: Install the moral harassment prevention tool app on the user's smartphone or PC. The user must grant the necessary permissions (audio recording, text analysis, etc.).

[0486] Server collaboration: Collaborate with the server to download the generative model, emotion engine, and necessary data.

[0487] Real-time monitoring and alert notifications

[0488] Terminal

[0489] Voice recording and text conversion: Record the user's voice in real time using the device's microphone and convert it into text data using voice recognition technology.

[0490] Text data collection: The content of chat messages and emails entered by users is monitored in real time to collect text data.

[0491] Data transmission: Collected text data (including text converted from speech) is sent to the server.

[0492] server

[0493] Data analysis: The server analyzes the received text data using a generative model to detect words and phrases that may be moral harassment.

[0494] Emotion analysis: Using an emotion engine, we recognize emotions from the user's voice and text data, such as anger and sadness.

[0495] Feedback generation: Based on the detection results and sentiment analysis results, feedback is generated to notify the user.

[0496] Terminal

[0497] Alert notifications: Notify users of generated feedback in real time using methods such as pop-up alerts or notification bars.

[0498] Suggesting improvement measures: Presents specific improvement measures to the user. For example, based on the results of sentiment analysis, the system displays advice such as, "What you just said may be perceived as offensive to the other person. We recommend that you take a moment to calm down."

[0499] Saving logs: Past detection results and notification contents are saved as logs so that users can check them at a later date.

[0500] Specific examples

[0501] Example 1: Statements made during a meeting

[0502] 1. A user says during a meeting, "You're the reason this project failed."

[0503] 2. The device records the audio and converts it into text.

[0504] 3. The server analyzes the converted text using a generative model and determines that the phrase "It's your fault" is likely to be moral harassment.

[0505] 4. The server uses the emotion engine to recognize the user's emotion, for example, recognizing that the user has an emotion of anger.

[0506] 5. The server generates feedback and informs the user, "Speaking in anger may come across as aggressive towards the other person. We recommend taking some time to calm down."

[0507] 6. The device displays an alert to the user and suggests specific measures to improve the situation.

[0508] Example 2: Message in chat

[0509] 1. A user sends a message in chat saying, "Your mistake is causing trouble for everyone."

[0510] 2. The device monitors and collects chat content in real time.

[0511] 3. The server analyzes the collected text using a generative model and determines that the phrase "Everyone is in trouble because of your mistake" is a possible example of moral harassment.

[0512] 4. The server uses an emotion engine to recognize the user's emotion, for example, recognizing that the user is dissatisfied.

[0513] 5. The server generates feedback and informs the user, "Sending this message out of frustration may come across as blaming. We recommend that you send the message again after you have calmed down a bit."

[0514] 6. The device will alert the user and provide specific advice to encourage them to stay calm.

[0515] Prompt Sentence Examples

[0516] "Does this statement take into consideration your customer's feelings? I suggest you try phrasing it differently."

[0517] "The emotion recognition engine detects negative emotions. Please try to use as polite language as possible."

[0518] In this way, the system can improve the quality of communication between staff and customers in physical stores.

[0519] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[0520] Step 1:

[0521] The device records the user's voice in real time. Specifically, it uses the device's microphone to capture the user's speech and obtains the voice data. This voice data becomes the input.

[0522] Step 2:

[0523] The device converts the voice data it receives into text. It uses voice recognition technology to analyze the voice data and generate corresponding text data. This text data is then output.

[0524] Step 3:

[0525] The device sends the converted text data to the server. Specifically, the device uploads the text data to the server via the Internet. This text data becomes the input.

[0526] Step 4:

[0527] The server analyzes the received text data using a generative model. The generative model has learned from past harassment cases and analyzes new text data to detect potentially harassing words and phrases. The results of this analysis are output.

[0528] Step 5:

[0529] The server uses an emotion engine to perform emotion analysis based on the analysis results. Specifically, the server analyzes the user's emotions (anger, dissatisfaction, etc.) from the text data. The analysis results are output.

[0530] Step 6:

[0531] The server generates feedback based on the analysis results and sentiment analysis results. The generative model and sentiment analysis results are combined to generate appropriate advice or warnings for the user. This feedback is the output.

[0532] Step 7:

[0533] The server sends the generated feedback to the terminal. Specifically, the server sends the generated feedback to the terminal via the Internet. This feedback becomes the input.

[0534] Step 8:

[0535] The device displays an alert to the user. Specifically, the feedback content is displayed on the device display using a pop-up alert or a notification bar. This feedback content becomes the output.

[0536] Step 9:

[0537] The device presents the user with improvement measures. Based on the results of the emotion analysis, it might display advice such as, "What you just said may be perceived as offensive to the other person. We recommend that you take a moment to calm down." This advice is the output.

[0538] Step 10:

[0539] The device saves past detection results and notification details as logs. Detection results and feedback details are saved in local storage or cloud storage so that users can check them later. This allows users to look back on past comments and alert details.

[0540] 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.

[0541] 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.

[0542] 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.

[0543] [Second embodiment]

[0544] FIG. 3 shows an example of the configuration of a data processing system 210 according to the second embodiment.

[0545] 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.

[0546] 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).

[0547] 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.

[0548] 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.

[0549] 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).

[0550] 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.

[0551] 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.

[0552] 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.

[0553] 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.

[0554] 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.

[0555] 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."

[0556] System Configuration

[0557] This invention is a system for preventing moral harassment using generative models, which is mainly composed of three parties: a server, a terminal, and a user. The server analyzes voice and text data collected from the terminal and aims to notify the user of an alert in real time.

[0558] Initial Setup

[0559] server

[0560] Integration and training of generative models: The server uses generative models to learn from past harassment cases, enabling analysis based on new data.

[0561] Database preparation: Prepare a database of past harassment cases and provide it as data to the generative model.

[0562] Terminal

[0563] Installing the app: Install the moral harassment prevention tool app on the user's smartphone or PC. The user must grant the necessary permissions (audio recording, text analysis, etc.).

[0564] Server collaboration: Collaborate with the server to download the generative model and necessary data.

[0565] Real-time monitoring and alert notifications

[0566] Terminal

[0567] Voice recording and text conversion: Record the user's voice in real time using the device's microphone and convert it into text data using voice recognition technology.

[0568] Text data collection: The contents of chat messages and emails entered by users are monitored in real time to collect text data.

[0569] Data transmission: The collected text data is sent to the server.

[0570] server

[0571] Data analysis: The server analyzes the received text data using a generative model to detect words and phrases that may be indicative of moral harassment.

[0572] Feedback generation: Based on the detection results, feedback is generated to notify the user.

[0573] Terminal

[0574] Alert Notification: Notify users of generated feedback in real time using pop-up alerts, notification bars, etc.

[0575] Suggesting improvement measures: Providing specific improvement measures to the user, such as "We recommend you take a break to calm down temporarily."

[0576] Saving logs: Past detection results and notification contents are saved as logs so that users can check them at a later date.

[0577] Specific examples

[0578] Example 1: Speaking during a meeting

[0579] 1. A user says during a meeting, "You're the reason this project failed."

[0580] 2. The device records the audio and converts it into text.

[0581] 3. The server analyzes the converted text using a generative model and determines that the phrase "It's your fault" is likely to be moral harassment.

[0582] 4. The server generates feedback and notifies the user that "Comments that blame others increase the mental burden on the other person and may constitute moral harassment."

[0583] 5. The device displays an alert to the user and suggests specific measures to improve the situation.

[0584] Example 2: Message in chat

[0585] 1. A user sends a message in chat saying, "Your mistake is causing trouble for everyone."

[0586] 2. The device monitors and collects chat content in real time.

[0587] 3. The server analyzes the collected text using a generative model and determines that the phrase "Everyone is in trouble because of your mistake" is a possible example of moral harassment.

[0588] 4. The server generates feedback stating that "attacks aimed at identifying and blaming individuals may constitute moral harassment."

[0589] 5. The device displays an alert to the user and a message encouraging them to stay calm.

[0590] This system can detect in advance whether comments or messages may constitute moral harassment, and by encouraging users to be aware of this, it can contribute to preventing moral harassment.

[0591] The processing flow will be explained below.

[0592] Step 1:

[0593] Users install the moral harassment prevention tool app on their smartphone or PC and grant the necessary permissions (audio recording, text analysis, etc.).

[0594] Step 2:

[0595] The device works in conjunction with the server to download the generative model and necessary data.

[0596] Step 3:

[0597] The device uses a microphone to record the user's voice in real time and converts it into text data using voice recognition technology.

[0598] Step 4:

[0599] The terminal monitors the contents of chat messages and emails entered by users in real time and collects text data.

[0600] Step 5:

[0601] The terminal transmits the collected text data (including the text converted from the speech) to the server.

[0602] Step 6:

[0603] The server analyzes the received text data using a generative model to detect words and phrases that may be indicative of moral harassment.

[0604] Step 7:

[0605] The server generates feedback to notify the user based on the detection results.

[0606] Step 8:

[0607] The device notifies the user of the feedback received from the server in real time using methods such as pop-up alerts and a notification bar.

[0608] Step 9:

[0609] The device will then suggest specific measures to improve the situation to the user, such as "We recommend you take a break to calm down temporarily."

[0610] Step 10:

[0611] The device stores past detection results and notification contents as a log so that the user can check them at a later date.

[0612] Example 1

[0613] 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."

[0614] In modern society, moral harassment (morahara) in the workplace and at home has become a serious problem. Morahara not only damages people's mental health and reduces productivity, but in severe cases can even lead to legal issues. Conventional systems and tools have made it difficult to detect and prevent moral harassment, and real-time monitoring and notification have rarely been realized. This has made it difficult to respond quickly and appropriately when moral harassment occurs.

[0615] 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.

[0616] In this invention, the server includes means for learning past cases of harassment using a generative model and monitoring speech and behavior in real time, means for detecting potentially harassing expressions, means for issuing an alert to notify the user when such expressions are detected, means for collecting messages and voice data entered by the user, means for transmitting the collected data to the server, means for analyzing the data received by the server and generating feedback, means for notifying the user of the generated feedback, and means for recording and saving past detection results and notification content. This makes it possible to detect possible moral harassment in real time and promptly notify the user.

[0617] A "generative model" is a type of artificial intelligence that learns from past data and analyzes and generates new data.

[0618] "Means for monitoring speech and behavior" refers to means for detecting user statements and behavior in real time and recording or analyzing the content.

[0619] "Means for detecting expressions" refers to means for detecting specific words or phrases from audio or text data and evaluating whether there is a problem.

[0620] "Means for issuing alerts" are means for notifying or warning users about detected problematic behavior or inappropriate comments.

[0621] "Means for collecting message and voice data" means means for recording text messages and voice input by users in real time and retaining them for subsequent processing.

[0622] The "means for transmitting data to a server" refers to a means for transmitting data collected by a terminal to a server via a network.

[0623] "Means for analyzing data" refers to means for analyzing received data in detail and detecting specific patterns or anomalies.

[0624] The "means for generating feedback" refers to the means for generating and notifying the user of countermeasures or warnings based on the analysis results.

[0625] "Means for notifying the user of feedback" refers to means for displaying or notifying the user of the generated feedback in real time.

[0626] "Means for recording and saving" refers to a means for saving past detection results and notification contents in a database or the like so that they can be referenced at a later date.

[0627] This invention is a moral harassment prevention system that uses a generative model. This system is mainly composed of three entities: a server, a terminal, and a user. A specific embodiment of the system will be described below.

[0628] Hardware and Software Configuration

[0629] server

[0630] The server is constructed using the following hardware and software:

[0631] Hardware: A server machine equipped with a high-performance CPU, GPU, and large memory capacity

[0632] Software: Python programming language, TensorFlow or PyTorch, MySQL or PostgreSQL database

[0633] Terminal

[0634] The terminal is a communication device such as a user's smartphone or PC.

[0635] Hardware: Smartphones (iOS, Android), PCs (Windows, macOS)

[0636] Software: Moral harassment prevention tool app, Google Speech-to-Text API, REST API client

[0637] System configuration

[0638] The system is realized by combining the following means:

[0639] 1. Generative Model Integration and Training

[0640] The server uses a generative model (e.g., GPT-4) to learn from past harassment cases, enabling real-time analysis based on new data.

[0641] 2. Prepare the database

[0642] The server prepares a database of past harassment cases and supplies it to the generative model as data. The database is MySQL or PostgreSQL.

[0643] 3. Install the app

[0644] Install a moral harassment prevention tool app on the user's smartphone or PC. The app is compatible with iOS and Android, and the user must grant the necessary permissions (voice recording, text analysis, etc.).

[0645] 4. Server Integration

[0646] The device connects to the server and downloads the generative model and necessary data. The data is obtained using a REST API.

[0647] Examples of concrete examples and prompts

[0648] Example 1: Statements made during a meeting

[0649] 1. A user says during a meeting, "You're the reason this project failed."

[0650] 2. The device records the audio and converts it to text using the Google Speech-to-Text API.

[0651] 3. The terminal sends the converted text to the server.

[0652] 4. The server analyzes the received text using the GPT-4 generative model and determines that the phrase "It's your fault" is likely to be moral harassment.

[0653] 5. The server generates feedback and notifies the user that the comment may be emotionally abusive.

[0654] 6. The device displays an alert to the user saying, "A statement that may be moral harassment has been detected."

[0655] Example 2: Chat messages

[0656] 1. A user sends a message in chat saying, "Your mistake is causing trouble for everyone."

[0657] 2. The device monitors and collects chat content in real time.

[0658] 3. The device sends the collected text to the server.

[0659] 4. The server analyzes the received text using the BERT generative model and determines that the phrase "Everyone is suffering because of your mistake" is a possible example of moral harassment.

[0660] 5. The server generates feedback stating that "attacks aimed at identifying and blaming individuals may constitute moral harassment."

[0661] 6. The device displays an alert to the user saying, "A statement that may be moral harassment has been detected."

[0662] Examples of prompt statements

[0663] "Respond to harassing phrases"

[0664] "Please determine whether this sentence constitutes moral harassment."

[0665] This invention can contribute to preventing moral harassment by detecting in real time whether a user's comments or messages may constitute moral harassment and quickly notifying the user.

[0666] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0667] Program processing flow

[0668] Step 1: Collecting audio data

[0669] Input: What the user says

[0670] Output: Collected audio data

[0671] Specific description:

[0672] The device collects the user's voice in real time through a microphone, specifically the built-in microphone of a smartphone or PC.

[0673] The app will launch the audio collection module and notify the user that "Audio recording has started."

[0674] Step 2: Convert audio data to text

[0675] Input: Collected audio data

[0676] Output: Converted text data

[0677] Specific description:

[0678] The device sends the collected voice data to the Google Speech-to-Text API, which converts the voice data into text.

[0679] Once the conversion is complete, the user is notified that "the audio data has been converted to text."

[0680] Step 3: Monitor and collect chat messages

[0681] Input: The chat message typed by the user

[0682] Output: Collected text messages

[0683] Specific description:

[0684] The device monitors the user's chat app (e.g., Slack or WhatsApp) and collects messages sent in real time.

[0685] Specifically, messages are retrieved using the chat app's API, and are temporarily stored on the device.

[0686] Step 4: Sending text data

[0687] Input: Collected text data

[0688] Output: Data sent to the server

[0689] Specific description:

[0690] The terminal transmits the collected text data to a server.

[0691] When sending, the data is AES encrypted to make it secure, and the user is notified that the data has been sent to the server.

[0692] Step 5: Data analysis

[0693] Input: Transmitted text data

[0694] Output: Analysis results

[0695] Specific description:

[0696] The text data received by the server is analyzed by a generative model (GPT-4 or BERT).

[0697] Specifically, to detect inappropriate words and phrases, prompts such as "Please respond to harassing phrases" are used.

[0698] When the processing is completed on the server, it records "Data analysis completed" in the log.

[0699] Step 6: Feedback generation

[0700] Input: Analysis results

[0701] Output: Generated feedback

[0702] Specific description:

[0703] The server generates feedback based on the analysis results of the generative model.

[0704] Specifically, it generates a notification message such as, "That comment may be emotional harassment. We recommend that you take a break to calm down."

[0705] Log "Feedback generated."

[0706] Step 7: Alert Notifications

[0707] Input: Generated feedback

[0708] Output: Alert notification to the user

[0709] Specific description:

[0710] The terminal notifies the user of the feedback received from the server in real time.

[0711] A pop-up alert or notification bar will be displayed to the user, informing them that "statements that may be emotionally abusive have been detected."

[0712] When a user acknowledges a notification, "User acknowledged notification" is logged.

