system

Abstract

Provide a system. A means for users to upload their own documents; means for converting said document into text; means for analyzing the text using natural language processing techniques; A method to calculate the copy-paste rate by comparing it with information on the Internet, a means for determining characteristics of the generative artificial intelligence; means for calculating a generation artificial intelligence rate based on the copy-paste rate and the determination result; A means for providing the results of the generated AI algorithm to a user; A system including:

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] With the evolution of generative AI (generative artificial intelligence), there has been an increase in cases where users use generative AI to create documents such as resumes and school reports, rather than creating them themselves. In such situations, it has become difficult to determine the legitimacy and originality of documents, posing a major problem, particularly for educational institutions and employers. The present invention aims to accurately assess such documents created using generative AI and provide a highly reliable evaluation. It also aims to provide a system for preventing fraudulent use of generative AI and evaluating legitimate learning and effort. [Means for solving the problem]

[0005] The present invention solves the above-mentioned problems by providing the following means. The system includes a means for a user to upload their own document, a means for converting the document into text, a means for analyzing the text using natural language processing technology, a means for calculating the copy-paste rate by comparing it with information on the Internet, a means for determining the characteristics of a generating AI, a means for calculating the generating AI rate based on the copy-paste rate and the determination result, and a means for providing the user with the result of the generating AI rate. The system also includes a means for notifying the user of resubmission or rejection if the generating AI rate exceeds a specific threshold, thereby enabling a highly reliable determination of the document's authenticity. Furthermore, the copy-paste rate is calculated by comparing the text with multiple information sources on the Internet, and by referencing various databases on the Internet, the originality of the document can be evaluated more accurately.

[0006] "User" means any person or organization that uses the System to upload its Documents.

[0007] "Document" refers to a file containing information in text format, such as a resume or school report.

[0008] "Upload method" refers to the interface or process by which a user submits a document to the system.

[0009] "Conversion Method" refers to the technology or process that converts uploaded document files into text format.

[0010] "Natural language processing technology" refers to computer science techniques for analyzing and understanding the sentence structure and meaning of text.

[0011] The "copy-paste rate" refers to the percentage of analyzed text that matches existing information on the Internet.

[0012] "Generative AI characteristics" refer to document characteristics that are likely to have been created by generative AI, such as a particular writing style or excessive use of keywords.

[0013] "Generative AI rate" refers to the overall probability that a document was created by a generative AI.

[0014] "Means of provision" refers to the interface or process for displaying or notifying users of the analysis results.

[0015] "Information on the Internet" refers to text data and information sources that are publicly available on the Web. [Brief explanation of the drawings]

[0016] [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

[0017] 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.

[0018] First, the terms used in the following description will be explained.

[0019] 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).

[0020] 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.

[0021] 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.

[0022] 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.

[0023] 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."

[0024] [First embodiment]

[0025] FIG. 1 shows an example of the configuration of a data processing system 10 according to the first embodiment.

[0026] 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.

[0027] 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).

[0028] 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.

[0029] 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.

[0030] 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.

[0031] 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.

[0032] FIG. 2 shows an example of the main functions of the data processing device 12 and the smart device 14.

[0033] 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.

[0034] 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.

[0035] 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.

[0036] 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."

[0037] This invention relates to a system that determines whether a document, such as a resume or a school report, was created by generative AI (generative artificial intelligence). The system performs a series of steps to determine whether the document was created by a user who uploads the document and analyzes it.

[0038] System Overview

[0039] 1. User uploads a document

[0040] Users upload their resumes and reports on the system's web interface (client terminal). Upload methods include a file selection button and meta-information input fields (university name, subjects, etc.).

[0041] Specific examples

[0042] A user accesses the web page, selects "Report.docx" in the upload form, enters the university name "General University" and the subject "Information Technology," and presses the upload button to send the document and information to the server.

[0043] 2. The server receives the file and converts it to text

[0044] The server receives the file sent by the user and converts the document from PDF or DOCX format to text format, for example using a PDF to text converter or a DOCX parser.

[0045] Specific examples

[0046] The server receives the "report.docx" and converts it to text using an appropriate library (e.g., Apache POI), which stores the converted text in memory.

[0047] 3. The server parses the text

[0048] The server applies natural language processing techniques to the converted text to analyze sentence structure and features, which may involve the use of NLP libraries (e.g., spaCy, NLTK).

[0049] Specific examples

[0050] The server analyzes the converted text and extracts grammatical elements such as the subject, predicate, and object of each sentence. It also evaluates the writing style and tone and calculates features (word frequency, sentence diversity, etc.).

[0051] 4. The server performs a generated AI check.

[0052] The server performs a spawned AI check by:

[0053] Copy-paste check: Use web scraping tools and APIs to collect information from the internet and compare it with text. Find matches and calculate the copy-paste rate.

[0054] AI document judgment: Using a machine learning model, we judge the possibility of generative AI based on writing style and characteristics. We score the possibility of generative AI for each sentence and calculate the overall generative AI rate.

[0055] Specific examples

[0056] The server uses a non-destructive scraping tool to collect information from the internet, calculates the match rate with the text, and uses a BERT-based machine learning model to analyze the style of the text and evaluate the likelihood that it is generative AI.

[0057] 5. The server provides the results to the user

[0058] The server generates a detailed report based on the analysis results and provides it to the user. If the generated AI rate exceeds a certain threshold (e.g., 70%), information including instructions for resubmission and a warning will be provided.

[0059] Specific examples

[0060] The server creates a report summarizing the results of the 70% AI generation rate and details (such as high match rate and characteristics of the AI-generated writing style) and displays it on the user's web interface, allowing the user to check the results and resubmit the document if necessary.

[0061] This concludes the description of a specific embodiment of the system according to the present invention, which is a powerful tool for evaluating the legitimacy of documents with high accuracy using generative AI.

[0062] The processing flow will be explained below.

[0063] Step 1:

[0064] A user accesses the upload form on the web page, selects "Report.docx", enters meta information such as the university name and subject (e.g., university name "General University", subject "Information Technology"), and presses the submit button.

[0065] Step 2:

[0066] The server saves the "report.docx" file and meta information received from the user to the file system, and then converts the saved file to a text format using an appropriate tool (e.g., Apache POI).

[0067] Step 3:

[0068] The server stores the converted text in memory and begins analyzing it using an NLP (Natural Language Processing) library (e.g., spaCy, NLTK). For each sentence, grammatical elements such as subject, predicate, and object are extracted, and features for evaluating style and tone are calculated.

[0069] Step 4:

[0070] To perform the copy-paste check, the server collects information on the Internet using a web scraping tool or API (e.g., Google (registered trademark) Custom Search API). The collected information is compared with the analyzed text to find matching parts, and the matching rate is calculated as the copy-paste rate.

[0071] Step 5:

[0072] The server uses a machine learning model (e.g., a BERT-based model) to perform AI document judgment based on the analyzed text. It scores the likelihood of generative AI based on features such as consistency of writing style and excessive use of specific keywords, and makes a judgment for each sentence.

[0073] Step 6:

[0074] The server combines the copy-paste rate and the AI ​​document judgment score to calculate the overall AI generation rate, and determines whether the AI ​​generation rate exceeds a certain threshold (e.g., 70%).