[0713] Step 8: Propose improvement measures

[0714] Input: Generated feedback

[0715] Output: Present specific improvement measures to the user

[0716] Specific description:

[0717] The device will present specific improvement measures to the user.

[0718] For example, it will display advice such as, "We recommend you take a short walk to calm down."

[0719] When a user confirms a remediation, "User confirmed remediation" is logged.

[0720] Step 9: Save the logs

[0721] Input: Detection results and notification content

[0722] Output: Saved log data

[0723] Specific description:

[0724] The device stores past detection results and notification contents in an SQLite database.

[0725] After saving, you will be notified that "The log has been saved to the database."

[0726] Users can check this log at any time.

[0727] (Application example 1)

[0728] 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."

[0729] There is a need to prevent harassment that occurs in communication between employees in commercial facilities and improve the working environment. In particular, a system that allows employees to review their own behavior and take appropriate action through real-time monitoring and immediate feedback is needed, but this type of system has not yet been fully developed.

[0730] 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.

[0731] In this invention, the server includes means for learning from past cases of harassment using a generative model and monitoring speech and behavior in real time, means for detecting potentially harassing words and sentences, means for issuing an alert to notify users when such words or sentences are detected, means for recording the voices of employees at commercial facilities and converting them into text data, means for analyzing the converted text data, and means for generating feedback and notifying alerts and remedial measures. This makes it possible to immediately detect harassing behavior in real-time communication between employees and propose remedial measures, thereby ensuring a safe and comfortable working environment.

[0732] A "generative model" is a machine learning model that can generate new data based on past data.

[0733] "Harassment cases" refer to cases of conduct that cause psychological or physical distress to another person.

[0734] "Real-time monitoring means" are methods for instantly observing and recording current data and behavior.

[0735] An "alert mechanism" is a mechanism that sends a warning or notification to the user when certain conditions are met.

[0736] A "commercial facility" is a business establishment that has the facilities and space to offer goods and services.

[0737] "Means for recording employee voice and converting it into text data" refers to technology that records employee speech using a microphone or other device and converts that voice data into text information.

[0738] "Means for analyzing text data" refers to techniques for evaluating and analyzing textual information and extracting specific patterns and meanings.

[0739] "Means for generating feedback and notifying alerts and improvement measures" refers to methods for generating warnings and advice to users based on the analysis results and immediately communicating them.

[0740] This invention is a system for improving communication between employees in commercial facilities and preventing harassment. This system is composed of three entities: a server, a terminal, and a user.

[0741] server

[0742] The server uses a generative model to learn from past harassment cases and analyze new data. This allows it to analyze incoming data in real time and detect potentially abusive words and phrases. The server also generates feedback based on the analysis results and sends alerts and recommendations to the device.

[0743] Specifically, the server uses the following software and hardware:

[0744] Hardware: High-performance processor, memory, and large storage capacity

[0745] Software: Python, TensorFlow, Keras, SpeechRecognition library, Requests library

[0746] Terminal

[0747] The devices include the employee's smartphone or smart glasses. The devices have the ability to record voice data and convert it into text. The converted text data is then sent to the server in real time, and the received feedback is notified to the user.

[0748] The specific operation method is as follows:

[0749] 1. Record your voice using a microphone

[0750] 2. Convert the recorded audio into text using the SpeechRecognition library

[0751] 3. Send the converted text data to the server using the Requests library

[0752] 4. Receive feedback from the server and display it to the user

[0753] User

[0754] Users are primarily employees of commercial facilities, and their role is to record voice in real time, convert it into text data, and send it to a server. They also receive feedback from the server and correct their behavior.

[0755] Specific examples

[0756] Example of audio data

[0757] If an employee says something like "Did you make another mistake?", it will be handled as follows:

[0758] 1. Record audio using the device's microphone

[0759] 2. Convert recorded audio to text

[0760] 3. Send text data to the server

[0761] 4. The server analyzes and detects keywords such as "mistake."

[0762] 5. The server generates feedback and sends it to the device.

[0763] 6. The device will send an alert saying, "Comments that emphasize personal mistakes may lead to moral harassment."

[0764] 7. As an improvement measure, the message reads, "Work with other staff to identify problems and come up with solutions together."

[0765] Prompt Sentence Examples

[0766] "User sent text: 'Did you make another mistake?' Evaluate potential abuse and provide appropriate feedback and remediation."

[0767] In this way, this invention can prevent harassment between employees in commercial facilities. By providing real-time feedback, employees can immediately correct their behavior and promote healthy communication.

[0768] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0769] Step 1:

[0770] The user records their voice using the microphone on their device. The input is the user's speech, and the output is voice data. Specifically, the user's words are recorded while they are talking to other employees in a commercial facility.

[0771] Step 2:

[0772] The device converts the recorded voice data into text. The input is voice data, and the output is text data that converts the voice into text information. Specifically, the voice-to-text conversion process is performed using the SpeechRecognition library.

[0773] Step 3:

[0774] The terminal sends text data to the server. The input is text data, and the output is an HTTP request to the server. Specifically, the Requests library is used to send text data to the server via a POST request.

[0775] Step 4:

[0776] The server analyzes the received text data. The input is the text data, and the output is the analysis result. Specifically, it uses a generative AI model to analyze the text data and detect words and phrases that may be abusive or harassing.

[0777] Step 5:

[0778] The server generates feedback based on the analysis results. The input is the analysis results, and the output is a feedback message. Specifically, based on the analysis results provided by the generative AI model, it creates feedback for the user that includes appropriate alerts and improvement measures.

[0779] Step 6:

[0780] The server sends the generated feedback to the terminal. The input is a feedback message, and the output is an HTTP response to the terminal. Specifically, data including the feedback message is sent to the terminal as an HTTP response.

[0781] Step 7:

[0782] The terminal notifies the user of the received feedback. The input is a feedback message, and the output is an alert or notification to the user. Specifically, the terminal displays the feedback message as a pop-up on the display or in the notification bar.

[0783] Step 8:

[0784] The user checks the notification and corrects their behavior. The input is the feedback message, and the output is the user's improved behavior. Specifically, the user understands the alert and the improvement measures notified to them, and corrects their statements and behavior based on them.

[0785] This allows real-time communication between employees within a commercial facility to be monitored, preventing harassment before it occurs.

[0786] 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.

[0787] System Configuration

[0788] This invention is a system for preventing moral harassment that uses a generative model and an emotion engine, and is primarily composed of three parties: a server, a device, and a user. The server analyzes voice and text data collected from the device and aims to notify the user of an alert in real time. In addition, by combining this with an emotion engine, the system can analyze the user's emotions and provide more appropriate alerts and feedback.

[0789] Initial Setup

[0790] server

[0791] Integration and training of generative models: The server uses generative models to learn from past harassment cases, enabling analysis based on new data.

[0792] Emotion engine integration: An emotion engine is integrated into the server to recognize emotions from user voice and text data.

[0793] Database preparation: Prepare a database of past harassment cases and provide it as data to the generative model.

[0794] Terminal

[0795] Installing the app: Install the moral harassment prevention tool app on the user's smartphone or PC. The user must grant the necessary permissions (audio recording, text analysis, etc.).

[0796] Server collaboration: Collaborate with the server to download the generative model, emotion engine, and necessary data.

[0797] Real-time monitoring and alert notifications

[0798] Terminal

[0799] Voice recording and text conversion: Record the user's voice in real time using the device's microphone and convert it into text data using voice recognition technology.

[0800] Text data collection: The contents of chat messages and emails entered by users are monitored in real time to collect text data.

[0801] Data transmission: The collected text data (including the text converted from the speech) is sent to the server.

[0802] server

[0803] Data analysis: The server analyzes the received text data using a generative model to detect words and phrases that may be indicative of moral harassment.

[0804] Emotion analysis: Using an emotion engine, we recognize emotions from the user's voice and text data, such as anger and sadness.

[0805] Feedback generation: Based on the detection results and sentiment analysis results, feedback is generated to notify the user.

[0806] Terminal

[0807] Alert notification: The generated feedback is notified to the user in real time using methods such as pop-up alerts and notification bars.

[0808] Suggesting improvement measures: Presents specific improvement measures to the user. For example, based on the results of sentiment analysis, the system displays advice such as, "What you just said may be perceived as offensive to the other person. We recommend that you take some time to calm down."

[0809] Saving logs: Past detection results and notification contents are saved as logs so that users can check them at a later date.

[0810] Specific examples

[0811] Example 1: Speaking during a meeting

[0812] 1. A user says during a meeting, "You're the reason this project failed."

[0813] 2. The device records the audio and converts it into text.

[0814] 3. The server analyzes the converted text using a generative model and determines that the phrase "It's your fault" is likely to be moral harassment.

[0815] 4. The server uses the emotion engine to recognize the user's emotion, for example, recognizing that the user has an emotion of anger.

[0816] 5. The server generates feedback and informs the user, "Speaking in anger may come across as aggressive towards the other person. We recommend taking some time to calm down."

[0817] 6. The device displays an alert to the user and suggests specific measures to improve the situation.

[0818] Example 2: Message in chat

[0819] 1. A user sends a message in chat saying, "Your mistake is causing trouble for everyone."

[0820] 2. The device monitors and collects chat content in real time.

[0821] 3. The server analyzes the collected text using a generative model and determines that the phrase "Everyone is in trouble because of your mistake" is a possible example of moral harassment.

[0822] 4. The server uses the emotion engine to recognize the user's emotion, for example, recognizing that the user is dissatisfied.

[0823] 5. The server generates feedback and informs the user, "Sending this message out of frustration may come across as blaming. We recommend that you send the message again after you have calmed down a bit."

[0824] 6. The device will alert the user and provide specific advice to encourage them to stay calm.

[0825] This system can further contribute to preventing moral harassment by detecting in advance comments or messages that may constitute moral harassment and providing feedback and improvement measures that take into account the user's emotional state.

[0826] The processing flow will be explained below.

[0827] System Configuration

[0828] This invention is a system for preventing moral harassment that uses a generative model and an emotion engine, and is primarily composed of three parties: a server, a device, and a user. The server analyzes voice and text data collected from the device and aims to notify the user of an alert in real time. In addition, by combining this with an emotion engine, the system can recognize the user's emotions and provide more appropriate alerts and feedback.

[0829] Initial Setup

[0830] Step 1:

[0831] The server uses a generative model to learn from past harassment cases, allowing it to analyze new data.

[0832] Step 2:

[0833] The server integrates an emotion engine to recognize emotions from the user's voice and text data.

[0834] Step 3:

[0835] The server prepares a database of past harassment cases and supplies it as data to the generative model.

[0836] Step 4:

[0837] The device involves installing a moral harassment prevention tool app on the user's smartphone or PC, and the user granting the necessary permissions (voice recording, text analysis, etc.).

[0838] Step 5:

[0839] The device connects to the server and downloads the generative model, emotion engine, and necessary data.

[0840] Real-time monitoring and alert notifications

[0841] Step 6:

[0842] The device uses a microphone to record the user's voice in real time and converts it into text data using voice recognition technology.

[0843] Step 7:

[0844] The terminal monitors the contents of chat messages and emails entered by users in real time and collects text data.

[0845] Step 8:

[0846] The terminal transmits the collected text data (including the text converted from the speech) to the server.

[0847] Step 9:

[0848] The server analyzes the received text data using a generative model to detect words and phrases that may be indicative of moral harassment.

[0849] Step 10:

[0850] The server uses an emotion engine to recognize emotions from the user's voice and text data, such as anger and sadness.

[0851] Step 11:

[0852] The server generates feedback to notify the user based on the analysis results of the generative model and the emotion analysis results.

[0853] Step 12:

[0854] The device notifies the user of the feedback received from the server in real time using methods such as pop-up alerts and a notification bar.

[0855] Step 13:

[0856] The device will then present the user with specific advice on how to improve the situation, such as "What you just said may be perceived as offensive to the other person. We recommend that you take a moment to calm down."

[0857] Step 14:

[0858] The device stores past detection results and notification contents as a log so that the user can check them at a later date.

[0859] Specific examples

[0860] Example 1: Speaking during a meeting

[0861] Step 1:

[0862] A user says during a meeting, "You are the reason this project failed."

[0863] Step 2:

[0864] The device will record the audio and convert it into text.

[0865] Step 3:

[0866] The server analyzes the converted text using a generative model and determines that the phrase "It's your fault" is likely to be moral harassment.

[0867] Step 4:

[0868] The server uses an emotion engine to recognize the user's emotion, for example, recognizing that the user has an emotion of anger.

[0869] Step 5:

[0870] The server generates feedback, informing the user that "Speaking in anger may come across as aggressive towards the other person. We recommend that you take some time to calm down."

[0871] Step 6:

[0872] The device will display an alert to the user and suggest specific remedial measures.

[0873] Example 2: Message in chat

[0874] Step 1:

[0875] A user sends a message in chat saying, "Your mistake is causing trouble for everyone."

[0876] Step 2:

[0877] The device monitors and collects chat content in real time.

[0878] Step 3:

[0879] The server analyzes the collected text using a generative model and determines that the phrase "Everyone is in trouble because of your mistake" may be an example of moral harassment.

[0880] Step 4:

[0881] The server uses an emotion engine to recognize the user's emotion, for example, recognizing that the user is dissatisfied.

[0882] Step 5:

[0883] The server generates feedback, informing the user, "Sending this message out of frustration may come across as blaming. We recommend that you send the message again after you have calmed down a bit."

[0884] Step 6:

[0885] The device will display an alert to the user and provide specific advice to encourage them to stay calm.

[0886] In this way, a system that combines a generative model and an emotion engine can detect in advance whether a statement or message may constitute moral harassment, encourage the user to be aware of it, and provide appropriate feedback and improvement measures tailored to the user's emotional state.

[0887] Example 2

[0888] 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."

[0889] In modern society, moral harassment (moral harassment) is becoming a serious problem in various settings, including the workplace and school. This not only causes victims psychological damage, stress, and anxiety, but also reduces their productivity and quality of life. While existing moral harassment prevention systems can address certain cases, they are not considered effective preventative measures due to insufficient real-time monitoring, analysis, and feedback. Therefore, there is a need for a more advanced and comprehensive moral harassment prevention system that can reduce stress in users' daily lives and work and improve their mental health.

[0890] The specific processing by the specific processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means.

[0891] In this invention, the server includes means for learning from past harassment cases using a generative model and monitoring speech and behavior in real time, means for detecting potentially harassing words and sentences, means for issuing an alert to notify the user when such words or sentences are detected, means for integrating an emotion engine that analyzes the user's emotions, and means for generating feedback based on the emotion analysis results. This makes it possible to detect in real time whether a statement or message may constitute moral harassment and provide appropriate feedback and improvement measures.

[0892] A "generative model" is a type of artificial intelligence that learns from past data and generates appropriate outputs for new data.

[0893] "Harassment cases" are specific cases of moral harassment or psychological aggression that have occurred in the past.

[0894] "Means of real-time monitoring" refers to technology that allows for immediate monitoring of users' behavior and input.

[0895] "Word and phrase detection methods" are technologies that identify specific signs of harassment in users' statements and messages.

[0896] An "alert mechanism" is a mechanism for notifying users of detected potential harassment.

[0897] An "emotion engine" is a system for analyzing user emotions from voice and text data.

[0898] "Means for generating feedback" refers to technology that creates advice for users based on the results of detected harassment and sentiment analysis.

[0899] This invention is a system for preventing moral harassment that uses a generative model and an emotion engine to monitor users' behavior in real time and provide appropriate feedback. This system is mainly composed of three parties: a server, a terminal, and a user.

[0900] System Configuration

[0901] server

[0902] 1. Generative model integration and training:

[0903] The server uses a generative model to learn from past harassment cases. The specific generative model used is a "generative AI model," a general language model. This model learns patterns from past case data and can be applied to new situations.

[0904] 2. Emotion engine integration:

[0905] The server will be integrated with Emotion Engine, a general-purpose emotion analysis software, which will enable real-time analysis of emotions from user voice and text data, enabling more appropriate alerts and feedback to be provided.

[0906] 3. Prepare the database:

[0907] The server prepares a database that accumulates past cases of harassment. This database contains text data and emotion tags, and is used to analyze cases of moral harassment. The specific database management system used is MongoDB or MySQL.

[0908] Terminal

[0909] 1. Install the app:

[0910] Users install the anti-moral harassment app on their smartphones or PCs. The app asks the user for permission to record voice and monitor chat logs, and once permission is granted, the system will function normally.

[0911] 2. Server integration:

[0912] The device connects to the server to download data for the generative model and emotion engine, using secure protocols such as HTTPS for communication.

[0913] Specific examples

[0914] Remarks made during the meeting

[0915] 1. A user says during a meeting, "You're the reason this project failed."

[0916] 2. The device records the audio and converts it into text. The speech recognition technology uses the Google Cloud Speech-to-Text API.

[0917] 3. The server analyzes the converted text using a generative model and determines that the phrase "It's your fault" is likely to be moral harassment.

[0918] 4. The server uses an emotion engine to recognize the user's emotion. For example, it determines that the user is angry.

[0919] 5. The server generates feedback and displays an alert saying, "Speaking in anger may come across as aggressive towards the other person. We recommend taking some time to calm down."

[0920] 6. The device provides generated feedback to the user as an alert notification, either via a pop-up alert or a notification bar.

[0921] Messages in chat

[0922] 1. A user sends a message in chat saying, "Your mistake is causing trouble for everyone."

[0923] 2. The device monitors chat content in real time and collects text data.

[0924] 3. The server analyzes the collected text using a generative model and determines that the phrase "Everyone is in trouble because of your mistake" is a possible example of moral harassment.

[0925] 4. The server uses an emotion engine to recognize the user's emotion. For example, it determines that the user is dissatisfied.

[0926] 5. The server generates feedback and displays an alert saying, "Sending this message out of frustration may come across as blaming. We recommend you send the message again after you have calmed down a bit."

[0927] 6. The device provides the generated feedback to the user as an alert notification, using the chat app's notification function.

[0928] This system not only prevents moral harassment in real time, but also contributes to reducing mental stress by providing appropriate feedback that takes into account the user's emotional state.The system also monitors users' comments and messages and provides immediate feedback, enabling early detection and resolution of problems.

[0929] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0930] Step 1:

[0931] The server integrates a generative model and learns from past harassment cases.

[0932] As input, the generative model is fed with historical harassment case data, which is in text format and contains a large number of examples.

[0933] As part of data processing and data calculation, a generative model (e.g., a generative AI model) analyzes this text data and learns patterns of harassment. Specifically, the model processes the text data and performs a learning process to adjust weights and biases.

[0934] The output is a trained generative model.

[0935] Step 2:

[0936] The server integrates the emotion engine.

[0937] As input, you install the software for sentiment analysis on the server and enter the necessary API keys and settings.

[0938] As part of data processing and data calculation, the emotion engine (e.g., emotion analysis software) performs the initial setup and integration for analyzing emotions from voice and text data. Specific operations include configuring APIs and installing emotion analysis models.

[0939] As an output, we obtain a server with an integrated emotion engine.

[0940] Step 3:

[0941] The device installs the application and performs initial setup.

[0942] The input is to download and install the moral harassment prevention tool application onto your smartphone or PC.

[0943] For data processing and data calculation, the application obtains permission from the user to access voice recordings and chat logs. Specifically, the user operates the application installation screen and grants the necessary permissions.

[0944] The output will be an application that has been successfully installed and has the necessary permissions.

[0945] Step 4:

[0946] The device collects the voice data and converts it into text.

[0947] As input, the user's speech is recorded in real time through the device's microphone.

[0948] Data processing and calculation involves converting recorded voice data into text data using the Google Cloud Speech-to-Text API, etc. Specific operations include sending the voice data to a cloud server and converting it into text.

[0949] The output is the text data converted from the speech.

[0950] Step 5:

[0951] The device collects the text data entered and sends it to the server.

[0952] As input, the user enters a chat message or email.

[0953] For data processing and data calculation, the device monitors the user's input and collects text data in real time, which is temporarily stored in the device's memory.

[0954] As an output, we get the collected text data, which is then sent to a server using the HTTPS protocol.

[0955] Step 6:

[0956] The server analyzes the received text data.

[0957] As input, it receives text data sent from the terminal.

[0958] As part of the data processing and data calculation process, a generative model is used to analyze text data and detect words and phrases that may be related to moral harassment. Specifically, the model analyzes the text data and detects specific patterns.

[0959] The output is a list of phrases and words that are judged to be likely to be moral harassment.

[0960] Step 7:

[0961] The server performs sentiment analysis.

[0962] It takes parsed text data and audio data as input.

[0963] For data processing and data calculation, an emotion engine is used to analyze user emotions, including emotions such as anger and sadness. Specifically, an emotion analysis model is executed to extract emotions from text and voice data.

[0964] As an output, information about the user's emotional state is obtained.

[0965] Step 8:

[0966] The server generates the feedback.

[0967] As input, it receives the results of moral harassment detection and emotion analysis.

[0968] Based on the results of data processing and calculation, feedback is generated to notify the user. Specifically, appropriate advice or warning messages are created using specific algorithms and templates.

[0969] As an output, we get the generated feedback message.

[0970] Step 9:

[0971] The device will alert the user and suggest remedial measures.

[0972] As input, it receives the feedback message sent by the server.

[0973] As data processing and calculations are performed, alert notifications are displayed to the user. Notification methods include pop-up alerts and notification bars. Specific operations include using the device's notification function to provide real-time feedback to the user.

[0974] The output is an alert notification and remediation provided to the user.

[0975] Step 10:

[0976] The device stores the log.

[0977] Past detection results and notification contents are collected as input.

[0978] This data is then processed and calculated and stored in a database (e.g., SQLite) within the device. Specifically, the data is added to the database and made accessible at a later date.

[0979] As an output, you get the saved log data.

[0980] (Application example 2)

[0981] 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."

[0982] In physical stores, there is a need to prevent problems caused by inappropriate comments or emotional interactions during communication between staff and customers. However, conventional methods have difficulty analyzing comments and emotions in real time, making it difficult to provide appropriate feedback quickly. To solve this problem, a system is needed that can monitor and analyze the behavior of store staff in real time and provide appropriate feedback instantly.

[0983] The identification process 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 learning past harassment cases using a generative model and monitoring speech and behavior in real time, means for detecting words and sentences that may be harassing, means for issuing an alert to notify the user when such words or sentences are detected, means for analyzing emotions from voice and text data using an emotion engine, and means for providing appropriate remedial measures based on the analysis results. This makes it possible to improve the quality of communication between staff and customers in physical stores and prevent problems from occurring.

[0984] A "generative model" is an artificial intelligence technology that learns from past cases of harassment and analyzes the possibility of harassment in new data.

[0985] A "harassment case" is one in which inappropriate words, actions or behaviour cause mental distress or discomfort to a specific individual.

[0986] The "emotion engine" is a system for analyzing users' emotions from voice and text data.

[0987] "Real-time monitoring measures" are technologies for instantly collecting and analyzing ongoing audio and text data.

[0988] "Means of issuing an alert" refers to a method of notifying users when potential harassment is detected.

[0989] "Means for recording voice and converting it into text" refers to technology for recording speech and converting it into text data.

[0990] "Means of analysis" are methods for analyzing collected data to detect specific patterns or risks.

[0991] "Means for providing improvements" refers to methods for providing appropriate advice and action suggestions to users based on the analysis results.

[0992] System Configuration

[0993] This invention is a moral harassment prevention system that uses a generative model and an emotion engine, and is primarily composed of three parties: a server, a terminal, and a user. The server analyzes voice and text data collected from the terminal and aims to notify the user of an alert in real time. In addition, by combining it with an emotion engine, the system can analyze the user's emotions and provide more appropriate alerts and feedback.

[0994] Initial Setup

[0995] server

[0996] Generative model integration and training: The server uses a generative model to learn from past harassment cases, enabling analysis based on new data.

[0997] Emotion engine integration: Integrate an emotion engine into the server to recognize emotions from user voice and text data.

[0998] Database preparation: Prepare a database of past harassment cases and feed it to the generative model as data.

[0999] Terminal

[1000] Installing the app: Install the moral harassment prevention tool app on the user's smartphone or PC. The user must grant the necessary permissions (audio recording, text analysis, etc.).

[1001] Server collaboration: Collaborate with the server to download the generative model, emotion engine, and necessary data.

[1002] Real-time monitoring and alert notifications

[1003] Terminal

[1004] Voice recording and text conversion: Record the user's voice in real time using the device's microphone and convert it into text data using voice recognition technology.

[1005] Text data collection: The content of chat messages and emails entered by users is monitored in real time to collect text data.

[1006] Data transmission: Collected text data (including text converted from speech) is sent to the server.

[1007] server

[1008] Data analysis: The server analyzes the received text data using a generative model to detect words and phrases that may be moral harassment.

[1009] Emotion analysis: Using an emotion engine, we recognize emotions from the user's voice and text data, such as anger and sadness.

[1010] Feedback generation: Based on the detection results and sentiment analysis results, feedback is generated to notify the user.

[1011] Terminal

[1012] Alert notifications: Notify users of generated feedback in real time using methods such as pop-up alerts or notification bars.

[1013] Suggesting improvement measures: Presents specific improvement measures to the user. For example, based on the results of sentiment analysis, the system displays advice such as, "What you just said may be perceived as offensive to the other person. We recommend that you take a moment to calm down."

[1014] Saving logs: Past detection results and notification contents are saved as logs so that users can check them at a later date.

[1015] Specific examples

[1016] Example 1: Statements made during a meeting

[1017] 1. A user says during a meeting, "You're the reason this project failed."

[1018] 2. The device records the audio and converts it into text.

[1019] 3. The server analyzes the converted text using a generative model and determines that the phrase "It's your fault" is likely to be moral harassment.

[1020] 4. The server uses the emotion engine to recognize the user's emotion, for example, recognizing that the user has an emotion of anger.

[1021] 5. The server generates feedback and informs the user, "Speaking in anger may come across as aggressive towards the other person. We recommend taking some time to calm down."

[1022] 6. The device displays an alert to the user and suggests specific measures to improve the situation.

[1023] Example 2: Message in chat

[1024] 1. A user sends a message in chat saying, "Your mistake is causing trouble for everyone."

[1025] 2. The device monitors and collects chat content in real time.

[1026] 3. The server analyzes the collected text using a generative model and determines that the phrase "Everyone is in trouble because of your mistake" is a possible example of moral harassment.

[1027] 4. The server uses an emotion engine to recognize the user's emotion, for example, recognizing that the user is dissatisfied.

[1028] 5. The server generates feedback and informs the user, "Sending this message out of frustration may come across as blaming. We recommend that you send the message again after you have calmed down a bit."

[1029] 6. The device will alert the user and provide specific advice to encourage them to stay calm.

[1030] Prompt Sentence Examples

[1031] "Does this statement take into consideration your customer's feelings? I suggest you try phrasing it differently."

[1032] "The emotion recognition engine detects negative emotions. Please try to use as polite language as possible."

[1033] In this way, the system can improve the quality of communication between staff and customers in physical stores.

[1034] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[1035] Step 1:

[1036] The device records the user's voice in real time. Specifically, it uses the device's microphone to capture the user's speech and obtains the voice data. This voice data becomes the input.

[1037] Step 2:

[1038] The device converts the voice data it receives into text. It uses voice recognition technology to analyze the voice data and generate corresponding text data. This text data is then output.

[1039] Step 3:

[1040] The device sends the converted text data to the server. Specifically, the device uploads the text data to the server via the Internet. This text data becomes the input.

[1041] Step 4:

[1042] The server analyzes the received text data using a generative model. The generative model has learned from past harassment cases and analyzes new text data to detect potentially harassing words and phrases. The results of this analysis are output.

[1043] Step 5:

[1044] The server uses an emotion engine to perform emotion analysis based on the analysis results. Specifically, the server analyzes the user's emotions (anger, dissatisfaction, etc.) from the text data. The analysis results are output.

[1045] Step 6:

[1046] The server generates feedback based on the analysis results and sentiment analysis results. The generative model and sentiment analysis results are combined to generate appropriate advice or warnings for the user. This feedback is the output.

[1047] Step 7:

[1048] The server sends the generated feedback to the terminal. Specifically, the server sends the generated feedback to the terminal via the Internet. This feedback becomes the input.

[1049] Step 8:

[1050] The device displays an alert to the user. Specifically, the feedback content is displayed on the device display using a pop-up alert or a notification bar. This feedback content becomes the output.

[1051] Step 9:

[1052] The device presents the user with improvement measures. Based on the results of the emotion analysis, it might display advice such as, "What you just said may be perceived as offensive to the other person. We recommend that you take a moment to calm down." This advice is the output.

[1053] Step 10:

[1054] The device saves past detection results and notification details as logs. Detection results and feedback details are saved in local storage or cloud storage so that users can check them later. This allows users to look back on past comments and alert details.

[1055] 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.

[1056] 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.

[1057] 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.

[1058] [Third embodiment]

[1059] FIG. 5 shows an example of the configuration of a data processing system 310 according to the third embodiment.

[1060] 5, the data processing system 310 includes the data processing device 12 and a headset type terminal 314. An example of the data processing device 12 is a server.

[1061] 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).

[1062] 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.

[1063] 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.

[1064] 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).

[1065] 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.

[1066] 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.

[1067] 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.

[1068] 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.

[1069] 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.

[1070] 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."

[1071] System Configuration

[1072] This invention is a system for preventing moral harassment using generative models, which is mainly composed of three parties: a server, a terminal, and a user. The server analyzes voice and text data collected from the terminal and aims to notify the user of an alert in real time.

[1073] Initial Setup

[1074] server

[1075] Integration and training of generative models: The server uses generative models to learn from past harassment cases, enabling analysis based on new data.

[1076] Database preparation: Prepare a database of past harassment cases and provide it as data to the generative model.

[1077] Terminal

[1078] Installing the app: Install the moral harassment prevention tool app on the user's smartphone or PC. The user must grant the necessary permissions (audio recording, text analysis, etc.).

[1079] Server collaboration: Collaborate with the server to download the generative model and necessary data.

[1080] Real-time monitoring and alert notifications

[1081] Terminal

[1082] Voice recording and text conversion: Record the user's voice in real time using the device's microphone and convert it into text data using voice recognition technology.

[1083] Text data collection: The contents of chat messages and emails entered by users are monitored in real time to collect text data.

[1084] Data transmission: The collected text data is sent to the server.

[1085] server

[1086] Data analysis: The server analyzes the received text data using a generative model to detect words and phrases that may be indicative of moral harassment.

[1087] Feedback generation: Based on the detection results, feedback is generated to notify the user.

[1088] Terminal

[1089] Alert Notification: Notify users of generated feedback in real time using pop-up alerts, notification bars, etc.

[1090] Suggesting improvement measures: Providing specific improvement measures to the user, such as "We recommend you take a break to calm down temporarily."

[1091] Saving logs: Past detection results and notification contents are saved as logs so that users can check them at a later date.

[1092] Specific examples

[1093] Example 1: Speaking during a meeting

[1094] 1. A user says during a meeting, "You're the reason this project failed."

[1095] 2. The device records the audio and converts it into text.

[1096] 3. The server analyzes the converted text using a generative model and determines that the phrase "It's your fault" is likely to be moral harassment.

[1097] 4. The server generates feedback and notifies the user that "Comments that blame others increase the mental burden on the other person and may constitute moral harassment."

[1098] 5. The device displays an alert to the user and suggests specific measures to improve the situation.

[1099] Example 2: Message in chat

[1100] 1. A user sends a message in chat saying, "Your mistake is causing trouble for everyone."

[1101] 2. The device monitors and collects chat content in real time.

[1102] 3. The server analyzes the collected text using a generative model and determines that the phrase "Everyone is in trouble because of your mistake" is a possible example of moral harassment.

[1103] 4. The server generates feedback stating that "attacks aimed at identifying and blaming individuals may constitute moral harassment."

[1104] 5. The device displays an alert to the user and a message encouraging them to stay calm.

[1105] This system can detect in advance whether comments or messages may constitute moral harassment, and by encouraging users to be aware of this, it can contribute to preventing moral harassment.

[1106] The processing flow will be explained below.

[1107] Step 1:

[1108] Users install the moral harassment prevention tool app on their smartphone or PC and grant the necessary permissions (audio recording, text analysis, etc.).

[1109] Step 2:

[1110] The device works in conjunction with the server to download the generative model and necessary data.

[1111] Step 3:

[1112] The device uses a microphone to record the user's voice in real time and converts it into text data using voice recognition technology.

[1113] Step 4:

[1114] The terminal monitors the contents of chat messages and emails entered by users in real time and collects text data.

[1115] Step 5:

[1116] The terminal transmits the collected text data (including the text converted from the speech) to the server.

[1117] Step 6:

[1118] The server analyzes the received text data using a generative model to detect words and phrases that may be indicative of moral harassment.

[1119] Step 7:

[1120] The server generates feedback to notify the user based on the detection results.

[1121] Step 8:

[1122] The device notifies the user of the feedback received from the server in real time using methods such as pop-up alerts and a notification bar.