[0075] Step 7:

[0076] The server generates the AI ​​generation rate results and a detailed report (e.g., high match rate, stylistic features of the AI-generated text) and sends them to the user's web interface. The user can review the results and receive instructions to resubmit the document if necessary.

[0077] This is the specific processing flow of the generative AI checker system, which makes it possible to evaluate the legitimacy and originality of documents with high accuracy.

[0078] Example 1

[0079] 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."

[0080] In recent years, advances in generative AI technology have led to an increase in the number of documents, such as resumes and reports, automatically generated by AI. These documents may lack originality and reliability, posing a major problem for educational institutions and businesses. However, existing methods lack effective methods for identifying and detecting such AI-generated documents, which poses the risk of incorrect evaluations and judgments.

[0081] 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.

[0082] In this invention, the server includes a means for users to upload their own documents, a means for converting the documents into text, a means for analyzing the text using natural language processing technology, a means for collecting information on the Internet and comparing it with the text to calculate a match rate, a means for determining the characteristics of the generative AI using a machine learning model, a means for calculating the generative AI rate based on the match rate and the determination result, and a means for providing the result of the generative AI rate to the user, thereby enabling highly accurate evaluation and determination of whether a document was created by a generative AI.

[0083] "User" means a person or entity that accesses the System and uploads their own documents for analysis.

[0084] "Document" refers to a file containing text data, such as a resume or report, that a user uploads to the system.

[0085] "Means of converting to text" refers to software or tools used to convert document formats such as PDF or DOCX into text format.

[0086] "Natural language processing technology" refers to technology for analyzing text data and extracting its grammatical structure and characteristics.

[0087] "Information on the Internet" refers to publicly available information accessible via the Internet, such as web pages and online databases.

[0088] "Match rate" refers to an indicator of the degree to which the text matches the content of the uploaded document and information on the Internet.

[0089] A "machine learning model" refers to an algorithm or mathematical model that learns from data and automatically performs specific tasks.

[0090] "Generative AI characteristics" refers to the stylistic and structural features that are unique to documents created by generative AI.

[0091] "Generative AI Rate" refers to a score or percentage that indicates the likelihood that a document was created by a generative AI.

[0092] "Warning" refers to a message or information that notifies users when a document is likely to have been created by generative AI.

[0093] "Resubmission" refers to a request by a user to submit a different document if the document is likely to have been created by generative AI.

[0094] This invention relates to a system that determines whether a document, such as a resume or a school report, was created by generative AI. The system performs this determination through a series of steps in which a user uploads the document and analyzes it.

[0095] System configuration

[0096] The system includes the following major components:

[0097] 1. User's device

[0098] 2. Server

[0099] 3. Natural Language Processing Libraries

[0100] 4. Machine Learning Models

[0101] System Details

[0102] Document upload

[0103] Users upload documents using a web interface that includes a file selection button and fields for entering meta information such as university name and subject name.

[0104] As a concrete example, a user accesses a web page, selects "Report.docx" in the upload form, enters the university name "General University" and the subject "Information Technology," and then presses the upload button, which sends the document and this information to the server.

[0105] Text conversion

[0106] The server converts the received file to text format using a PDF to text converter or a DOCX parser (e.g., Apache POI).

[0107] For example, the server receives a "report.docx" and uses the Apache POI library to parse the document into text format, which is then stored in memory.

[0108] Text Analysis

[0109] The server analyzes the text using natural language processing techniques, using natural language processing libraries (e.g., spaCy, NLTK).

[0110] For example, the server analyzes text data and extracts grammatical elements such as subject, predicate, and object of each sentence. It also evaluates writing style and tone and calculates features such as word frequency and sentence diversity.

[0111] Generation AI Check

[0112] The server performs a generated AI check using the following means:

[0113] 1. Copy-paste check: Use web scraping tools and APIs to collect information from the internet and compare it with text.

[0114] 2. AI document assessment: Using a machine learning model (e.g., a BERT-based model), we assess the potential for generative AI based on writing style and features.

[0115] For example, the server uses a non-destructive scraping tool to collect information from the internet, calculates the match rate with the text, and uses a BERT-based machine learning model to analyze the style of the text and score the potential of the generative AI.

[0116] Providing results

[0117] The server generates a detailed report based on the analysis results and provides it to the user. If the generated AI rate exceeds a certain threshold (e.g., 70%), the server provides the user with information including instructions for resubmission and a warning.

[0118] For example, the server generates a report summarizing the results of a 70% AI generation rate and details (such as high match rates and characteristics of the AI-generated writing style) and displays it on the user's web interface, allowing the user to check the results and resubmit the document if necessary.

[0119] This system is a powerful tool for evaluating the legitimacy of documents with high accuracy using generative AI, and will contribute to ensuring the reliability of documents in educational institutions and companies.

[0120] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0121] Step 1:

[0122] The user uploads their own document. The user accesses the web interface, selects the document file (e.g., report.docx) using the file selection button, and enters meta information such as the university name and subject in the input fields. Specifically, the user uploads "report.docx," enters "general university" and "information technology," and presses the upload button. This sends the file and meta information as input to the server.

[0123] Step 2:

[0124] The server converts the received file into text format. The server uses a PDF to text conversion tool or a DOCX parser (e.g., Apache POI) to convert the received document file from PDF or DOCX format to text format. Specifically, the server converts "report.docx" to text using the conversion tool and saves it in memory. The input of this step is the file data, and the output is the converted text data.

[0125] Step 3:

[0126] The server analyzes the converted text. It uses natural language processing techniques (e.g., spaCy, NLTK) to analyze grammatical structure and stylistic features. Specifically, the server passes the text data to an NLP library to extract grammatical elements such as the subject, predicate, and object of each sentence. It also calculates word frequency and sentence diversity. The input for this step is the text data, and the output is the analysis results (grammatical element and feature data).

[0127] Step 4:

[0128] The server performs a generation AI check. The server determines whether the document was created by a generation AI using two methods:

[0129] Copy-paste check: Collects information from the Internet using web scraping tools and APIs, compares it with the text, and calculates the match rate. Specifically, the server compares the text with multiple sources collected by the scraping tool and calculates the match rate.

[0130] AI document assessment: A BERT-based machine learning model is used to determine the likelihood of generative AI based on writing style and features. Specifically, the server uses the BERT model to score the likelihood of generative AI using the analysis results as input. The input for this step is the analysis results and information data on the Internet, and the output is the copy-paste rate and the score of the generative AI.

[0131] Step 5:

[0132] The server calculates the generation AI rate based on the analysis results and provides it to the user. If the generation AI rate exceeds a certain threshold (e.g., 70%), a detailed report is generated that includes information such as instructions for resubmission and warnings. Specifically, the server creates a report and displays the results on the user's web interface. The inputs to this step are the copy-paste rate and the generation AI score, and the output is a detailed report.

[0133] Through the above processing steps, this system can accurately evaluate the possibility that a document can be created by AI, and provide users with reliable information.

[0134] (Application example 1)

[0135] 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."

[0136] In recent years, the use of generative AI to generate documents has become commonplace, leading to an increase in fraudulent use and acts of fraud. In particular, in electronic payments, the risk of fraudulent transactions is increasing when important documents such as contracts and invoices are forged by generative AI. Effective measures to prevent such fraudulent use and acts are needed, but existing technologies cannot adequately address this issue.

[0137] 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.