[1123] Step 9:

[1124] The device will then suggest specific measures to improve the situation to the user, such as "We recommend you take a break to calm down temporarily."

[1125] Step 10:

[1126] The device stores past detection results and notification contents as a log so that the user can check them at a later date.

[1127] Example 1

[1128] 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."

[1129] In modern society, moral harassment (morahara) in the workplace and at home has become a serious problem. Morahara not only damages people's mental health and reduces productivity, but in severe cases can even lead to legal issues. Conventional systems and tools have made it difficult to detect and prevent moral harassment, and real-time monitoring and notification have rarely been realized. This has made it difficult to respond quickly and appropriately when moral harassment occurs.

[1130] 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.

[1131] In this invention, the server includes means for learning past cases of harassment using a generative model and monitoring speech and behavior in real time, means for detecting potentially harassing expressions, means for issuing an alert to notify the user when such expressions are detected, means for collecting messages and voice data entered by the user, means for transmitting the collected data to the server, means for analyzing the data received by the server and generating feedback, means for notifying the user of the generated feedback, and means for recording and saving past detection results and notification content. This makes it possible to detect possible moral harassment in real time and promptly notify the user.

[1132] A "generative model" is a type of artificial intelligence that learns from past data and analyzes and generates new data.

[1133] "Means for monitoring speech and behavior" refers to means for detecting user statements and behavior in real time and recording or analyzing the content.

[1134] "Means for detecting expressions" refers to means for detecting specific words or phrases from audio or text data and evaluating whether there is a problem.

[1135] "Means for issuing alerts" are means for notifying or warning users about detected problematic behavior or inappropriate comments.

[1136] "Means for collecting message and voice data" means means for recording text messages and voice input by users in real time and retaining them for subsequent processing.

[1137] The "means for transmitting data to a server" refers to a means for transmitting data collected by a terminal to a server via a network.

[1138] "Means for analyzing data" refers to means for analyzing received data in detail and detecting specific patterns or anomalies.

[1139] The "means for generating feedback" refers to the means for generating and notifying the user of countermeasures or warnings based on the analysis results.

[1140] "Means for notifying the user of feedback" refers to means for displaying or notifying the user of the generated feedback in real time.

[1141] "Means for recording and saving" refers to a means for saving past detection results and notification contents in a database or the like so that they can be referenced at a later date.

[1142] This invention is a moral harassment prevention system that uses a generative model. This system is mainly composed of three entities: a server, a terminal, and a user. A specific embodiment of the system will be described below.

[1143] Hardware and Software Configuration

[1144] server

[1145] The server is constructed using the following hardware and software:

[1146] Hardware: A server machine equipped with a high-performance CPU, GPU, and large memory capacity

[1147] Software: Python programming language, TensorFlow or PyTorch, MySQL or PostgreSQL database

[1148] Terminal

[1149] The terminal is a communication device such as a user's smartphone or PC.

[1150] Hardware: Smartphones (iOS, Android), PCs (Windows, macOS)

[1151] Software: Moral harassment prevention tool app, Google Speech-to-Text API, REST API client

[1152] System configuration

[1153] The system is realized by combining the following means:

[1154] 1. Generative Model Integration and Training

[1155] The server uses a generative model (e.g., GPT-4) to learn from past harassment cases, enabling real-time analysis based on new data.

[1156] 2. Prepare the database

[1157] The server prepares a database of past harassment cases and supplies it to the generative model as data. The database is MySQL or PostgreSQL.

[1158] 3. Install the app

[1159] Install a moral harassment prevention tool app on the user's smartphone or PC. The app is compatible with iOS and Android, and the user must grant the necessary permissions (voice recording, text analysis, etc.).

[1160] 4. Server Integration

[1161] The device connects to the server and downloads the generative model and necessary data. The data is obtained using a REST API.

[1162] Examples of concrete examples and prompts

[1163] Example 1: Statements made during a meeting

[1164] 1. A user says during a meeting, "You're the reason this project failed."

[1165] 2. The device records the audio and converts it to text using the Google Speech-to-Text API.

[1166] 3. The terminal sends the converted text to the server.

[1167] 4. The server analyzes the received text using the GPT-4 generative model and determines that the phrase "It's your fault" is likely to be moral harassment.

[1168] 5. The server generates feedback and notifies the user that the comment may be emotionally abusive.

[1169] 6. The device displays an alert to the user saying, "A statement that may be moral harassment has been detected."

[1170] Example 2: Chat messages

[1171] 1. A user sends a message in chat saying, "Your mistake is causing trouble for everyone."

[1172] 2. The device monitors and collects chat content in real time.

[1173] 3. The device sends the collected text to the server.

[1174] 4. The server analyzes the received text using the BERT generative model and determines that the phrase "Everyone is suffering because of your mistake" is a possible example of moral harassment.

[1175] 5. The server generates feedback stating that "attacks aimed at identifying and blaming individuals may constitute moral harassment."

[1176] 6. The device displays an alert to the user saying, "A statement that may be moral harassment has been detected."

[1177] Examples of prompt statements

[1178] "Respond to harassing phrases"

[1179] "Please determine whether this sentence constitutes moral harassment."

[1180] This invention can contribute to preventing moral harassment by detecting in real time whether a user's comments or messages may constitute moral harassment and quickly notifying the user.

[1181] The flow of the identification process in the first embodiment will be described with reference to FIG.

[1182] Program processing flow

[1183] Step 1: Collecting audio data

[1184] Input: What the user says

[1185] Output: Collected audio data

[1186] Specific description:

[1187] The device collects the user's voice in real time through a microphone, specifically the built-in microphone of a smartphone or PC.

[1188] The app will launch the audio collection module and notify the user that "Audio recording has started."

[1189] Step 2: Convert audio data to text

[1190] Input: Collected audio data

[1191] Output: Converted text data

[1192] Specific description:

[1193] The device sends the collected voice data to the Google Speech-to-Text API, which converts the voice data into text.

[1194] Once the conversion is complete, the user is notified that "the audio data has been converted to text."

[1195] Step 3: Monitor and collect chat messages

[1196] Input: The chat message typed by the user

[1197] Output: Collected text messages

[1198] Specific description:

[1199] The device monitors the user's chat app (e.g., Slack or WhatsApp) and collects messages sent in real time.

[1200] Specifically, messages are retrieved using the chat app's API, and are temporarily stored on the device.

[1201] Step 4: Sending text data

[1202] Input: Collected text data

[1203] Output: Data sent to the server

[1204] Specific description:

[1205] The terminal transmits the collected text data to a server.

[1206] When sending, the data is AES encrypted to make it secure, and the user is notified that the data has been sent to the server.

[1207] Step 5: Data analysis

[1208] Input: Transmitted text data

[1209] Output: Analysis results

[1210] Specific description:

[1211] The text data received by the server is analyzed by a generative model (GPT-4 or BERT).

[1212] Specifically, to detect inappropriate words and phrases, prompts such as "Please respond to harassing phrases" are used.

[1213] When the processing is completed on the server, it records "Data analysis completed" in the log.

[1214] Step 6: Feedback generation

[1215] Input: Analysis results

[1216] Output: Generated feedback

[1217] Specific description:

[1218] The server generates feedback based on the analysis results of the generative model.

[1219] Specifically, it generates a notification message such as, "That comment may be emotional harassment. We recommend that you take a break to calm down."

[1220] Log "Feedback generated."

[1221] Step 7: Alert Notifications

[1222] Input: Generated feedback

[1223] Output: Alert notification to the user

[1224] Specific description:

[1225] The terminal notifies the user of the feedback received from the server in real time.

[1226] A pop-up alert or notification bar will be displayed to the user, informing them that "statements that may be emotionally abusive have been detected."

[1227] When a user acknowledges a notification, "User acknowledged notification" is logged.

[1228] Step 8: Propose improvement measures

[1229] Input: Generated feedback

[1230] Output: Present specific improvement measures to the user

[1231] Specific description:

[1232] The device will present specific improvement measures to the user.

[1233] For example, it will display advice such as, "We recommend you take a short walk to calm down."

[1234] When a user confirms a remediation, "User confirmed remediation" is logged.

[1235] Step 9: Save the logs

[1236] Input: Detection results and notification content

[1237] Output: Saved log data

[1238] Specific description:

[1239] The device stores past detection results and notification contents in an SQLite database.

[1240] After saving, you will be notified that "The log has been saved to the database."

[1241] Users can check this log at any time.

[1242] (Application example 1)

[1243] 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."

[1244] There is a need to prevent harassment that occurs in communication between employees in commercial facilities and improve the working environment. In particular, a system that allows employees to review their own behavior and take appropriate action through real-time monitoring and immediate feedback is needed, but this type of system has not yet been fully developed.

[1245] 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.

[1246] In this invention, the server includes means for learning from past cases of harassment using a generative model and monitoring speech and behavior in real time, means for detecting potentially harassing words and sentences, means for issuing an alert to notify users when such words or sentences are detected, means for recording the voices of employees at commercial facilities and converting them into text data, means for analyzing the converted text data, and means for generating feedback and notifying alerts and remedial measures. This makes it possible to immediately detect harassing behavior in real-time communication between employees and propose remedial measures, thereby ensuring a safe and comfortable working environment.

[1247] A "generative model" is a machine learning model that can generate new data based on past data.

[1248] "Harassment cases" refer to cases of conduct that cause psychological or physical distress to another person.

[1249] "Real-time monitoring means" are methods for instantly observing and recording current data and behavior.

[1250] An "alert mechanism" is a mechanism that sends a warning or notification to the user when certain conditions are met.

[1251] A "commercial facility" is a business establishment that has the facilities and space to offer goods and services.

[1252] "Means for recording employee voice and converting it into text data" refers to technology that records employee speech using a microphone or other device and converts that voice data into text information.

[1253] "Means for analyzing text data" refers to techniques for evaluating and analyzing textual information and extracting specific patterns and meanings.

[1254] "Means for generating feedback and notifying alerts and improvement measures" refers to methods for generating warnings and advice to users based on the analysis results and immediately communicating them.

[1255] This invention is a system for improving communication between employees in commercial facilities and preventing harassment. This system is composed of three entities: a server, a terminal, and a user.

[1256] server

[1257] The server uses a generative model to learn from past harassment cases and analyze new data. This allows it to analyze incoming data in real time and detect potentially abusive words and phrases. The server also generates feedback based on the analysis results and sends alerts and recommendations to the device.

[1258] Specifically, the server uses the following software and hardware:

[1259] Hardware: High-performance processor, memory, and large storage capacity

[1260] Software: Python, TensorFlow, Keras, SpeechRecognition library, Requests library

[1261] Terminal

[1262] The devices include the employee's smartphone or smart glasses. The devices have the ability to record voice data and convert it into text. The converted text data is then sent to the server in real time, and the received feedback is notified to the user.

[1263] The specific operation method is as follows:

[1264] 1. Record your voice using a microphone

[1265] 2. Convert the recorded audio into text using the SpeechRecognition library

[1266] 3. Send the converted text data to the server using the Requests library

[1267] 4. Receive feedback from the server and display it to the user

[1268] User

[1269] Users are primarily employees of commercial facilities, and their role is to record voice in real time, convert it into text data, and send it to a server. They also receive feedback from the server and correct their behavior.

[1270] Specific examples

[1271] Example of audio data

[1272] If an employee says something like "Did you make another mistake?", it will be handled as follows:

[1273] 1. Record audio using the device's microphone

[1274] 2. Convert recorded audio to text

[1275] 3. Send text data to the server

[1276] 4. The server analyzes and detects keywords such as "mistake."

[1277] 5. The server generates feedback and sends it to the device.

[1278] 6. The device will send an alert saying, "Comments that emphasize personal mistakes may lead to moral harassment."

[1279] 7. As an improvement measure, the message reads, "Work with other staff to identify problems and come up with solutions together."

[1280] Prompt Sentence Examples

[1281] "User sent text: 'Did you make another mistake?' Evaluate potential abuse and provide appropriate feedback and remediation."

[1282] In this way, this invention can prevent harassment between employees in commercial facilities. By providing real-time feedback, employees can immediately correct their behavior and promote healthy communication.

[1283] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[1284] Step 1:

[1285] The user records their voice using the microphone on their device. The input is the user's speech, and the output is voice data. Specifically, the user's words are recorded while they are talking to other employees in a commercial facility.

[1286] Step 2:

[1287] The device converts the recorded voice data into text. The input is voice data, and the output is text data that converts the voice into text information. Specifically, the voice-to-text conversion process is performed using the SpeechRecognition library.

[1288] Step 3:

[1289] The terminal sends text data to the server. The input is text data, and the output is an HTTP request to the server. Specifically, the Requests library is used to send text data to the server via a POST request.

[1290] Step 4:

[1291] The server analyzes the received text data. The input is the text data, and the output is the analysis result. Specifically, it uses a generative AI model to analyze the text data and detect words and phrases that may be abusive or harassing.

[1292] Step 5:

[1293] The server generates feedback based on the analysis results. The input is the analysis results, and the output is a feedback message. Specifically, based on the analysis results provided by the generative AI model, it creates feedback for the user that includes appropriate alerts and improvement measures.

[1294] Step 6:

[1295] The server sends the generated feedback to the terminal. The input is a feedback message, and the output is an HTTP response to the terminal. Specifically, data including the feedback message is sent to the terminal as an HTTP response.

[1296] Step 7:

[1297] The terminal notifies the user of the received feedback. The input is a feedback message, and the output is an alert or notification to the user. Specifically, the terminal displays the feedback message as a pop-up on the display or in the notification bar.

[1298] Step 8:

[1299] The user checks the notification and corrects their behavior. The input is the feedback message, and the output is the user's improved behavior. Specifically, the user understands the alert and the improvement measures notified to them, and corrects their statements and behavior based on them.

[1300] This allows real-time communication between employees within a commercial facility to be monitored, preventing harassment before it occurs.

[1301] 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.

[1302] System Configuration

[1303] This invention is a system for preventing moral harassment that uses a generative model and an emotion engine, and is primarily composed of three parties: a server, a device, and a user. The server analyzes voice and text data collected from the device and aims to notify the user of an alert in real time. In addition, by combining this with an emotion engine, the system can analyze the user's emotions and provide more appropriate alerts and feedback.

[1304] Initial Setup

[1305] server

[1306] Integration and training of generative models: The server uses generative models to learn from past harassment cases, enabling analysis based on new data.

[1307] Emotion engine integration: An emotion engine is integrated into the server to recognize emotions from user voice and text data.

[1308] Database preparation: Prepare a database of past harassment cases and provide it as data to the generative model.

[1309] Terminal

[1310] Installing the app: Install the moral harassment prevention tool app on the user's smartphone or PC. The user must grant the necessary permissions (audio recording, text analysis, etc.).

[1311] Server collaboration: Collaborate with the server to download the generative model, emotion engine, and necessary data.

[1312] Real-time monitoring and alert notifications

[1313] Terminal

[1314] Voice recording and text conversion: Record the user's voice in real time using the device's microphone and convert it into text data using voice recognition technology.

[1315] Text data collection: The contents of chat messages and emails entered by users are monitored in real time to collect text data.

[1316] Data transmission: The collected text data (including the text converted from the speech) is sent to the server.

[1317] server

[1318] Data analysis: The server analyzes the received text data using a generative model to detect words and phrases that may be indicative of moral harassment.

[1319] Emotion analysis: Using an emotion engine, we recognize emotions from the user's voice and text data, such as anger and sadness.

[1320] Feedback generation: Based on the detection results and sentiment analysis results, feedback is generated to notify the user.

[1321] Terminal

[1322] Alert notification: The generated feedback is notified to the user in real time using methods such as pop-up alerts and notification bars.

[1323] Suggesting improvement measures: Presents specific improvement measures to the user. For example, based on the results of sentiment analysis, the system displays advice such as, "What you just said may be perceived as offensive to the other person. We recommend that you take some time to calm down."

[1324] Saving logs: Past detection results and notification contents are saved as logs so that users can check them at a later date.

[1325] Specific examples

[1326] Example 1: Speaking during a meeting

[1327] 1. A user says during a meeting, "You're the reason this project failed."

[1328] 2. The device records the audio and converts it into text.

[1329] 3. The server analyzes the converted text using a generative model and determines that the phrase "It's your fault" is likely to be moral harassment.

[1330] 4. The server uses the emotion engine to recognize the user's emotion, for example, recognizing that the user has an emotion of anger.

[1331] 5. The server generates feedback and informs the user, "Speaking in anger may come across as aggressive towards the other person. We recommend taking some time to calm down."

[1332] 6. The device displays an alert to the user and suggests specific measures to improve the situation.

[1333] Example 2: Message in chat

[1334] 1. A user sends a message in chat saying, "Your mistake is causing trouble for everyone."

[1335] 2. The device monitors and collects chat content in real time.

[1336] 3. The server analyzes the collected text using a generative model and determines that the phrase "Everyone is in trouble because of your mistake" is a possible example of moral harassment.

[1337] 4. The server uses the emotion engine to recognize the user's emotion, for example, recognizing that the user is dissatisfied.

[1338] 5. The server generates feedback and informs the user, "Sending this message out of frustration may come across as blaming. We recommend that you send the message again after you have calmed down a bit."

[1339] 6. The device will alert the user and provide specific advice to encourage them to stay calm.

[1340] This system can further contribute to preventing moral harassment by detecting in advance comments or messages that may constitute moral harassment and providing feedback and improvement measures that take into account the user's emotional state.

[1341] The processing flow will be explained below.

[1342] System Configuration

[1343] This invention is a system for preventing moral harassment that uses a generative model and an emotion engine, and is primarily composed of three parties: a server, a device, and a user. The server analyzes voice and text data collected from the device and aims to notify the user of an alert in real time. In addition, by combining this with an emotion engine, the system can recognize the user's emotions and provide more appropriate alerts and feedback.

[1344] Initial Setup

[1345] Step 1:

[1346] The server uses a generative model to learn from past harassment cases, allowing it to analyze new data.

[1347] Step 2:

[1348] The server integrates an emotion engine to recognize emotions from the user's voice and text data.

[1349] Step 3:

[1350] The server prepares a database of past harassment cases and supplies it as data to the generative model.

[1351] Step 4:

[1352] The device involves installing a moral harassment prevention tool app on the user's smartphone or PC, and the user granting the necessary permissions (voice recording, text analysis, etc.).

[1353] Step 5:

[1354] The device connects to the server and downloads the generative model, emotion engine, and necessary data.

[1355] Real-time monitoring and alert notifications

[1356] Step 6:

[1357] The device uses a microphone to record the user's voice in real time and converts it into text data using voice recognition technology.

[1358] Step 7:

[1359] The terminal monitors the contents of chat messages and emails entered by users in real time and collects text data.

[1360] Step 8:

[1361] The terminal transmits the collected text data (including the text converted from the speech) to the server.

[1362] Step 9:

[1363] The server analyzes the received text data using a generative model to detect words and phrases that may be indicative of moral harassment.

[1364] Step 10:

[1365] The server uses an emotion engine to recognize emotions from the user's voice and text data, such as anger and sadness.

[1366] Step 11:

[1367] The server generates feedback to notify the user based on the analysis results of the generative model and the emotion analysis results.

[1368] Step 12:

[1369] The device notifies the user of the feedback received from the server in real time using methods such as pop-up alerts and a notification bar.

[1370] Step 13:

[1371] The device will then present the user with specific advice on how to improve the situation, such as "What you just said may be perceived as offensive to the other person. We recommend that you take a moment to calm down."

[1372] Step 14:

[1373] The device stores past detection results and notification contents as a log so that the user can check them at a later date.

[1374] Specific examples

[1375] Example 1: Speaking during a meeting

[1376] Step 1:

[1377] A user says during a meeting, "You are the reason this project failed."

[1378] Step 2:

[1379] The device will record the audio and convert it into text.

[1380] Step 3:

[1381] The server analyzes the converted text using a generative model and determines that the phrase "It's your fault" is likely to be moral harassment.

[1382] Step 4:

[1383] The server uses an emotion engine to recognize the user's emotion, for example, recognizing that the user has an emotion of anger.

[1384] Step 5:

[1385] The server generates feedback, informing the user that "Speaking in anger may come across as aggressive towards the other person. We recommend that you take some time to calm down."

[1386] Step 6:

[1387] The device will display an alert to the user and suggest specific remedial measures.

[1388] Example 2: Message in chat

[1389] Step 1:

[1390] A user sends a message in chat saying, "Your mistake is causing trouble for everyone."

[1391] Step 2:

[1392] The device monitors and collects chat content in real time.

[1393] Step 3:

[1394] The server analyzes the collected text using a generative model and determines that the phrase "Everyone is in trouble because of your mistake" may be an example of moral harassment.

[1395] Step 4:

[1396] The server uses an emotion engine to recognize the user's emotion, for example, recognizing that the user is dissatisfied.

[1397] Step 5:

[1398] The server generates feedback, informing the user, "Sending this message out of frustration may come across as blaming. We recommend that you send the message again after you have calmed down a bit."

[1399] Step 6:

[1400] The device will display an alert to the user and provide specific advice to encourage them to stay calm.

[1401] In this way, a system that combines a generative model and an emotion engine can detect in advance whether a statement or message may constitute moral harassment, encourage the user to be aware of it, and provide appropriate feedback and improvement measures tailored to the user's emotional state.

[1402] Example 2

[1403] 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."

[1404] In modern society, moral harassment (moral harassment) is becoming a serious problem in various settings, including the workplace and school. This not only causes victims psychological damage, stress, and anxiety, but also reduces their productivity and quality of life. While existing moral harassment prevention systems can address certain cases, they are not considered effective preventative measures due to insufficient real-time monitoring, analysis, and feedback. Therefore, there is a need for a more advanced and comprehensive moral harassment prevention system that can reduce stress in users' daily lives and work and improve their mental health.

[1405] The specific processing by the specific processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means.

[1406] In this invention, the server includes means for learning from past harassment cases using a generative model and monitoring speech and behavior in real time, means for detecting potentially harassing words and sentences, means for issuing an alert to notify the user when such words or sentences are detected, means for integrating an emotion engine that analyzes the user's emotions, and means for generating feedback based on the emotion analysis results. This makes it possible to detect in real time whether a statement or message may constitute moral harassment and provide appropriate feedback and improvement measures.

[1407] A "generative model" is a type of artificial intelligence that learns from past data and generates appropriate outputs for new data.

[1408] "Harassment cases" are specific cases of moral harassment or psychological aggression that have occurred in the past.

[1409] "Means of real-time monitoring" refers to technology that allows for immediate monitoring of users' behavior and input.

[1410] "Word and phrase detection methods" are technologies that identify specific signs of harassment in users' statements and messages.

[1411] An "alert mechanism" is a mechanism for notifying users of detected potential harassment.

[1412] An "emotion engine" is a system for analyzing user emotions from voice and text data.

[1413] "Means for generating feedback" refers to technology that creates advice for users based on the results of detected harassment and sentiment analysis.

[1414] This invention is a system for preventing moral harassment that uses a generative model and an emotion engine to monitor users' behavior in real time and provide appropriate feedback. This system is mainly composed of three parties: a server, a terminal, and a user.

[1415] System Configuration

[1416] server

[1417] 1. Generative model integration and training:

[1418] The server uses a generative model to learn from past harassment cases. The specific generative model used is a "generative AI model," a general language model. This model learns patterns from past case data and can be applied to new situations.

[1419] 2. Emotion engine integration:

[1420] The server will be integrated with Emotion Engine, a general-purpose emotion analysis software, which will enable real-time analysis of emotions from user voice and text data, enabling more appropriate alerts and feedback to be provided.

[1421] 3. Prepare the database:

[1422] The server prepares a database that accumulates past cases of harassment. This database contains text data and emotion tags, and is used to analyze cases of moral harassment. The specific database management system used is MongoDB or MySQL.

[1423] Terminal

[1424] 1. Install the app:

[1425] Users install the anti-moral harassment app on their smartphones or PCs. The app asks the user for permission to record voice and monitor chat logs, and once permission is granted, the system will function normally.

[1426] 2. Server integration:

[1427] The device connects to the server to download data for the generative model and emotion engine, using secure protocols such as HTTPS for communication.

[1428] Specific examples

[1429] Remarks made during the meeting

[1430] 1. A user says during a meeting, "You're the reason this project failed."

[1431] 2. The device records the audio and converts it into text. The speech recognition technology uses the Google Cloud Speech-to-Text API.

[1432] 3. The server analyzes the converted text using a generative model and determines that the phrase "It's your fault" is likely to be moral harassment.

[1433] 4. The server uses an emotion engine to recognize the user's emotion. For example, it determines that the user is angry.

[1434] 5. The server generates feedback and displays an alert saying, "Speaking in anger may come across as aggressive towards the other person. We recommend taking some time to calm down."

[1435] 6. The device provides generated feedback to the user as an alert notification, either via a pop-up alert or a notification bar.

[1436] Messages in chat

[1437] 1. A user sends a message in chat saying, "Your mistake is causing trouble for everyone."

[1438] 2. The device monitors chat content in real time and collects text data.

[1439] 3. The server analyzes the collected text using a generative model and determines that the phrase "Everyone is in trouble because of your mistake" is a possible example of moral harassment.

[1440] 4. The server uses an emotion engine to recognize the user's emotion. For example, it determines that the user is dissatisfied.

[1441] 5. The server generates feedback and displays an alert saying, "Sending this message out of frustration may come across as blaming. We recommend you send the message again after you have calmed down a bit."

[1442] 6. The device provides the generated feedback to the user as an alert notification, using the chat app's notification function.

[1443] This system not only prevents moral harassment in real time, but also contributes to reducing mental stress by providing appropriate feedback that takes into account the user's emotional state.The system also monitors users' comments and messages and provides immediate feedback, enabling early detection and resolution of problems.

[1444] The flow of the identification process in the second embodiment will be described with reference to FIG.

[1445] Step 1:

[1446] The server integrates a generative model and learns from past harassment cases.

[1447] As input, the generative model is fed with historical harassment case data, which is in text format and contains a large number of examples.

[1448] As part of data processing and data calculation, a generative model (e.g., a generative AI model) analyzes this text data and learns patterns of harassment. Specifically, the model processes the text data and performs a learning process to adjust weights and biases.

[1449] The output is a trained generative model.

[1450] Step 2:

[1451] The server integrates the emotion engine.

[1452] As input, you install the software for sentiment analysis on the server and enter the necessary API keys and settings.

[1453] As part of data processing and data calculation, the emotion engine (e.g., emotion analysis software) performs the initial setup and integration for analyzing emotions from voice and text data. Specific operations include configuring APIs and installing emotion analysis models.

[1454] As an output, we obtain a server with an integrated emotion engine.

[1455] Step 3:

[1456] The device installs the application and performs initial setup.

[1457] The input is to download and install the moral harassment prevention tool application onto your smartphone or PC.

[1458] For data processing and data calculation, the application obtains permission from the user to access voice recordings and chat logs. Specifically, the user operates the application installation screen and grants the necessary permissions.

[1459] The output will be an application that has been successfully installed and has the necessary permissions.

[1460] Step 4:

[1461] The device collects the voice data and converts it into text.

[1462] As input, the user's speech is recorded in real time through the device's microphone.

[1463] Data processing and calculation involves converting recorded voice data into text data using the Google Cloud Speech-to-Text API, etc. Specific operations include sending the voice data to a cloud server and converting it into text.

[1464] The output is the text data converted from the speech.

[1465] Step 5:

[1466] The device collects the text data entered and sends it to the server.

[1467] As input, the user enters a chat message or email.

[1468] For data processing and data calculation, the device monitors the user's input and collects text data in real time, which is temporarily stored in the device's memory.

[1469] As an output, we get the collected text data, which is then sent to a server using the HTTPS protocol.

[1470] Step 6:

[1471] The server analyzes the received text data.

[1472] As input, it receives text data sent from the terminal.

[1473] As part of the data processing and data calculation process, a generative model is used to analyze text data and detect words and phrases that may be related to moral harassment. Specifically, the model analyzes the text data and detects specific patterns.

[1474] The output is a list of phrases and words that are judged to be likely to be moral harassment.

[1475] Step 7:

[1476] The server performs sentiment analysis.

[1477] It takes parsed text data and audio data as input.

[1478] For data processing and data calculation, an emotion engine is used to analyze user emotions, including emotions such as anger and sadness. Specifically, an emotion analysis model is executed to extract emotions from text and voice data.

[1479] As an output, information about the user's emotional state is obtained.

[1480] Step 8:

[1481] The server generates the feedback.

[1482] As input, it receives the results of moral harassment detection and emotion analysis.

[1483] Based on the results of data processing and calculation, feedback is generated to notify the user. Specifically, appropriate advice or warning messages are created using specific algorithms and templates.

[1484] As an output, we get the generated feedback message.

[1485] Step 9:

[1486] The device will alert the user and suggest remedial measures.

[1487] As input, it receives the feedback message sent by the server.

[1488] As data processing and calculations are performed, alert notifications are displayed to the user. Notification methods include pop-up alerts and notification bars. Specific operations include using the device's notification function to provide real-time feedback to the user.

[1489] The output is an alert notification and remediation provided to the user.

[1490] Step 10:

[1491] The device stores the log.

[1492] Past detection results and notification contents are collected as input.

[1493] This data is then processed and calculated and stored in a database (e.g., SQLite) within the device. Specifically, the data is added to the database and made accessible at a later date.

[1494] As an output, you get the saved log data.

[1495] (Application example 2)

[1496] 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."

[1497] In physical stores, there is a need to prevent problems caused by inappropriate comments or emotional interactions during communication between staff and customers. However, conventional methods have difficulty analyzing comments and emotions in real time, making it difficult to provide appropriate feedback quickly. To solve this problem, a system is needed that can monitor and analyze the behavior of store staff in real time and provide appropriate feedback instantly.

[1498] The identification process 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 learning past harassment cases using a generative model and monitoring speech and behavior in real time, means for detecting words and sentences that may be harassing, means for issuing an alert to notify the user when such words or sentences are detected, means for analyzing emotions from voice and text data using an emotion engine, and means for providing appropriate remedial measures based on the analysis results. This makes it possible to improve the quality of communication between staff and customers in physical stores and prevent problems from occurring.

[1499] A "generative model" is an artificial intelligence technology that learns from past cases of harassment and analyzes the possibility of harassment in new data.

[1500] A "harassment case" is one in which inappropriate words, actions or behaviour cause mental distress or discomfort to a specific individual.

[1501] The "emotion engine" is a system for analyzing users' emotions from voice and text data.

[1502] "Real-time monitoring measures" are technologies for instantly collecting and analyzing ongoing audio and text data.

[1503] "Means of issuing an alert" refers to a method of notifying users when potential harassment is detected.

[1504] "Means for recording voice and converting it into text" refers to technology for recording speech and converting it into text data.

[1505] "Means of analysis" are methods for analyzing collected data to detect specific patterns or risks.

[1506] "Means for providing improvements" refers to methods for providing appropriate advice and action suggestions to users based on the analysis results.

[1507] System Configuration

[1508] This invention is a moral harassment prevention system that uses a generative model and an emotion engine, and is primarily composed of three parties: a server, a terminal, and a user. The server analyzes voice and text data collected from the terminal and aims to notify the user of an alert in real time. In addition, by combining it with an emotion engine, the system can analyze the user's emotions and provide more appropriate alerts and feedback.

[1509] Initial Setup

[1510] server

[1511] Generative model integration and training: The server uses a generative model to learn from past harassment cases, enabling analysis based on new data.

[1512] Emotion engine integration: Integrate an emotion engine into the server to recognize emotions from user voice and text data.

[1513] Database preparation: Prepare a database of past harassment cases and feed it to the generative model as data.

[1514] Terminal

[1515] Installing the app: Install the moral harassment prevention tool app on the user's smartphone or PC. The user must grant the necessary permissions (audio recording, text analysis, etc.).

[1516] Server collaboration: Collaborate with the server to download the generative model, emotion engine, and necessary data.

[1517] Real-time monitoring and alert notifications

[1518] Terminal

[1519] Voice recording and text conversion: Record the user's voice in real time using the device's microphone and convert it into text data using voice recognition technology.

[1520] Text data collection: The content of chat messages and emails entered by users is monitored in real time to collect text data.

[1521] Data transmission: Collected text data (including text converted from speech) is sent to the server.

[1522] server

[1523] Data analysis: The server analyzes the received text data using a generative model to detect words and phrases that may be moral harassment.

[1524] Emotion analysis: Using an emotion engine, we recognize emotions from the user's voice and text data, such as anger and sadness.

[1525] Feedback generation: Based on the detection results and sentiment analysis results, feedback is generated to notify the user.

[1526] Terminal

[1527] Alert notifications: Notify users of generated feedback in real time using methods such as pop-up alerts or notification bars.

[1528] Suggesting improvement measures: Presents specific improvement measures to the user. For example, based on the results of sentiment analysis, the system displays advice such as, "What you just said may be perceived as offensive to the other person. We recommend that you take a moment to calm down."

[1529] Saving logs: Past detection results and notification contents are saved as logs so that users can check them at a later date.