[0138] In this invention, the server includes means for users to upload their own documents, means for converting the documents into text, means for analyzing the text using natural language processing technology, means for calculating the copy-paste rate by comparing it with information on the Internet, means for determining the characteristics of the generating AI, means for calculating the generating AI rate based on the copy-paste rate and the determination result, means for providing the generating AI rate result to the user, and means for determining whether documents related to payment have been created by a generating AI in order to prevent fraudulent use and misuse during electronic payments. This makes it possible to prevent fraudulent use and misuse during electronic payments.

[0139] "User" refers to any individual or entity that uses the Service or System.

[0140] "Document" refers to any paper or file containing information, including formats such as PDF and DOCX.

[0141] "Means of converting to text" refers to tools and technologies used to analyze the contents of document files and convert them to plain text. Specifically, text extraction libraries are used.

[0142] "Natural language processing technology" refers to a set of techniques and methods that enable computers to understand and analyze human language, including NLP libraries and machine learning models.

[0143] The "copy-paste rate" is a measure of how closely the text in a document matches other sources on the Internet.

[0144] "Generative AI characteristics" refers to specific patterns, style, structure, and other characteristics found in documents created by generative AI.

[0145] The "AI generation rate" is a numerical indicator that indicates the likelihood that a document was created by a generative AI.

[0146] "Unauthorized use" refers to the intentional and unfair use of a system or service.

[0147] "Fraud" is the act of deceiving others through dishonest means in order to gain profits.

[0148] "Payment" refers to the process of making monetary transactions for goods and services.

[0149] "Means of determination" refers to the technology or algorithms used to assess whether a document was generated by AI.

[0150] MODE FOR CARRYING OUT THE INVENTION

[0151] This invention provides a system that allows users to upload documents related to electronic payments and performs a series of processes to determine whether the documents were created by a generation AI. The specific configuration and processing steps of the system are described below.

[0152] System configuration and operation

[0153] The system consists of the following components:

[0154] 1. User device: Provides an interface that runs on a smartphone, tablet, or PC and allows users to upload documents.

[0155] 2. Server: The central processing unit that receives uploaded documents, converts them to text, and performs analysis.

[0156] 3. Text conversion tools: Software that converts documents into text format (e.g., textract).

[0157] 4. Natural Language Processing (NLP) models: Libraries for analyzing text and extracting grammatical structures and features (e.g., spaCy).

[0158] 5. Generative AI Verification Model: A machine learning model to determine whether a document was created by generative AI (e.g., a BERT-based model).

[0159] 6. Copy-paste check tools: Web scraping tools and APIs that compare information on the internet to calculate copy-paste rates.

[0160] Processing flow

[0161] 1. The user uploads relevant documents (e.g., invoices, contracts) through a web interface or application provided on the user's device.

[0162] 2. The server receives the document sent by the user and converts the document into text format using a text conversion tool (e.g., textract).

[0163] 3. The server analyzes the converted text using a natural language processing (NLP) model (e.g., spaCy) to extract grammatical structures and features.

[0164] 4. The server uses a copy-paste check tool to compare the text with information on the Internet and calculate the copy-paste rate.

[0165] 5. The server uses a generative AI decision model (e.g., a BERT-based model) to determine whether the text was created by a generative AI and scores the likelihood.

[0166] 6. The server calculates the generative AI rate based on the copy-paste rate and the results of the generative AI judgment model.

[0167] 7. The server provides the user with the results of the generated AI rating and, if necessary, provides additional authentication measures and instructions for resubmission.

[0168] Hardware and software used

[0169] Server: A computer system that performs major processing, often deployed on a cloud service.

[0170] User devices: smartphones, tablets, computers, etc.

[0171] software:

[0172] Textract: A tool that converts documents into text.

[0173] SpaCy: A library for natural language processing.

[0174] BERT: A machine learning model that determines whether a document was created by generative AI.

[0175] Web scraping tools or APIs: Tools for copy-paste checking.

[0176] Specific examples

[0177] Specific examples of prompts include the following scenarios:

[0178] "Let's say a user uploads a file named 'invoice.pdf'. This file is uploaded from a smartphone and analyzed by an artificial intelligence (AI)-based fraud detection system. The server receives the file, converts and analyzes the text content, and then makes a judgment on the AI-generated document. The analysis result is sent to the payment provider, which then takes the necessary action."

[0179] As described above, by using this system, it is possible to prevent unauthorized use and fraudulent acts related to electronic payments.

[0180] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0181] Step 1:

[0182] A user uploads a document

[0183] Users upload relevant documents (e.g., invoices or contracts) from an application or web interface on their smartphone, tablet, or PC. The input is a document file such as PDF or DOCX, and the output is the uploaded document file. Specific operations involve selecting a document using the file selection button and clicking the upload button.

[0184] Step 2:

[0185] The server receives the document

[0186] The server receives the document sent by the user. The input is the document file sent by the user, and the output is a document file stored on the server side. Specifically, the server receives an HTTP POST request and retrieves the document file contained therein.

[0187] Step 3:

[0188] The server converts the document into text

[0189] The server uses a text conversion tool (e.g., textract) to convert the document file into a text format. The input is a document file in PDF or DOCX format, and the output is plain text data. Specifically, the text extraction library parses the document file and generates plain text.

[0190] Step 4:

[0191] The server parses the text

[0192] The server uses a natural language processing (NLP) model (e.g., spaCy) to analyze the converted text. The input is plain text data, and the output is an analysis result including grammatical elements and features. Specifically, the NLP model tokenizes the text and extracts grammatical elements such as subject, predicate, and object.

[0193] Step 5:

[0194] The server compares the information on the Internet and calculates the copy-paste rate

[0195] The server uses copy-paste checking tools and APIs to match the text with online information and calculate the copy-paste rate. The input is the analyzed text and multiple online sources, and the output is a number indicating the copy-paste rate. Specifically, the server sends the text to the web scraping tool, finds matches, and calculates the match rate.

[0196] Step 6:

[0197] The server uses the generative AI judgment model to determine the characteristics of the generative AI.

[0198] The server uses a generative AI judgment model (e.g., a BERT-based model) to determine whether a document was created by a generative AI. The input is the parsed text, and the output is a score indicating the likelihood of the generative AI. Specifically, the server inputs the text into the judgment model and obtains the scored result.

[0199] Step 7:

[0200] The server calculates the AI ​​generation rate

[0201] The server calculates the AI ​​generation rate based on the copy-paste rate and the results of the AI ​​generation judgment model. The input is the copy-paste rate and a score indicating the AI ​​generation potential, and the output is a number indicating the AI ​​generation rate. Specifically, the server applies an algorithm that combines the copy-paste rate and the score to calculate the AI ​​generation rate.

[0202] Step 8:

[0203] The server provides the results to the user

[0204] The server provides users with the results of the AI ​​generation rate and detailed analysis results. The input is a numerical value indicating the AI ​​generation rate and detailed analysis data, and the output is a report format that can be viewed by users. Specifically, the server formats the analysis results into a report in HTML or PDF format and displays it on the user's terminal interface.

[0205] The above is a specific description of the processing steps in the system of the invention.

[0206] 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.