[1530] Specific examples

[1531] Example 1: Statements made during a meeting

[1532] 1. A user says during a meeting, "You're the reason this project failed."

[1533] 2. The device records the audio and converts it into text.

[1534] 3. The server analyzes the converted text using a generative model and determines that the phrase "It's your fault" is likely to be moral harassment.

[1535] 4. The server uses the emotion engine to recognize the user's emotion, for example, recognizing that the user has an emotion of anger.

[1536] 5. The server generates feedback and informs the user, "Speaking in anger may come across as aggressive towards the other person. We recommend taking some time to calm down."

[1537] 6. The device displays an alert to the user and suggests specific measures to improve the situation.

[1538] Example 2: Message in chat

[1539] 1. A user sends a message in chat saying, "Your mistake is causing trouble for everyone."

[1540] 2. The device monitors and collects chat content in real time.

[1541] 3. The server analyzes the collected text using a generative model and determines that the phrase "Everyone is in trouble because of your mistake" is a possible example of moral harassment.

[1542] 4. The server uses an emotion engine to recognize the user's emotion, for example, recognizing that the user is dissatisfied.

[1543] 5. The server generates feedback and informs the user, "Sending this message out of frustration may come across as blaming. We recommend that you send the message again after you have calmed down a bit."

[1544] 6. The device will alert the user and provide specific advice to encourage them to stay calm.

[1545] Prompt Sentence Examples

[1546] "Does this statement take into consideration your customer's feelings? I suggest you try phrasing it differently."

[1547] "The emotion recognition engine detects negative emotions. Please try to use as polite language as possible."

[1548] In this way, the system can improve the quality of communication between staff and customers in physical stores.

[1549] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[1550] Step 1:

[1551] The device records the user's voice in real time. Specifically, it uses the device's microphone to capture the user's speech and obtains the voice data. This voice data becomes the input.

[1552] Step 2:

[1553] The device converts the voice data it receives into text. It uses voice recognition technology to analyze the voice data and generate corresponding text data. This text data is then output.

[1554] Step 3:

[1555] The device sends the converted text data to the server. Specifically, the device uploads the text data to the server via the Internet. This text data becomes the input.

[1556] Step 4:

[1557] The server analyzes the received text data using a generative model. The generative model has learned from past harassment cases and analyzes new text data to detect potentially harassing words and phrases. The results of this analysis are output.

[1558] Step 5:

[1559] The server uses an emotion engine to perform emotion analysis based on the analysis results. Specifically, the server analyzes the user's emotions (anger, dissatisfaction, etc.) from the text data. The analysis results are output.

[1560] Step 6:

[1561] The server generates feedback based on the analysis results and sentiment analysis results. The generative model and sentiment analysis results are combined to generate appropriate advice or warnings for the user. This feedback is the output.

[1562] Step 7:

[1563] The server sends the generated feedback to the terminal. Specifically, the server sends the generated feedback to the terminal via the Internet. This feedback becomes the input.

[1564] Step 8:

[1565] The device displays an alert to the user. Specifically, the feedback content is displayed on the device display using a pop-up alert or a notification bar. This feedback content becomes the output.

[1566] Step 9:

[1567] The device presents the user with improvement measures. Based on the results of the emotion analysis, it might display advice such as, "What you just said may be perceived as offensive to the other person. We recommend that you take a moment to calm down." This advice is the output.

[1568] Step 10:

[1569] The device saves past detection results and notification details as logs. Detection results and feedback details are saved in local storage or cloud storage so that users can check them later. This allows users to look back on past comments and alert details.

[1570] 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.

[1571] 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.

[1572] 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.

[1573] [Fourth embodiment]

[1574] FIG. 7 shows an example of the configuration of a data processing system 410 according to the fourth embodiment.

[1575] 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.

[1576] 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).

[1577] 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.

[1578] 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.

[1579] 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).

[1580] 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.

[1581] 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.

[1582] 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.

[1583] 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.

[1584] 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.

[1585] 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.

[1586] 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."

[1587] System Configuration

[1588] This invention is a system for preventing moral harassment using generative models, which is mainly composed of three parties: a server, a terminal, and a user. The server analyzes voice and text data collected from the terminal and aims to notify the user of an alert in real time.

[1589] Initial Setup

[1590] server

[1591] Integration and training of generative models: The server uses generative models to learn from past harassment cases, enabling analysis based on new data.

[1592] Database preparation: Prepare a database of past harassment cases and provide it as data to the generative model.

[1593] Terminal

[1594] Installing the app: Install the moral harassment prevention tool app on the user's smartphone or PC. The user must grant the necessary permissions (audio recording, text analysis, etc.).

[1595] Server collaboration: Collaborate with the server to download the generative model and necessary data.

[1596] Real-time monitoring and alert notifications

[1597] Terminal

[1598] Voice recording and text conversion: Record the user's voice in real time using the device's microphone and convert it into text data using voice recognition technology.

[1599] Text data collection: The contents of chat messages and emails entered by users are monitored in real time to collect text data.

[1600] Data transmission: The collected text data is sent to the server.

[1601] server

[1602] Data analysis: The server analyzes the received text data using a generative model to detect words and phrases that may be indicative of moral harassment.

[1603] Feedback generation: Based on the detection results, feedback is generated to notify the user.

[1604] Terminal

[1605] Alert Notification: Notify users of generated feedback in real time using pop-up alerts, notification bars, etc.

[1606] Suggesting improvement measures: Providing specific improvement measures to the user, such as "We recommend you take a break to calm down temporarily."

[1607] Saving logs: Past detection results and notification contents are saved as logs so that users can check them at a later date.

[1608] Specific examples

[1609] Example 1: Speaking during a meeting

[1610] 1. A user says during a meeting, "You're the reason this project failed."

[1611] 2. The device records the audio and converts it into text.

[1612] 3. The server analyzes the converted text using a generative model and determines that the phrase "It's your fault" is likely to be moral harassment.

[1613] 4. The server generates feedback and notifies the user that "Comments that blame others increase the mental burden on the other person and may constitute moral harassment."

[1614] 5. The device displays an alert to the user and suggests specific measures to improve the situation.

[1615] Example 2: Message in chat

[1616] 1. A user sends a message in chat saying, "Your mistake is causing trouble for everyone."

[1617] 2. The device monitors and collects chat content in real time.

[1618] 3. The server analyzes the collected text using a generative model and determines that the phrase "Everyone is in trouble because of your mistake" is a possible example of moral harassment.

[1619] 4. The server generates feedback stating that "attacks aimed at identifying and blaming individuals may constitute moral harassment."

[1620] 5. The device displays an alert to the user and a message encouraging them to stay calm.

[1621] This system can detect in advance whether comments or messages may constitute moral harassment, and by encouraging users to be aware of this, it can contribute to preventing moral harassment.

[1622] The processing flow will be explained below.

[1623] Step 1:

[1624] Users install the moral harassment prevention tool app on their smartphone or PC and grant the necessary permissions (audio recording, text analysis, etc.).

[1625] Step 2:

[1626] The device works in conjunction with the server to download the generative model and necessary data.

[1627] Step 3:

[1628] The device uses a microphone to record the user's voice in real time and converts it into text data using voice recognition technology.

[1629] Step 4:

[1630] The terminal monitors the contents of chat messages and emails entered by users in real time and collects text data.

[1631] Step 5:

[1632] The terminal transmits the collected text data (including the text converted from the speech) to the server.

[1633] Step 6:

[1634] The server analyzes the received text data using a generative model to detect words and phrases that may be indicative of moral harassment.

[1635] Step 7:

[1636] The server generates feedback to notify the user based on the detection results.

[1637] Step 8:

[1638] The device notifies the user of the feedback received from the server in real time using methods such as pop-up alerts and a notification bar.

[1639] Step 9:

[1640] The device will then suggest specific measures to improve the situation to the user, such as "We recommend you take a break to calm down temporarily."

[1641] Step 10:

[1642] The device stores past detection results and notification contents as a log so that the user can check them at a later date.

[1643] Example 1

[1644] 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."

[1645] In modern society, moral harassment (morahara) in the workplace and at home has become a serious problem. Morahara not only damages people's mental health and reduces productivity, but in severe cases can even lead to legal issues. Conventional systems and tools have made it difficult to detect and prevent moral harassment, and real-time monitoring and notification have rarely been realized. This has made it difficult to respond quickly and appropriately when moral harassment occurs.

[1646] 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.

[1647] In this invention, the server includes means for learning past cases of harassment using a generative model and monitoring speech and behavior in real time, means for detecting potentially harassing expressions, means for issuing an alert to notify the user when such expressions are detected, means for collecting messages and voice data entered by the user, means for transmitting the collected data to the server, means for analyzing the data received by the server and generating feedback, means for notifying the user of the generated feedback, and means for recording and saving past detection results and notification content. This makes it possible to detect possible moral harassment in real time and promptly notify the user.

[1648] A "generative model" is a type of artificial intelligence that learns from past data and analyzes and generates new data.

[1649] "Means for monitoring speech and behavior" refers to means for detecting user statements and behavior in real time and recording or analyzing the content.

[1650] "Means for detecting expressions" refers to means for detecting specific words or phrases from audio or text data and evaluating whether there is a problem.

[1651] "Means for issuing alerts" are means for notifying or warning users about detected problematic behavior or inappropriate comments.

[1652] "Means for collecting message and voice data" means means for recording text messages and voice input by users in real time and retaining them for subsequent processing.

[1653] The "means for transmitting data to a server" refers to a means for transmitting data collected by a terminal to a server via a network.

[1654] "Means for analyzing data" refers to means for analyzing received data in detail and detecting specific patterns or anomalies.

[1655] The "means for generating feedback" refers to the means for generating and notifying the user of countermeasures or warnings based on the analysis results.

[1656] "Means for notifying the user of feedback" refers to means for displaying or notifying the user of the generated feedback in real time.

[1657] "Means for recording and saving" refers to a means for saving past detection results and notification contents in a database or the like so that they can be referenced at a later date.

[1658] This invention is a moral harassment prevention system that uses a generative model. This system is mainly composed of three entities: a server, a terminal, and a user. A specific embodiment of the system will be described below.

[1659] Hardware and Software Configuration

[1660] server

[1661] The server is constructed using the following hardware and software:

[1662] Hardware: A server machine equipped with a high-performance CPU, GPU, and large memory capacity

[1663] Software: Python programming language, TensorFlow or PyTorch, MySQL or PostgreSQL database

[1664] Terminal

[1665] The terminal is a communication device such as a user's smartphone or PC.

[1666] Hardware: Smartphones (iOS, Android), PCs (Windows, macOS)

[1667] Software: Moral harassment prevention tool app, Google Speech-to-Text API, REST API client

[1668] System configuration

[1669] The system is realized by combining the following means:

[1670] 1. Generative Model Integration and Training

[1671] The server uses a generative model (e.g., GPT-4) to learn from past harassment cases, enabling real-time analysis based on new data.

[1672] 2. Prepare the database

[1673] The server prepares a database of past harassment cases and supplies it to the generative model as data. The database is MySQL or PostgreSQL.

[1674] 3. Install the app

[1675] Install a moral harassment prevention tool app on the user's smartphone or PC. The app is compatible with iOS and Android, and the user must grant the necessary permissions (voice recording, text analysis, etc.).

[1676] 4. Server Integration

[1677] The device connects to the server and downloads the generative model and necessary data. The data is obtained using a REST API.

[1678] Examples of concrete examples and prompts

[1679] Example 1: Statements made during a meeting

[1680] 1. A user says during a meeting, "You're the reason this project failed."

[1681] 2. The device records the audio and converts it to text using the Google Speech-to-Text API.

[1682] 3. The terminal sends the converted text to the server.

[1683] 4. The server analyzes the received text using the GPT-4 generative model and determines that the phrase "It's your fault" is likely to be moral harassment.

[1684] 5. The server generates feedback and notifies the user that the comment may be emotionally abusive.

[1685] 6. The device displays an alert to the user saying, "A statement that may be moral harassment has been detected."

[1686] Example 2: Chat messages

[1687] 1. A user sends a message in chat saying, "Your mistake is causing trouble for everyone."

[1688] 2. The device monitors and collects chat content in real time.

[1689] 3. The device sends the collected text to the server.

[1690] 4. The server analyzes the received text using the BERT generative model and determines that the phrase "Everyone is suffering because of your mistake" is a possible example of moral harassment.

[1691] 5. The server generates feedback stating that "attacks aimed at identifying and blaming individuals may constitute moral harassment."

[1692] 6. The device displays an alert to the user saying, "A statement that may be moral harassment has been detected."

[1693] Examples of prompt statements

[1694] "Respond to harassing phrases"

[1695] "Please determine whether this sentence constitutes moral harassment."

[1696] This invention can contribute to preventing moral harassment by detecting in real time whether a user's comments or messages may constitute moral harassment and quickly notifying the user.

[1697] The flow of the identification process in the first embodiment will be described with reference to FIG.

[1698] Program processing flow

[1699] Step 1: Collecting audio data

[1700] Input: What the user says

[1701] Output: Collected audio data

[1702] Specific description:

[1703] The device collects the user's voice in real time through a microphone, specifically the built-in microphone of a smartphone or PC.

[1704] The app will launch the audio collection module and notify the user that "Audio recording has started."

[1705] Step 2: Convert audio data to text

[1706] Input: Collected audio data

[1707] Output: Converted text data

[1708] Specific description:

[1709] The device sends the collected voice data to the Google Speech-to-Text API, which converts the voice data into text.

[1710] Once the conversion is complete, the user is notified that "the audio data has been converted to text."

[1711] Step 3: Monitor and collect chat messages

[1712] Input: The chat message typed by the user

[1713] Output: Collected text messages

[1714] Specific description:

[1715] The device monitors the user's chat app (e.g., Slack or WhatsApp) and collects messages sent in real time.

[1716] Specifically, messages are retrieved using the chat app's API, and are temporarily stored on the device.

[1717] Step 4: Sending text data

[1718] Input: Collected text data

[1719] Output: Data sent to the server

[1720] Specific description:

[1721] The terminal transmits the collected text data to a server.

[1722] When sending, the data is AES encrypted to make it secure, and the user is notified that the data has been sent to the server.

[1723] Step 5: Data analysis

[1724] Input: Transmitted text data

[1725] Output: Analysis results

[1726] Specific description:

[1727] The text data received by the server is analyzed by a generative model (GPT-4 or BERT).

[1728] Specifically, to detect inappropriate words and phrases, prompts such as "Please respond to harassing phrases" are used.

[1729] When the processing is completed on the server, it records "Data analysis completed" in the log.

[1730] Step 6: Feedback generation

[1731] Input: Analysis results

[1732] Output: Generated feedback

[1733] Specific description:

[1734] The server generates feedback based on the analysis results of the generative model.

[1735] Specifically, it generates a notification message such as, "That comment may be emotional harassment. We recommend that you take a break to calm down."

[1736] Log "Feedback generated."

[1737] Step 7: Alert Notifications

[1738] Input: Generated feedback

[1739] Output: Alert notification to the user

[1740] Specific description:

[1741] The terminal notifies the user of the feedback received from the server in real time.

[1742] A pop-up alert or notification bar will be displayed to the user, informing them that "statements that may be emotionally abusive have been detected."

[1743] When a user acknowledges a notification, "User acknowledged notification" is logged.

[1744] Step 8: Propose improvement measures

[1745] Input: Generated feedback

[1746] Output: Present specific improvement measures to the user

[1747] Specific description:

[1748] The device will present specific improvement measures to the user.

[1749] For example, it will display advice such as, "We recommend you take a short walk to calm down."

[1750] When a user confirms a remediation, "User confirmed remediation" is logged.

[1751] Step 9: Save the logs

[1752] Input: Detection results and notification content

[1753] Output: Saved log data

[1754] Specific description:

[1755] The device stores past detection results and notification contents in an SQLite database.

[1756] After saving, you will be notified that "The log has been saved to the database."

[1757] Users can check this log at any time.

[1758] (Application example 1)

[1759] 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."

[1760] There is a need to prevent harassment that occurs in communication between employees in commercial facilities and improve the working environment. In particular, a system that allows employees to review their own behavior and take appropriate action through real-time monitoring and immediate feedback is needed, but this type of system has not yet been fully developed.

[1761] 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.

[1762] In this invention, the server includes means for learning from past cases of harassment using a generative model and monitoring speech and behavior in real time, means for detecting potentially harassing words and sentences, means for issuing an alert to notify users when such words or sentences are detected, means for recording the voices of employees at commercial facilities and converting them into text data, means for analyzing the converted text data, and means for generating feedback and notifying alerts and remedial measures. This makes it possible to immediately detect harassing behavior in real-time communication between employees and propose remedial measures, thereby ensuring a safe and comfortable working environment.