[0207] This invention relates to a system that determines whether documents such as resumes and school reports were created by generative AI (generative artificial intelligence), and by combining it with an emotion engine that recognizes the user's emotions, it provides more advanced judgment and feedback. This system improves the user experience by adding analysis using the user's emotion data in addition to the series of processes from uploading the document to analyzing it and providing the results.

[0208] System Overview

[0209] 1. User uploads a document

[0210] Users upload their resumes and reports on the system's web interface (client terminal). Upload methods include a file selection button and meta-information input fields (university name, subjects, etc.).

[0211] Specific examples

[0212] A user accesses the web page, selects "Report.docx" in the upload form, enters the university name "General University" and the subject "Information Technology," and presses the upload button to send the document and information to the server.

[0213] 2. The server receives the file and converts it to text

[0214] The server receives the file sent by the user and converts the document from PDF or DOCX format to text format, for example using a PDF to text converter or a DOCX parser.

[0215] Specific examples

[0216] The server receives the "report.docx" and converts it to text using an appropriate library (e.g. Apache POI), which stores the converted text in memory.

[0217] 3. The server parses the text

[0218] The server applies natural language processing techniques to the converted text to analyze sentence structure and features, which may involve the use of NLP libraries (e.g., spaCy, NLTK).

[0219] Specific examples

[0220] The server analyzes the converted text and extracts grammatical elements such as the subject, predicate, and object of each sentence. It also evaluates the writing style and tone and calculates features (word frequency, sentence diversity, etc.).

[0221] 4. The server performs a generated AI check.

[0222] The server performs a spawned AI check by:

[0223] Copy-paste check: Use web scraping tools and APIs to collect information from the internet and compare it with text. Find matches and calculate the copy-paste rate.

[0224] AI document judgment: Using a machine learning model, we judge the possibility of generative AI based on writing style and characteristics. We score the possibility of generative AI for each sentence and calculate the overall generative AI rate.

[0225] Specific examples

[0226] The server uses a non-destructive scraping tool to collect information from the internet, calculates the match rate with the text, and uses a BERT-based machine learning model to analyze the style of the text and evaluate the likelihood that it is generative AI.

[0227] 5. The server performs analysis using the emotion engine

[0228] When a user uploads a document, the emotion engine recognizes the user's emotions from their facial expressions and input. This emotion data is then incorporated into the analysis and used to calculate the AI ​​generation rate and provide feedback.

[0229] Specific examples

[0230] When a user uploads a document, their facial expressions are analyzed via a webcam to obtain emotional data. If the user shows signs of anxiety, the server uses this information to adjust the threshold for AI generation rates and provide accurate feedback.

[0231] 6. The server provides the results to the user

[0232] The server generates a detailed report based on the analysis results and provides it to the user. If the AI ​​generation rate exceeds a certain threshold (e.g., 70%), information including instructions for resubmission and a warning is provided. The server can also adjust the way results are presented based on emotion data.

[0233] Specific examples

[0234] The server creates a report summarizing the results of 70% AI generation rates and their details (such as high match rates and characteristics of the AI-generated writing style) and displays it on the user's web interface. If the user shows any anxiety, it provides a gentle explanation and offers additional support options.

[0235] This concludes the description of a specific embodiment of the system according to the present invention. This system uses generative AI to evaluate the validity of documents with high accuracy, and also takes into account the user's emotions to provide more personalized feedback.

[0236] The processing flow will be explained below.

[0237] Step 1:

[0238] A user accesses the upload form on the web page, selects "Report.docx", enters meta information such as the university name and subject (e.g., university name "General University", subject "Information Technology"), and presses the submit button.

[0239] Step 2:

[0240] The server saves the "report.docx" file and meta information received from the user to the file system, and then converts the saved file to a text format using an appropriate tool (e.g., Apache POI).

[0241] Step 3:

[0242] The server stores the converted text in memory and begins analyzing it using an NLP (Natural Language Processing) library (e.g., spaCy, NLTK). For each sentence, grammatical elements such as subject, predicate, and object are extracted, and features for evaluating style and tone are calculated.

[0243] Step 4:

[0244] When a user uploads a document, the emotion engine recognizes the user's emotion from the webcam, keyboard typing speed, etc. This emotion data is analyzed using specific algorithms (e.g., facial expression recognition algorithms and typing speed analysis).

[0245] Step 5:

[0246] To perform the copy-paste check, the server collects information from the Internet using web scraping tools or APIs (e.g., Google Custom Search API). The collected information is compared with the analyzed text to find matching parts, and the matching rate is calculated as the copy-paste rate.

[0247] Step 6:

[0248] The server uses a machine learning model (e.g., a BERT-based model) to perform AI document judgment based on the analyzed text. It scores the likelihood of generative AI based on features such as consistency of writing style and excessive use of specific keywords, and makes a judgment for each sentence.

[0249] Step 7:

[0250] The server combines the copy-paste rate and the AI ​​document judgment score to calculate the overall generative AI rate, where it adjusts the threshold based on the recognized user emotion data to enable personalized feedback.

[0251] Step 8:

[0252] The server generates the AI ​​generation rate results and a detailed report (e.g., high match rate, stylistic features of the AI-generated writing) and sends it to the user's web interface. If the user shows signs of nervousness or anxiety, the report includes additional support options. The user is then instructed to review the results and resubmit the document if necessary.

[0253] This is the specific processing flow of the generative AI checker system that combines an emotion engine. This system can evaluate the validity and originality of documents with high accuracy, and provide personalized feedback that takes the user's emotions into account.

[0254] Example 2

[0255] 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."

[0256] Conventional document assessment systems have difficulty accurately determining whether a document was created by a generative AI. Furthermore, they lack a mechanism for taking the user's feelings into consideration when providing the user with a result of whether a document was AI-generated. Therefore, there is a need to improve the reliability of assessment results and the user experience.

[0257] 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.

[0258] In this invention, the server includes means for users to upload their own documents, means for converting the documents into text, means for analyzing the text using natural language processing technology, means for comparing the text with information on the Internet to calculate the copy-paste rate, means for determining the possibility of a generative AI from writing style and characteristics using a machine learning model, means for acquiring user emotion data and incorporating it into the analysis, means for calculating a generative AI rate based on the copy-paste rate, the generative AI determination result, and the emotion data, and means for providing the result of the generative AI rate to the user. This increases the accuracy of determining documents created by the generative AI and makes it possible to provide feedback that takes into account the user's emotions.

[0259] "Document" refers to a formal document such as a resume or report uploaded by a User.

[0260] "Text" refers to character string data converted from document formats such as PDF and DOCX.

[0261] "Natural language processing technology" refers to the technology of analyzing text and evaluating grammatical elements and writing style, and uses NLP libraries, etc.

[0262] The "copy-paste rate" refers to the rate at which the text matches when compared with information on the Internet.

[0263] "Generative AI" refers to text or documents created using a generative AI model.

[0264] A "machine learning model" refers to a technology that uses algorithms to learn patterns from large amounts of data and determine the characteristics of the generated AI.

[0265] "Emotion data" refers to data that indicates the emotional state of the user that can be obtained from their facial expressions and input content.

[0266] "Generative AI Rate" refers to the percentage that indicates the likelihood that the text was created by a generative AI.

[0267] "Feedback" refers to information that provides analysis results and advice to users.

[0268] This invention is a system that determines whether documents such as resumes and school reports were created by generative AI, and by combining it with an emotion engine that recognizes the user's emotions, it provides more advanced judgment and feedback. This system improves the user experience by adding analysis using the user's emotion data in addition to the series of processes from uploading the document to analyzing it and providing the results.

[0269] First, the user uploads their document through the device's web interface. The uploading method includes a file selection button and meta-information input fields (university name, subject, etc.). Specifically, the user accesses the web page, selects "Report.docx" in the upload form, and enters the university name "General University" and the subject "Information Technology." Then, by pressing the upload button, the document and information are sent to the server.

[0270] The server receives the file sent by the user and converts it from file format (PDF or DOCX) to text format using the appropriate library. For example, the server receives a "report.docx" and converts it to text using the Apache POI library. This converted text is stored in memory.

[0271] The server then applies natural language processing techniques to the converted text to analyze the structure and characteristics of the sentences. NLP libraries such as spaCy and NLTK are used. For example, the server analyzes the converted text and extracts grammatical elements such as the subject, predicate, and object of each sentence. It also evaluates the writing style and tone and calculates features (word frequency, sentence diversity, etc.).

[0272] In addition, the server performs copy-paste checks and AI document assessments. Copy-paste checks involve collecting information on the Internet using web scraping tools and APIs and calculating the match rate with the text. As a specific example, the server collects information on the Internet using a non-destructive scraping tool and calculates the match rate with the text. AI document assessment uses a BERT-based machine learning model to analyze the writing style and evaluate the possibility that the text is generative.

[0273] The system also performs analysis using an emotion engine. When a user uploads a document, their facial expressions are analyzed via a webcam to capture emotional data. This emotional data is incorporated into the analysis and reflected in the calculation of the AI ​​generation rate and the feedback content. For example, if the user expresses anxiety, the server can use this information to adjust the AI ​​generation rate threshold and provide more accurate feedback.

[0274] Finally, the server generates a detailed report based on the analysis results and provides it to the user. If the AI ​​generation rate exceeds a certain threshold (e.g., 70%), information including instructions for resubmission and a warning is provided. The server can also adjust the way results are presented based on emotional data. For example, the server could create a report summarizing the results of a 70% AI generation rate and its details (high match rate, characteristics of the AI-generated writing style, etc.) and display it on the user's web interface. Users who express anxiety could be offered gentle explanations and additional support options.

[0275] The system uses generative AI to accurately assess the legitimacy of documents and also takes user sentiment into account to provide more personalized feedback.

[0276] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0277] Step 1:

[0278] The user inputs the document file and meta information and presses the upload button on the device's web interface. This is the process of inputting a file called "report.docx" and meta information such as the university name "general university" and the subject "information technology," and then clicking the upload button.

[0279] Input: Document file (e.g., report.docx), meta information (university name, subject)

[0280] Output: The file and meta information are sent to the server.

[0281] Step 2:

[0282] The server receives the file sent by the user and converts it from PDF or DOCX format to a text format using an appropriate library (e.g., Apache POI). This converted text data is temporarily stored in memory.

[0283] Input: Document file

[0284] Data processing: Converting from file format to text format (using Apache POI)

[0285] Output: Text format data (temporarily saved in memory)

[0286] Step 3:

[0287] The server analyzes the converted text, using natural language processing techniques (e.g., spaCy, NLTK) to extract grammatical elements (e.g., subject, predicate, object) and calculate features of style and tone (e.g., word frequency, sentence diversity).

[0288] Input: Text format data

[0289] Data processing: Extracting grammatical elements and analyzing style and tone (using spaCy and NLTK)

[0290] Output: Analysis results (grammatical elements, style and tone features)

[0291] Step 4:

[0292] The server performs copy-paste checks, collects information from the internet using web scraping tools and APIs, compares it with text, and calculates the match rate. In addition, it analyzes writing style using a BERT-based machine learning model to evaluate the potential of generative AI.

[0293] Input: Text format data

[0294] Data Computing: Web scraping and text matching, matching rate calculation, stylistic analysis with the BERT model

[0295] Output: Match rate, possibility evaluation score of the generated AI

[0296] Step 5:

[0297] When a user uploads a document, their facial expression is captured using the device's webcam. The server uses an emotion engine to analyze the facial expression data and obtain emotional data, which is then used for analysis.

[0298] Input: Facial expression capture data

[0299] Data calculation: Analysis of facial expression data (using emotion engine)

[0300] Output: Emotion data

[0301] Step 6:

[0302] The server calculates the AI ​​generation rate by integrating and evaluating the copy-paste rate, stylistic analysis results, and emotional data. A specific algorithm is used to calculate the AI ​​generation rate.

[0303] Input: Match rate, generated AI evaluation score, emotion data

[0304] Data calculation: Algorithm-based data integration and generation AI rate calculation

[0305] Output: Generation AI rate

[0306] Step 7:

[0307] The server generates a detailed report based on the analysis results and provides it to the user. If the AI ​​generation rate exceeds a certain threshold (e.g., 70%), the report will include information such as instructions for resubmission and warnings. The server will also adjust the presentation of results based on emotion data.

[0308] Input: AI generation rate, emotion data

[0309] Output: Detailed report, feedback (resubmission instructions, warnings, etc.)

[0310] A feature of this system is that it takes user emotions into consideration while performing highly accurate document judgment using generative AI and providing personalized feedback to users.

[0311] (Application example 2)

[0312] 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."

[0313] In modern logistics centers, verifying the authenticity of package documents is important, but there is a risk of documents created by generative AI being mixed in. Furthermore, the emotions and stress of inspectors can affect work efficiency and accuracy, but there is no system that effectively monitors and provides feedback. Therefore, a system that combines package document authenticity verification and worker emotion recognition is needed.

[0314] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 2 is realized by the following means.

[0315] In this invention, the server includes means for users to upload their own documents, means for converting the documents into text, means for analyzing the text using natural language processing technology, means for calculating the copy-paste rate by comparing it with information on the Internet, means for determining the characteristics of the generating AI, means for calculating the generating AI rate based on the copy-paste rate and the determination result, means for providing the generating AI rate result to the user, means for recognizing the user's emotions, and means for adjusting the method for presenting the generating AI rate result based on the emotion data. This makes it possible to evaluate the legitimacy of package documents with high accuracy and provide feedback that takes into account the emotions of inspection workers.

[0316] "User" means a person or institution that uses the System to upload documents and obtain analysis results thereof.

[0317] "Document" means any electronic or paper information that can be converted into text format, including, for example, a resume, report, or shipping instructions.

[0318] The "means for converting to text" is a mechanism for converting an image document into text data using optical character recognition (OCR) technology.

[0319] "Natural language processing technology" is a technology for analyzing text data and extracting and analyzing grammatical structure and meaning, and specifically includes morphological analysis, grammatical analysis, and semantic analysis.

[0320] The "copy-paste rate" is the percentage of a document that has been copied in whole or in part from another existing document, and is calculated by comparing it with sources on the Internet.

[0321] "Means for determining characteristics of generative artificial intelligence" refers to techniques that use machine learning models to analyze the style and structure of documents in order to identify whether they were created by generative AI.

[0322] The "generated AI rate" is an indicator of the likelihood that a document was created by a generative AI, and is calculated based on the copy-paste rate and the AI ​​judgment results.

[0323] "Means for recognizing the user's emotions" refers to technology for extracting and analyzing emotions from the user's facial expressions and voice data, and typically involves using a camera or microphone.