[1763] A "generative model" is a machine learning model that can generate new data based on past data.

[1764] "Harassment cases" refer to cases of conduct that cause psychological or physical distress to another person.

[1765] "Real-time monitoring means" are methods for instantly observing and recording current data and behavior.

[1766] An "alert mechanism" is a mechanism that sends a warning or notification to the user when certain conditions are met.

[1767] A "commercial facility" is a business establishment that has the facilities and space to offer goods and services.

[1768] "Means for recording employee voice and converting it into text data" refers to technology that records employee speech using a microphone or other device and converts that voice data into text information.

[1769] "Means for analyzing text data" refers to techniques for evaluating and analyzing textual information and extracting specific patterns and meanings.

[1770] "Means for generating feedback and notifying alerts and improvement measures" refers to methods for generating warnings and advice to users based on the analysis results and immediately communicating them.

[1771] This invention is a system for improving communication between employees in commercial facilities and preventing harassment. This system is composed of three entities: a server, a terminal, and a user.

[1772] server

[1773] The server uses a generative model to learn from past harassment cases and analyze new data. This allows it to analyze incoming data in real time and detect potentially abusive words and phrases. The server also generates feedback based on the analysis results and sends alerts and recommendations to the device.

[1774] Specifically, the server uses the following software and hardware:

[1775] Hardware: High-performance processor, memory, and large storage capacity

[1776] Software: Python, TensorFlow, Keras, SpeechRecognition library, Requests library

[1777] Terminal

[1778] The devices include the employee's smartphone or smart glasses. The devices have the ability to record voice data and convert it into text. The converted text data is then sent to the server in real time, and the received feedback is notified to the user.

[1779] The specific operation method is as follows:

[1780] 1. Record your voice using a microphone

[1781] 2. Convert the recorded audio into text using the SpeechRecognition library

[1782] 3. Send the converted text data to the server using the Requests library

[1783] 4. Receive feedback from the server and display it to the user

[1784] User

[1785] Users are primarily employees of commercial facilities, and their role is to record voice in real time, convert it into text data, and send it to a server. They also receive feedback from the server and correct their behavior.

[1786] Specific examples

[1787] Example of audio data

[1788] If an employee says something like "Did you make another mistake?", it will be handled as follows:

[1789] 1. Record audio using the device's microphone

[1790] 2. Convert recorded audio to text

[1791] 3. Send text data to the server

[1792] 4. The server analyzes and detects keywords such as "mistake."

[1793] 5. The server generates feedback and sends it to the device.

[1794] 6. The device will send an alert saying, "Comments that emphasize personal mistakes may lead to moral harassment."

[1795] 7. As an improvement measure, the message reads, "Work with other staff to identify problems and come up with solutions together."

[1796] Prompt Sentence Examples

[1797] "User sent text: 'Did you make another mistake?' Evaluate potential abuse and provide appropriate feedback and remediation."

[1798] In this way, this invention can prevent harassment between employees in commercial facilities. By providing real-time feedback, employees can immediately correct their behavior and promote healthy communication.

[1799] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[1800] Step 1:

[1801] The user records their voice using the microphone on their device. The input is the user's speech, and the output is voice data. Specifically, the user's words are recorded while they are talking to other employees in a commercial facility.

[1802] Step 2:

[1803] The device converts the recorded voice data into text. The input is voice data, and the output is text data that converts the voice into text information. Specifically, the voice-to-text conversion process is performed using the SpeechRecognition library.

[1804] Step 3:

[1805] The terminal sends text data to the server. The input is text data, and the output is an HTTP request to the server. Specifically, the Requests library is used to send text data to the server via a POST request.

[1806] Step 4:

[1807] The server analyzes the received text data. The input is the text data, and the output is the analysis result. Specifically, it uses a generative AI model to analyze the text data and detect words and phrases that may be abusive or harassing.

[1808] Step 5:

[1809] The server generates feedback based on the analysis results. The input is the analysis results, and the output is a feedback message. Specifically, based on the analysis results provided by the generative AI model, it creates feedback for the user that includes appropriate alerts and improvement measures.

[1810] Step 6:

[1811] The server sends the generated feedback to the terminal. The input is a feedback message, and the output is an HTTP response to the terminal. Specifically, data including the feedback message is sent to the terminal as an HTTP response.

[1812] Step 7:

[1813] The terminal notifies the user of the received feedback. The input is a feedback message, and the output is an alert or notification to the user. Specifically, the terminal displays the feedback message as a pop-up on the display or in the notification bar.

[1814] Step 8:

[1815] The user checks the notification and corrects their behavior. The input is the feedback message, and the output is the user's improved behavior. Specifically, the user understands the alert and the improvement measures notified to them, and corrects their statements and behavior based on them.

[1816] This allows real-time communication between employees within a commercial facility to be monitored, preventing harassment before it occurs.

[1817] 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.

[1818] System Configuration

[1819] This invention is a system for preventing moral harassment that uses a generative model and an emotion engine, and is primarily composed of three parties: a server, a device, and a user. The server analyzes voice and text data collected from the device and aims to notify the user of an alert in real time. In addition, by combining this with an emotion engine, the system can analyze the user's emotions and provide more appropriate alerts and feedback.

[1820] Initial Setup

[1821] server

[1822] Integration and training of generative models: The server uses generative models to learn from past harassment cases, enabling analysis based on new data.

[1823] Emotion engine integration: An emotion engine is integrated into the server to recognize emotions from user voice and text data.

[1824] Database preparation: Prepare a database of past harassment cases and provide it as data to the generative model.

[1825] Terminal

[1826] Installing the app: Install the moral harassment prevention tool app on the user's smartphone or PC. The user must grant the necessary permissions (audio recording, text analysis, etc.).

[1827] Server collaboration: Collaborate with the server to download the generative model, emotion engine, and necessary data.

[1828] Real-time monitoring and alert notifications

[1829] Terminal

[1830] Voice recording and text conversion: Record the user's voice in real time using the device's microphone and convert it into text data using voice recognition technology.

[1831] Text data collection: The contents of chat messages and emails entered by users are monitored in real time to collect text data.

[1832] Data transmission: The collected text data (including the text converted from the speech) is sent to the server.

[1833] server

[1834] Data analysis: The server analyzes the received text data using a generative model to detect words and phrases that may be indicative of moral harassment.

[1835] Emotion analysis: Using an emotion engine, we recognize emotions from the user's voice and text data, such as anger and sadness.

[1836] Feedback generation: Based on the detection results and sentiment analysis results, feedback is generated to notify the user.

[1837] Terminal

[1838] Alert notification: The generated feedback is notified to the user in real time using methods such as pop-up alerts and notification bars.

[1839] Suggesting improvement measures: Presents specific improvement measures to the user. For example, based on the results of sentiment analysis, the system displays advice such as, "What you just said may be perceived as offensive to the other person. We recommend that you take some time to calm down."

[1840] Saving logs: Past detection results and notification contents are saved as logs so that users can check them at a later date.

[1841] Specific examples

[1842] Example 1: Speaking during a meeting

[1843] 1. A user says during a meeting, "You're the reason this project failed."

[1844] 2. The device records the audio and converts it into text.

[1845] 3. The server analyzes the converted text using a generative model and determines that the phrase "It's your fault" is likely to be moral harassment.

[1846] 4. The server uses the emotion engine to recognize the user's emotion, for example, recognizing that the user has an emotion of anger.

[1847] 5. The server generates feedback and informs the user, "Speaking in anger may come across as aggressive towards the other person. We recommend taking some time to calm down."

[1848] 6. The device displays an alert to the user and suggests specific measures to improve the situation.

[1849] Example 2: Message in chat

[1850] 1. A user sends a message in chat saying, "Your mistake is causing trouble for everyone."

[1851] 2. The device monitors and collects chat content in real time.

[1852] 3. The server analyzes the collected text using a generative model and determines that the phrase "Everyone is in trouble because of your mistake" is a possible example of moral harassment.

[1853] 4. The server uses the emotion engine to recognize the user's emotion, for example, recognizing that the user is dissatisfied.

[1854] 5. The server generates feedback and informs the user, "Sending this message out of frustration may come across as blaming. We recommend that you send the message again after you have calmed down a bit."

[1855] 6. The device will alert the user and provide specific advice to encourage them to stay calm.

[1856] This system can further contribute to preventing moral harassment by detecting in advance comments or messages that may constitute moral harassment and providing feedback and improvement measures that take into account the user's emotional state.

[1857] The processing flow will be explained below.

[1858] System Configuration

[1859] This invention is a system for preventing moral harassment that uses a generative model and an emotion engine, and is primarily composed of three parties: a server, a device, and a user. The server analyzes voice and text data collected from the device and aims to notify the user of an alert in real time. In addition, by combining this with an emotion engine, the system can recognize the user's emotions and provide more appropriate alerts and feedback.

[1860] Initial Setup

[1861] Step 1:

[1862] The server uses a generative model to learn from past harassment cases, allowing it to analyze new data.

[1863] Step 2:

[1864] The server integrates an emotion engine to recognize emotions from the user's voice and text data.

[1865] Step 3:

[1866] The server prepares a database of past harassment cases and supplies it as data to the generative model.

[1867] Step 4:

[1868] The device involves installing a moral harassment prevention tool app on the user's smartphone or PC, and the user granting the necessary permissions (voice recording, text analysis, etc.).

[1869] Step 5:

[1870] The device connects to the server and downloads the generative model, emotion engine, and necessary data.

[1871] Real-time monitoring and alert notifications

[1872] Step 6:

[1873] The device uses a microphone to record the user's voice in real time and converts it into text data using voice recognition technology.

[1874] Step 7:

[1875] The terminal monitors the contents of chat messages and emails entered by users in real time and collects text data.

[1876] Step 8:

[1877] The terminal transmits the collected text data (including the text converted from the speech) to the server.

[1878] Step 9:

[1879] The server analyzes the received text data using a generative model to detect words and phrases that may be indicative of moral harassment.

[1880] Step 10:

[1881] The server uses an emotion engine to recognize emotions from the user's voice and text data, such as anger and sadness.

[1882] Step 11:

[1883] The server generates feedback to notify the user based on the analysis results of the generative model and the emotion analysis results.

[1884] Step 12:

[1885] The device notifies the user of the feedback received from the server in real time using methods such as pop-up alerts and a notification bar.

[1886] Step 13:

[1887] The device will then present the user with specific advice on how to improve the situation, such as "What you just said may be perceived as offensive to the other person. We recommend that you take a moment to calm down."

[1888] Step 14:

[1889] The device stores past detection results and notification contents as a log so that the user can check them at a later date.

[1890] Specific examples

[1891] Example 1: Speaking during a meeting

[1892] Step 1:

[1893] A user says during a meeting, "You are the reason this project failed."

[1894] Step 2:

[1895] The device will record the audio and convert it into text.

[1896] Step 3:

[1897] The server analyzes the converted text using a generative model and determines that the phrase "It's your fault" is likely to be moral harassment.

[1898] Step 4:

[1899] The server uses an emotion engine to recognize the user's emotion, for example, recognizing that the user has an emotion of anger.

[1900] Step 5:

[1901] The server generates feedback, informing the user that "Speaking in anger may come across as aggressive towards the other person. We recommend that you take some time to calm down."

[1902] Step 6:

[1903] The device will display an alert to the user and suggest specific remedial measures.

[1904] Example 2: Message in chat

[1905] Step 1:

[1906] A user sends a message in chat saying, "Your mistake is causing trouble for everyone."

[1907] Step 2:

[1908] The device monitors and collects chat content in real time.

[1909] Step 3:

[1910] The server analyzes the collected text using a generative model and determines that the phrase "Everyone is in trouble because of your mistake" may be an example of moral harassment.

[1911] Step 4:

[1912] The server uses an emotion engine to recognize the user's emotion, for example, recognizing that the user is dissatisfied.

[1913] Step 5:

[1914] The server generates feedback, informing the user, "Sending this message out of frustration may come across as blaming. We recommend that you send the message again after you have calmed down a bit."

[1915] Step 6:

[1916] The device will display an alert to the user and provide specific advice to encourage them to stay calm.

[1917] In this way, a system that combines a generative model and an emotion engine can detect in advance whether a statement or message may constitute moral harassment, encourage the user to be aware of it, and provide appropriate feedback and improvement measures tailored to the user's emotional state.

[1918] Example 2

[1919] 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."

[1920] In modern society, moral harassment (moral harassment) is becoming a serious problem in various settings, including the workplace and school. This not only causes victims psychological damage, stress, and anxiety, but also reduces their productivity and quality of life. While existing moral harassment prevention systems can address certain cases, they are not considered effective preventative measures due to insufficient real-time monitoring, analysis, and feedback. Therefore, there is a need for a more advanced and comprehensive moral harassment prevention system that can reduce stress in users' daily lives and work and improve their mental health.

[1921] The specific processing by the specific processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means.

[1922] In this invention, the server includes means for learning from past harassment cases using a generative model and monitoring speech and behavior in real time, means for detecting potentially harassing words and sentences, means for issuing an alert to notify the user when such words or sentences are detected, means for integrating an emotion engine that analyzes the user's emotions, and means for generating feedback based on the emotion analysis results. This makes it possible to detect in real time whether a statement or message may constitute moral harassment and provide appropriate feedback and improvement measures.

[1923] A "generative model" is a type of artificial intelligence that learns from past data and generates appropriate outputs for new data.

[1924] "Harassment cases" are specific cases of moral harassment or psychological aggression that have occurred in the past.

[1925] "Means of real-time monitoring" refers to technology that allows for immediate monitoring of users' behavior and input.

[1926] "Word and phrase detection methods" are technologies that identify specific signs of harassment in users' statements and messages.

[1927] An "alert mechanism" is a mechanism for notifying users of detected potential harassment.

[1928] An "emotion engine" is a system for analyzing user emotions from voice and text data.

[1929] "Means for generating feedback" refers to technology that creates advice for users based on the results of detected harassment and sentiment analysis.

[1930] This invention is a system for preventing moral harassment that uses a generative model and an emotion engine to monitor users' behavior in real time and provide appropriate feedback. This system is mainly composed of three parties: a server, a terminal, and a user.

[1931] System Configuration

[1932] server

[1933] 1. Generative model integration and training:

[1934] The server uses a generative model to learn from past harassment cases. The specific generative model used is a "generative AI model," a general language model. This model learns patterns from past case data and can be applied to new situations.

[1935] 2. Emotion engine integration:

[1936] The server will be integrated with Emotion Engine, a general-purpose emotion analysis software, which will enable real-time analysis of emotions from user voice and text data, enabling more appropriate alerts and feedback to be provided.

[1937] 3. Prepare the database:

[1938] The server prepares a database that accumulates past cases of harassment. This database contains text data and emotion tags, and is used to analyze cases of moral harassment. The specific database management system used is MongoDB or MySQL.

[1939] Terminal

[1940] 1. Install the app:

[1941] Users install the anti-moral harassment app on their smartphones or PCs. The app asks the user for permission to record voice and monitor chat logs, and once permission is granted, the system will function normally.

[1942] 2. Server integration:

[1943] The device connects to the server to download data for the generative model and emotion engine, using secure protocols such as HTTPS for communication.

[1944] Specific examples

[1945] Remarks made during the meeting

[1946] 1. A user says during a meeting, "You're the reason this project failed."

[1947] 2. The device records the audio and converts it into text. The speech recognition technology uses the Google Cloud Speech-to-Text API.

[1948] 3. The server analyzes the converted text using a generative model and determines that the phrase "It's your fault" is likely to be moral harassment.

[1949] 4. The server uses an emotion engine to recognize the user's emotion. For example, it determines that the user is angry.

[1950] 5. The server generates feedback and displays an alert saying, "Speaking in anger may come across as aggressive towards the other person. We recommend taking some time to calm down."

[1951] 6. The device provides generated feedback to the user as an alert notification, either via a pop-up alert or a notification bar.

[1952] Messages in chat

[1953] 1. A user sends a message in chat saying, "Your mistake is causing trouble for everyone."

[1954] 2. The device monitors chat content in real time and collects text data.

[1955] 3. The server analyzes the collected text using a generative model and determines that the phrase "Everyone is in trouble because of your mistake" is a possible example of moral harassment.

[1956] 4. The server uses an emotion engine to recognize the user's emotion. For example, it determines that the user is dissatisfied.

[1957] 5. The server generates feedback and displays an alert saying, "Sending this message out of frustration may come across as blaming. We recommend you send the message again after you have calmed down a bit."