[0324] The "means for adjusting the method of presenting results" is a mechanism that changes the way the analysis results are displayed based on the user's emotional data and provides optimal feedback.

[0325] A head-mounted display is a device worn on the head that displays information directly into the field of view. It also has built-in cameras and sensors that can acquire data on the external environment in real time.

[0326] The system for implementing this invention aims to control quality in logistics centers. Specifically, it determines whether shipping instructions and content descriptions included in packages were created using generative AI, and further improves work efficiency by providing feedback using the emotion data of inspection workers.

[0327] 1. System Configuration

[0328] The system includes the following main components:

[0329] 1.1 Head-mounted display (HMD)

[0330] An inspection worker wearing an HMD (e.g., a typical smart HMD device) scans the shipping instructions and contents description inside each package.

[0331] 1.2 Camera and microphone

[0332] The HMD's built-in camera scans documents and converts the images into text using OCR technology (e.g., Tesseract OCR).The camera and microphone also capture the inspector's facial expressions and voice data.

[0333] 1.3 Server

[0334] The server converts the received image data into text, analyzes grammatical structure and writing style using natural language processing (NLP) technology (e.g., spaCy), and uses a generative AI model (BERT-based model) to determine whether the document was created by generative AI and calculates the generative AI rate.

[0335] 1.4 Emotion Engine

[0336] An emotion engine running on the server analyzes the emotional data of inspection workers and adjusts the feedback content.

[0337] 2. Explanation of the processing flow

[0338] The system process flow is as follows:

[0339] 2.1 Image to Text Conversion

[0340] An inspector wearing an HMD scans the package documents with a camera and sends the image data to a server, which then uses OCR technology to convert the image data into text.

[0341] 2.2 Analysis using natural language processing

[0342] The server applies NLP techniques to the converted text to extract grammatical structures and features, including analyzing the subject, predicate, and object of sentences, and assessing sentence variety and tone.

[0343] 2.3 Generation AI judgment

[0344] The server uses the generative AI model to determine whether a document was created by generative AI, performs copy-paste checks and AI document determination functions, and calculates the generative AI rate.

[0345] 2.4 Sentiment Analysis and Feedback

[0346] The server uses an emotion engine to analyze the inspector's emotional data, and if the inspector shows signs of stress or fatigue, it adjusts the feedback method and provides appropriate alerts and suggestions (e.g., to take a break).

[0347] 3. Specific Examples

[0348] As a concrete example, an inspection worker scans a shipping instruction sheet with the file name "Doc123.jpg" using an HMD and sends it to the server. The server then uses OCR technology to convert the image into text and calculates the AI ​​generation rate. At the same time, emotional data displayed by the worker is also collected and analyzed, and feedback based on the results is displayed on the HMD display.

[0349] "Please determine if this report was created by a generative AI." is an example of a prompt sentence, which is used as input to a generative AI model.

[0350] Names of specific hardware and software used

[0351] HMD: General smart HMD devices (e.g., Microsoft® HoloLens®, etc.)

[0352] Camera and microphone: Camera and microphone built into the HMD

[0353] OCR technology: Tesseract OCR

[0354] NLP technology: spaCy

[0355] Generative AI model: BERT-based generative AI decision model

[0356] Emotion Engine: Emotion Recognition Library (OpenCV and Dlib)

[0357] In this way, the present invention provides an efficient and highly accurate system for evaluating package documents through quality control and worker emotion recognition in logistics centers.

[0358] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[0359] Step 1:

[0360] The user wears a head-mounted display (HMD) and scans shipping instructions and content descriptions.

[0361] Input: Image of shipping instructions or contents description

[0362] Output: Image data

[0363] Specific operation: The camera built into the HMD is used to take a photo of the document inside the package and obtain image data.

[0364] Step 2:

[0365] The server receives the image data and converts the image into text using optical character recognition (OCR) technology.

[0366] Input: Image data

[0367] Output: Text data

[0368] What happens: The server uses an OCR library (e.g., Tesseract) to analyze the image data and extract the text within the document.

[0369] Step 3:

[0370] The server analyzes the extracted text using natural language processing (NLP) techniques.

[0371] Input: Text data

[0372] Output: Grammatical structure and features

[0373] What it does: The server applies NLP libraries (e.g., spaCy) to extract grammatical elements (e.g., subject, predicate, object) in the text and evaluate sentence variety and tone.

[0374] Step 4:

[0375] The server checks for copy-paste and determines the AI ​​that generated the text data.

[0376] Input: Text data

[0377] Output: Generated AI rate

[0378] How it works: The server uses a copy-paste check tool to check against information on the internet and calculates the percentage of matches. It also uses a generative AI model (based on BERT) to analyze the style of the text and calculates the generative AI rate.

[0379] Step 5:

[0380] The server acquires the emotion data of the inspector using the camera and microphone built into the HMD.

[0381] Input: Facial expression and voice data

[0382] Output: Emotion data

[0383] How it works: The server uses the Emotion Recognition Library (OpenCV and Dlib) to analyze facial expressions and voice captured by the camera and microphone to identify the worker's emotional state.

[0384] Step 6:

[0385] The server adjusts the way results are presented based on the emotional data and provides optimal feedback to the user.

[0386] Input: Generated AI rate and emotion data

[0387] Output: Feedback results

[0388] Specific operation: The server integrates the generated AI rate and emotion data and generates a report based on the analysis results. Based on the emotion data, the format and tone of the results displayed on the HMD are adjusted to provide optimal information to the user. For example, if a worker is feeling stressed, a gentle message or a suggestion to take a break is displayed.

[0389] Through these steps, the system is able to evaluate the validity of package documents and provide feedback to improve work efficiency.

[0390] 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.

[0391] The data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of the data generation model 58 is ChatGPT (registered trademark) (Internet search engine).<URL: https: / / openai.com / blog / chatgpt> ), Gemini (registered trademark) (Internet search <url: https: gemini.google.com ?hl="ja">) and other generation AIs. The data generation model 58 is obtained by performing deep learning on a neural network. A prompt including an instruction is input to the data generation model 58, and inference data such as voice data indicating voice, text data indicating text, and image data indicating an image is also input. The data generation model 58 performs inference on the input inference data in accordance with the instruction indicated by the prompt, and outputs the inference result in a data format such as voice data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[0392] 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.

[0393] [Second embodiment]

[0394] FIG. 3 shows an example of the configuration of a data processing system 210 according to the second embodiment.

[0395] 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.

[0396] 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).

[0397] 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.

[0398] 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.

[0399] 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).

[0400] 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.

[0401] 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.

[0402] 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.

[0403] 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.

[0404] In the smart glasses 214, 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.

[0405] 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."

[0406] This invention relates to a system that determines whether a document, such as a resume or a school report, was created by generative AI (generative artificial intelligence). The system performs a series of steps to determine whether the document was created by a user who uploads the document and analyzes it.

[0407] System Overview

[0408] 1. User uploads a document

[0409] Users upload their resumes and reports on the system's web interface (client terminal). Upload methods include a file selection button and meta-information input fields (university name, subjects, etc.).

[0410] Specific examples

[0411] A user accesses the web page, selects "Report.docx" in the upload form, enters the university name "General University" and the subject "Information Technology," and presses the upload button to send the document and information to the server.

[0412] 2. The server receives the file and converts it to text

[0413] The server receives the file sent by the user and converts the document from PDF or DOCX format to text format, for example using a PDF to text converter or a DOCX parser.

[0414] Specific examples

[0415] The server receives the "report.docx" and converts it to text using an appropriate library (e.g. Apache POI), which stores the converted text in memory.

[0416] 3. The server parses the text

[0417] The server applies natural language processing techniques to the converted text to analyze sentence structure and features, which may involve the use of NLP libraries (e.g., spaCy, NLTK).

[0418] Specific examples

[0419] The server analyzes the converted text and extracts grammatical elements such as the subject, predicate, and object of each sentence. It also evaluates the writing style and tone and calculates features (word frequency, sentence diversity, etc.).

[0420] 4. The server performs a generated AI check.

[0421] The server performs a spawned AI check by:

[0422] Copy-paste check: Use web scraping tools and APIs to collect information from the internet and compare it with text. Find matches and calculate the copy-paste rate.

[0423] AI document judgment: Using a machine learning model, we judge the possibility of generative AI based on writing style and characteristics. We score the possibility of generative AI for each sentence and calculate the overall generative AI rate.

[0424] Specific examples

[0425] The server uses a non-destructive scraping tool to collect information from the internet, calculates the match rate with the text, and uses a BERT-based machine learning model to analyze the style of the text and evaluate the likelihood that it is generative AI.

[0426] 5. The server provides the results to the user

[0427] The server generates a detailed report based on the analysis results and provides it to the user. If the generated AI rate exceeds a certain threshold (e.g., 70%), information including instructions for resubmission and a warning will be provided.

[0428] Specific examples

[0429] The server creates a report summarizing the results of the 70% AI generation rate and details (such as high match rate and characteristics of the AI-generated writing style) and displays it on the user's web interface, allowing the user to check the results and resubmit the document if necessary.

[0430] This concludes the description of a specific embodiment of the system according to the present invention, which is a powerful tool for evaluating the legitimacy of documents with high accuracy using generative AI.

[0431] The processing flow will be explained below.

[0432] Step 1:

[0433] A user accesses the upload form on the web page, selects "Report.docx", enters meta information such as the university name and subject (e.g., university name "General University", subject "Information Technology"), and presses the submit button.

[0434] Step 2:

[0435] The server saves the "report.docx" file and meta information received from the user to the file system, and then converts the saved file to a text format using an appropriate tool (e.g., Apache POI).

[0436] Step 3:

[0437] The server stores the converted text in memory and begins analyzing it using an NLP (Natural Language Processing) library (e.g., spaCy, NLTK). For each sentence, grammatical elements such as subject, predicate, and object are extracted, and features for evaluating style and tone are calculated.

[0438] Step 4:

[0439] To perform the copy-paste check, the server collects information from the Internet using web scraping tools or APIs (e.g., Google Custom Search API). The collected information is compared with the analyzed text to find matching parts, and the matching rate is calculated as the copy-paste rate.

[0440] Step 5:

[0441] The server uses a machine learning model (e.g., a BERT-based model) to perform AI document judgment based on the analyzed text. It scores the likelihood of generative AI based on features such as consistency of writing style and excessive use of specific keywords, and makes a judgment for each sentence.

[0442] Step 6:

[0443] The server combines the copy-paste rate and the AI ​​document judgment score to calculate the overall AI generation rate, and determines whether the AI ​​generation rate exceeds a certain threshold (e.g., 70%).

[0444] Step 7:

[0445] The server generates the AI ​​generation rate results and a detailed report (e.g., high match rate, stylistic features of the AI-generated text) and sends them to the user's web interface. The user can review the results and receive instructions to resubmit the document if necessary.

[0446] This is the specific processing flow of the generative AI checker system, which makes it possible to evaluate the legitimacy and originality of documents with high accuracy.

[0447] Example 1

[0448] 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."

[0449] In recent years, advances in generative AI technology have led to an increase in the number of documents, such as resumes and reports, automatically generated by AI. These documents may lack originality and reliability, posing a major problem for educational institutions and businesses. However, existing methods lack effective methods for identifying and detecting such AI-generated documents, which poses the risk of incorrect evaluations and judgments.

[0450] 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.

[0451] In this invention, the server includes a means for users to upload their own documents, a means for converting the documents into text, a means for analyzing the text using natural language processing technology, a means for collecting information on the Internet and comparing it with the text to calculate a match rate, a means for determining the characteristics of the generative AI using a machine learning model, a means for calculating the generative AI rate based on the match rate and the determination result, and a means for providing the result of the generative AI rate to the user, thereby enabling highly accurate evaluation and determination of whether a document was created by a generative AI.

[0452] "User" means a person or entity that accesses the System and uploads their own documents for analysis.

[0453] "Document" refers to a file containing text data, such as a resume or report, that a user uploads to the system.

[0454] "Means of converting to text" refers to software or tools used to convert document formats such as PDF or DOCX into text format.

[0455] "Natural language processing technology" refers to technology for analyzing text data and extracting its grammatical structure and characteristics.

[0456] "Information on the Internet" refers to publicly available information accessible via the Internet, such as web pages and online databases.

[0457] "Match rate" refers to an indicator of the degree to which the text matches the content of the uploaded document and information on the Internet.

[0458] A "machine learning model" refers to an algorithm or mathematical model that learns from data and automatically performs specific tasks.

[0459] "Generative AI characteristics" refers to the stylistic and structural features that are unique to documents created by generative AI.

[0460] "Generative AI Rate" refers to a score or percentage that indicates the likelihood that a document was created by a generative AI.

[0461] "Warning" refers to a message or information that notifies users when a document is likely to have been created by generative AI.

[0462] "Resubmission" refers to a request by a user to submit a different document if the document is likely to have been created by generative AI.

[0463] This invention relates to a system that determines whether a document, such as a resume or a school report, was created by generative AI. The system performs this determination through a series of steps in which a user uploads the document and analyzes it.

[0464] System configuration

[0465] The system includes the following major components:

[0466] 1. User's device

[0467] 2. Server

[0468] 3. Natural Language Processing Libraries

[0469] 4. Machine Learning Models

[0470] System Details

[0471] Document upload

[0472] Users upload documents using a web interface that includes a file selection button and fields for entering meta information such as university name and subject name.

[0473] As a concrete example, a user accesses a web page, selects "Report.docx" in the upload form, enters the university name "General University" and the subject "Information Technology," and then presses the upload button, which sends the document and this information to the server.

[0474] Text conversion

[0475] The server converts the received file to text format using a PDF to text converter or a DOCX parser (e.g., Apache POI).

[0476] For example, the server receives a "report.docx" and uses the Apache POI library to parse the document into text format, which is then stored in memory.

[0477] Text Analysis

[0478] The server analyzes the text using natural language processing techniques, using natural language processing libraries (e.g., spaCy, NLTK).

[0479] For example, the server analyzes text data and extracts grammatical elements such as subject, predicate, and object of each sentence. It also evaluates writing style and tone and calculates features such as word frequency and sentence diversity.

[0480] Generation AI Check

[0481] The server performs a generated AI check using the following means:

[0482] 1. Copy-paste check: Use web scraping tools and APIs to collect information from the internet and compare it with text.