[1958] 6. The device provides the generated feedback to the user as an alert notification, using the chat app's notification function.

[1959] This system not only prevents moral harassment in real time, but also contributes to reducing mental stress by providing appropriate feedback that takes into account the user's emotional state.The system also monitors users' comments and messages and provides immediate feedback, enabling early detection and resolution of problems.

[1960] The flow of the identification process in the second embodiment will be described with reference to FIG.

[1961] Step 1:

[1962] The server integrates a generative model and learns from past harassment cases.

[1963] As input, the generative model is fed with historical harassment case data, which is in text format and contains a large number of examples.

[1964] As part of data processing and data calculation, a generative model (e.g., a generative AI model) analyzes this text data and learns patterns of harassment. Specifically, the model processes the text data and performs a learning process to adjust weights and biases.

[1965] The output is a trained generative model.

[1966] Step 2:

[1967] The server integrates the emotion engine.

[1968] As input, you install the software for sentiment analysis on the server and enter the necessary API keys and settings.

[1969] As part of data processing and data calculation, the emotion engine (e.g., emotion analysis software) performs the initial setup and integration for analyzing emotions from voice and text data. Specific operations include configuring APIs and installing emotion analysis models.

[1970] As an output, we obtain a server with an integrated emotion engine.

[1971] Step 3:

[1972] The device installs the application and performs initial setup.

[1973] The input is to download and install the moral harassment prevention tool application onto your smartphone or PC.

[1974] For data processing and data calculation, the application obtains permission from the user to access voice recordings and chat logs. Specifically, the user operates the application installation screen and grants the necessary permissions.

[1975] The output will be an application that has been successfully installed and has the necessary permissions.

[1976] Step 4:

[1977] The device collects the voice data and converts it into text.

[1978] As input, the user's speech is recorded in real time through the device's microphone.

[1979] Data processing and calculation involves converting recorded voice data into text data using the Google Cloud Speech-to-Text API, etc. Specific operations include sending the voice data to a cloud server and converting it into text.

[1980] The output is the text data converted from the speech.

[1981] Step 5:

[1982] The device collects the text data entered and sends it to the server.

[1983] As input, the user enters a chat message or email.

[1984] For data processing and data calculation, the device monitors the user's input and collects text data in real time, which is temporarily stored in the device's memory.

[1985] As an output, we get the collected text data, which is then sent to a server using the HTTPS protocol.

[1986] Step 6:

[1987] The server analyzes the received text data.

[1988] As input, it receives text data sent from the terminal.

[1989] As part of the data processing and data calculation process, a generative model is used to analyze text data and detect words and phrases that may be related to moral harassment. Specifically, the model analyzes the text data and detects specific patterns.

[1990] The output is a list of phrases and words that are judged to be likely to be moral harassment.

[1991] Step 7:

[1992] The server performs sentiment analysis.

[1993] It takes parsed text data and audio data as input.

[1994] For data processing and data calculation, an emotion engine is used to analyze user emotions, including emotions such as anger and sadness. Specifically, an emotion analysis model is executed to extract emotions from text and voice data.

[1995] As an output, information about the user's emotional state is obtained.

[1996] Step 8:

[1997] The server generates the feedback.

[1998] As input, it receives the results of moral harassment detection and emotion analysis.

[1999] Based on the results of data processing and calculation, feedback is generated to notify the user. Specifically, appropriate advice or warning messages are created using specific algorithms and templates.

[2000] As an output, we get the generated feedback message.

[2001] Step 9:

[2002] The device will alert the user and suggest remedial measures.

[2003] As input, it receives the feedback message sent by the server.

[2004] As data processing and calculations are performed, alert notifications are displayed to the user. Notification methods include pop-up alerts and notification bars. Specific operations include using the device's notification function to provide real-time feedback to the user.

[2005] The output is an alert notification and remediation provided to the user.

[2006] Step 10:

[2007] The device stores the log.

[2008] Past detection results and notification contents are collected as input.

[2009] This data is then processed and calculated and stored in a database (e.g., SQLite) within the device. Specifically, the data is added to the database and made accessible at a later date.

[2010] As an output, you get the saved log data.

[2011] (Application example 2)

[2012] 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."

[2013] In physical stores, there is a need to prevent problems caused by inappropriate comments or emotional interactions during communication between staff and customers. However, conventional methods have difficulty analyzing comments and emotions in real time, making it difficult to provide appropriate feedback quickly. To solve this problem, a system is needed that can monitor and analyze the behavior of store staff in real time and provide appropriate feedback instantly.

[2014] The identification process 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 learning past harassment cases using a generative model and monitoring speech and behavior in real time, means for detecting words and sentences that may be harassing, means for issuing an alert to notify the user when such words or sentences are detected, means for analyzing emotions from voice and text data using an emotion engine, and means for providing appropriate remedial measures based on the analysis results. This makes it possible to improve the quality of communication between staff and customers in physical stores and prevent problems from occurring.

[2015] A "generative model" is an artificial intelligence technology that learns from past cases of harassment and analyzes the possibility of harassment in new data.

[2016] A "harassment case" is one in which inappropriate words, actions or behaviour cause mental distress or discomfort to a specific individual.

[2017] The "emotion engine" is a system for analyzing users' emotions from voice and text data.

[2018] "Real-time monitoring measures" are technologies for instantly collecting and analyzing ongoing audio and text data.

[2019] "Means of issuing an alert" refers to a method of notifying users when potential harassment is detected.

[2020] "Means for recording voice and converting it into text" refers to technology for recording speech and converting it into text data.

[2021] "Means of analysis" are methods for analyzing collected data to detect specific patterns or risks.

[2022] "Means for providing improvements" refers to methods for providing appropriate advice and action suggestions to users based on the analysis results.

[2023] System Configuration

[2024] This invention is a moral harassment prevention system that uses a generative model and an emotion engine, and is primarily composed of three parties: a server, a terminal, and a user. The server analyzes voice and text data collected from the terminal and aims to notify the user of an alert in real time. In addition, by combining it with an emotion engine, the system can analyze the user's emotions and provide more appropriate alerts and feedback.

[2025] Initial Setup

[2026] server

[2027] Generative model integration and training: The server uses a generative model to learn from past harassment cases, enabling analysis based on new data.

[2028] Emotion engine integration: Integrate an emotion engine into the server to recognize emotions from user voice and text data.

[2029] Database preparation: Prepare a database of past harassment cases and feed it to the generative model as data.

[2030] Terminal

[2031] Installing the app: Install the moral harassment prevention tool app on the user's smartphone or PC. The user must grant the necessary permissions (audio recording, text analysis, etc.).

[2032] Server collaboration: Collaborate with the server to download the generative model, emotion engine, and necessary data.

[2033] Real-time monitoring and alert notifications

[2034] Terminal

[2035] Voice recording and text conversion: Record the user's voice in real time using the device's microphone and convert it into text data using voice recognition technology.

[2036] Text data collection: The content of chat messages and emails entered by users is monitored in real time to collect text data.

[2037] Data transmission: Collected text data (including text converted from speech) is sent to the server.

[2038] server

[2039] Data analysis: The server analyzes the received text data using a generative model to detect words and phrases that may be moral harassment.

[2040] Emotion analysis: Using an emotion engine, we recognize emotions from the user's voice and text data, such as anger and sadness.

[2041] Feedback generation: Based on the detection results and sentiment analysis results, feedback is generated to notify the user.

[2042] Terminal

[2043] Alert notifications: Notify users of generated feedback in real time using methods such as pop-up alerts or notification bars.

[2044] Suggesting improvement measures: Presents specific improvement measures to the user. For example, based on the results of sentiment analysis, the system displays advice such as, "What you just said may be perceived as offensive to the other person. We recommend that you take a moment to calm down."

[2045] Saving logs: Past detection results and notification contents are saved as logs so that users can check them at a later date.

[2046] Specific examples

[2047] Example 1: Statements made during a meeting

[2048] 1. A user says during a meeting, "You're the reason this project failed."

[2049] 2. The device records the audio and converts it into text.

[2050] 3. The server analyzes the converted text using a generative model and determines that the phrase "It's your fault" is likely to be moral harassment.

[2051] 4. The server uses the emotion engine to recognize the user's emotion, for example, recognizing that the user has an emotion of anger.

[2052] 5. The server generates feedback and informs the user, "Speaking in anger may come across as aggressive towards the other person. We recommend taking some time to calm down."

[2053] 6. The device displays an alert to the user and suggests specific measures to improve the situation.

[2054] Example 2: Message in chat

[2055] 1. A user sends a message in chat saying, "Your mistake is causing trouble for everyone."

[2056] 2. The device monitors and collects chat content in real time.

[2057] 3. The server analyzes the collected text using a generative model and determines that the phrase "Everyone is in trouble because of your mistake" is a possible example of moral harassment.

[2058] 4. The server uses an emotion engine to recognize the user's emotion, for example, recognizing that the user is dissatisfied.

[2059] 5. The server generates feedback and informs the user, "Sending this message out of frustration may come across as blaming. We recommend that you send the message again after you have calmed down a bit."

[2060] 6. The device will alert the user and provide specific advice to encourage them to stay calm.

[2061] Prompt Sentence Examples

[2062] "Does this statement take into consideration your customer's feelings? I suggest you try phrasing it differently."

[2063] "The emotion recognition engine detects negative emotions. Please try to use as polite language as possible."

[2064] In this way, the system can improve the quality of communication between staff and customers in physical stores.

[2065] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[2066] Step 1:

[2067] The device records the user's voice in real time. Specifically, it uses the device's microphone to capture the user's speech and obtains the voice data. This voice data becomes the input.

[2068] Step 2:

[2069] The device converts the voice data it receives into text. It uses voice recognition technology to analyze the voice data and generate corresponding text data. This text data is then output.

[2070] Step 3:

[2071] The device sends the converted text data to the server. Specifically, the device uploads the text data to the server via the Internet. This text data becomes the input.

[2072] Step 4:

[2073] The server analyzes the received text data using a generative model. The generative model has learned from past harassment cases and analyzes new text data to detect potentially harassing words and phrases. The results of this analysis are output.

[2074] Step 5:

[2075] The server uses an emotion engine to perform emotion analysis based on the analysis results. Specifically, the server analyzes the user's emotions (anger, dissatisfaction, etc.) from the text data. The analysis results are output.

[2076] Step 6:

[2077] The server generates feedback based on the analysis results and sentiment analysis results. The generative model and sentiment analysis results are combined to generate appropriate advice or warnings for the user. This feedback is the output.

[2078] Step 7:

[2079] The server sends the generated feedback to the terminal. Specifically, the server sends the generated feedback to the terminal via the Internet. This feedback becomes the input.

[2080] Step 8:

[2081] The device displays an alert to the user. Specifically, the feedback content is displayed on the device display using a pop-up alert or a notification bar. This feedback content becomes the output.

[2082] Step 9:

[2083] The device presents the user with improvement measures. Based on the results of the emotion analysis, it might display advice such as, "What you just said may be perceived as offensive to the other person. We recommend that you take a moment to calm down." This advice is the output.

[2084] Step 10:

[2085] The device saves past detection results and notification details as logs. Detection results and feedback details are saved in local storage or cloud storage so that users can check them later. This allows users to look back on past comments and alert details.

[2086] 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.

[2087] 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.

[2088] 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.

[2089] 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.

[2090] 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.

[2091] 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.

[2092] 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).

[2093] 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.

[2094] 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."

[2095] 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.

[2096] 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).

[2097] 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.

[2098] 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.

[2099] 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.

[2100] 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.

[2101] 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.

[2102] 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.

[2103] 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.

[2104] 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.

[2105] 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.

[2106] 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.

[2107] The following is further disclosed regarding the above embodiment.

[2108] (Claim 1)

[2109] A means to monitor behavior in real time by learning from past harassment cases using a generative model,

[2110] A means of detecting potentially abusive words and phrases;

[2111] A means of issuing an alert to notify the user if detected;

[2112] A system including:

[2113] (Claim 2)

[2114] A means of recording the user's voice and converting it into text;

[2115] means for analyzing the converted text data;

[2116] 10. The system of claim 1.

[2117] (Claim 3)

[2118] a means for monitoring text messages entered by users in real time;

[2119] means for transmitting the collected text data to a server;

[2120] 10. The system of claim 1.

[2121] (Claim 4)

[2122] A means for generating feedback based on the analysis results of the generative model;

[2123] a means of notifying users of their feedback;

[2124] 10. The system of claim 1.

[2125] (Claim 5)

[2126] A means of presenting specific improvement measures to users, and

[2127] A means of saving past detection results as a log;

[2128] 10. The system of claim 1.

[2129] "Example 1"

[2130] (Claim 1)

[2131] A means to learn from past harassment cases using a generative model and monitor behavior in real time;

[2132] A means of detecting potentially abusive speech; and

[2133] A means of issuing an alert to notify the user if detected;

[2134] A means of collecting user-typed messages and voice data;

[2135] means for transmitting the collected data to a server;

[2136] a means for analyzing the received data at the server and generating feedback;

[2137] a means of notifying the user of the generated feedback;

[2138] A means to record and save past detection results and notification contents,

[2139] ...

[2140] A system including:

[2141] (Claim 2)

[2142] A means of recording the user's voice and converting it into text;

[2143] means for transmitting the converted text data to a server and analyzing the data;

[2144] 10. The system of claim 1.

[2145] (Claim 3)

[2146] a means for monitoring text messages entered by users in real time;

[2147] means for transmitting the collected text data to a server;

[2148] means for generating feedback based on the parsed results at the server;

[2149] 10. The system of claim 1.

[2150] "Application Example 1"

[2151] (Claim 1)

[2152] A means to monitor behavior in real time by learning from past harassment cases using a generative model,

[2153] A means of detecting potentially abusive words and phrases;

[2154] A means of issuing an alert to notify the user if detected;

[2155] A means for recording the voices of employees at a commercial facility and converting the voices into text data;

[2156] A means for analyzing the converted text data and sending an alert;

[2157] A system including:

[2158] (Claim 2)

[2159] A means of recording the user's voice and converting it into text;

[2160] means for analyzing the converted text data;

[2161] means for transmitting data to a server and generating feedback;

[2162] A means of communicating alerts and remediation measures;

[2163] 10. The system of claim 1.

[2164] (Claim 3)

[2165] a means for monitoring text messages entered by users in real time;

[2166] means for transmitting the collected text data to a server;

[2167] A means of generating feedback and communicating it in real time;

[2168] 10. The system of claim 1.

[2169] "Example 2: Combining Emotion Engines"

[2170] (Claim 1)

[2171] A means to monitor behavior in real time by learning from past harassment cases using a generative model,

[2172] A means of detecting potentially abusive words and phrases;

[2173] A means of issuing an alert to notify the user if detected;

[2174] means for integrating an emotion engine for analyzing user emotions;

[2175] A means for generating feedback based on the sentiment analysis result;

[2176] A system including:

[2177] (Claim 2)

[2178] A means of recording the user's voice and converting it into text;

[2179] means for analyzing the converted text data;

[2180] 10. The system of claim 1.

[2181] (Claim 3)

[2182] a means for monitoring text messages entered by users in real time;

[2183] means for transmitting the collected text data to a server;

[2184] 10. The system of claim 1.

[2185] "Application example 2 when combining emotion engines"

[2186] (Claim 1)

[2187] A means to monitor behavior in real time by learning from past harassment cases using a generative model,

[2188] A means of detecting potentially abusive words and phrases;

[2189] a means for issuing an alert to notify users when detected;

[2190] A means for analyzing emotions from voice and text data using an emotion engine;

[2191] A means of providing appropriate remedial measures based on the analysis results;

[2192] A system including:

[2193] (Claim 2)

[2194] a means for recording the user's voice and converting it into text;

[2195] means for analyzing the converted text data;

[2196] 10. The system of claim 1.

[2197] (Claim 3)

[2198] a means for monitoring text messages entered by a user in real time;

[2199] means for transmitting the collected text data to a server;

[2200] 10. The system of claim 1. [Explanation of symbols]

[2201] 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 to monitor behavior in real time by learning from past harassment cases using a generative model, A means of detecting potentially abusive words and phrases; A means of issuing an alert to notify the user if detected; A system including:

2. A means of recording the user's voice and converting it into text; means for analyzing the converted text data; The system of claim 1 .

3. a means for monitoring text messages entered by users in real time; means for transmitting the collected text data to a server; The system of claim 1 .

4. A means for generating feedback based on the analysis results of the generative model; a means of notifying users of their feedback; The system of claim 1 .

5. A means of presenting specific improvement measures to users, and A means of saving past detection results as a log; The system of claim 1 .

Images