[0483] 2. AI document assessment: Using a machine learning model (e.g., a BERT-based model), we assess the potential for generative AI based on writing style and features.

[0484] For example, the server uses a non-destructive scraping tool to collect information from the internet, calculates the match rate with the text, and uses a BERT-based machine learning model to analyze the style of the text and score the potential of the generative AI.

[0485] Providing results

[0486] The server generates a detailed report based on the analysis results and provides it to the user. If the generated AI rate exceeds a certain threshold (e.g., 70%), the server provides the user with information including instructions for resubmission and a warning.

[0487] For example, the server generates a report summarizing the results of a 70% AI generation rate and details (such as high match rates and characteristics of the AI-generated writing style) and displays it on the user's web interface, allowing the user to check the results and resubmit the document if necessary.

[0488] This system is a powerful tool for evaluating the legitimacy of documents with high accuracy using generative AI, and will contribute to ensuring the reliability of documents in educational institutions and companies.

[0489] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0490] Step 1:

[0491] The user uploads their own document. The user accesses the web interface, selects the document file (e.g., report.docx) using the file selection button, and enters meta information such as the university name and subject in the input fields. Specifically, the user uploads "report.docx," enters "general university" and "information technology," and presses the upload button. This sends the file and meta information as input to the server.

[0492] Step 2:

[0493] The server converts the received file into text format. The server uses a PDF to text conversion tool or a DOCX parser (e.g., Apache POI) to convert the received document file from PDF or DOCX format to text format. Specifically, the server converts "report.docx" to text using the conversion tool and saves it in memory. The input of this step is the file data, and the output is the converted text data.

[0494] Step 3:

[0495] The server analyzes the converted text. It uses natural language processing techniques (e.g., spaCy, NLTK) to analyze grammatical structure and stylistic features. Specifically, the server passes the text data to an NLP library to extract grammatical elements such as the subject, predicate, and object of each sentence. It also calculates word frequency and sentence diversity. The input for this step is the text data, and the output is the analysis results (grammatical element and feature data).

[0496] Step 4:

[0497] The server performs a generation AI check. The server determines whether the document was created by a generation AI using two methods:

[0498] Copy-paste check: Collects information from the Internet using web scraping tools and APIs, compares it with the text, and calculates the match rate. Specifically, the server compares the text with multiple sources collected by the scraping tool and calculates the match rate.

[0499] AI document assessment: A BERT-based machine learning model is used to determine the likelihood of generative AI based on writing style and features. Specifically, the server uses the BERT model to score the likelihood of generative AI using the analysis results as input. The input for this step is the analysis results and information data on the Internet, and the output is the copy-paste rate and the score of the generative AI.

[0500] Step 5:

[0501] The server calculates the generation AI rate based on the analysis results and provides it to the user. If the generation AI rate exceeds a certain threshold (e.g., 70%), a detailed report is generated that includes information such as instructions for resubmission and warnings. Specifically, the server creates a report and displays the results on the user's web interface. The inputs to this step are the copy-paste rate and the generation AI score, and the output is a detailed report.

[0502] Through the above processing steps, this system can accurately evaluate the possibility that a document can be created by AI, and provide users with reliable information.

[0503] (Application example 1)

[0504] 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."

[0505] In recent years, the use of generative AI to generate documents has become commonplace, leading to an increase in fraudulent use and acts of fraud. In particular, in electronic payments, the risk of fraudulent transactions is increasing when important documents such as contracts and invoices are forged by generative AI. Effective measures to prevent such fraudulent use and acts are needed, but existing technologies cannot adequately address this issue.

[0506] 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.

[0507] In this invention, the server includes means for users to upload their own documents, means for converting the documents into text, means for analyzing the text using natural language processing technology, means for calculating the copy-paste rate by comparing it with information on the Internet, means for determining the characteristics of the generating AI, means for calculating the generating AI rate based on the copy-paste rate and the determination result, means for providing the generating AI rate result to the user, and means for determining whether documents related to payment have been created by a generating AI in order to prevent fraudulent use and misuse during electronic payments. This makes it possible to prevent fraudulent use and misuse during electronic payments.

[0508] "User" refers to any individual or entity that uses the Service or System.

[0509] "Document" refers to any paper or file containing information, including formats such as PDF and DOCX.

[0510] "Means of converting to text" refers to tools and technologies used to analyze the contents of document files and convert them to plain text. Specifically, text extraction libraries are used.

[0511] "Natural language processing technology" refers to a set of techniques and methods that enable computers to understand and analyze human language, including NLP libraries and machine learning models.

[0512] The "copy-paste rate" is a measure of how closely the text in a document matches other sources on the Internet.

[0513] "Generative AI characteristics" refers to specific patterns, style, structure, and other characteristics found in documents created by generative AI.

[0514] The "AI generation rate" is a numerical indicator that indicates the likelihood that a document was created by a generative AI.

[0515] "Unauthorized use" refers to the intentional and unfair use of a system or service.

[0516] "Fraud" is the act of deceiving others through dishonest means in order to gain profits.

[0517] "Payment" refers to the process of making monetary transactions for goods and services.

[0518] "Means of determination" refers to the technology or algorithms used to assess whether a document was generated by AI.

[0519] MODE FOR CARRYING OUT THE INVENTION

[0520] This invention provides a system that allows users to upload documents related to electronic payments and performs a series of processes to determine whether the documents were created by a generation AI. The specific configuration and processing steps of the system are described below.

[0521] System configuration and operation

[0522] The system consists of the following components:

[0523] 1. User device: Provides an interface that runs on a smartphone, tablet, or PC and allows users to upload documents.

[0524] 2. Server: The central processing unit that receives uploaded documents, converts them to text, and performs analysis.

[0525] 3. Text conversion tools: Software that converts documents into text format (e.g., textract).

[0526] 4. Natural Language Processing (NLP) models: Libraries for analyzing text and extracting grammatical structures and features (e.g., spaCy).

[0527] 5. Generative AI Verification Model: A machine learning model to determine whether a document was created by generative AI (e.g., a BERT-based model).

[0528] 6. Copy-paste check tools: Web scraping tools and APIs that compare information on the internet to calculate copy-paste rates.

[0529] Processing flow

[0530] 1. The user uploads relevant documents (e.g., invoices, contracts) through a web interface or application provided on the user's device.

[0531] 2. The server receives the document sent by the user and converts the document into text format using a text conversion tool (e.g., textract).

[0532] 3. The server analyzes the converted text using a natural language processing (NLP) model (e.g., spaCy) to extract grammatical structures and features.

[0533] 4. The server uses a copy-paste check tool to compare the text with information on the Internet and calculate the copy-paste rate.

[0534] 5. The server uses a generative AI decision model (e.g., a BERT-based model) to determine whether the text was created by a generative AI and scores the likelihood.

[0535] 6. The server calculates the generative AI rate based on the copy-paste rate and the results of the generative AI judgment model.

[0536] 7. The server provides the user with the results of the generated AI rating and, if necessary, provides additional authentication measures and instructions for resubmission.

[0537] Hardware and software used

[0538] Server: A computer system that performs major processing, often deployed on a cloud service.

[0539] User devices: smartphones, tablets, computers, etc.

[0540] software:

[0541] Textract: A tool that converts documents into text.

[0542] SpaCy: A library for natural language processing.

[0543] BERT: A machine learning model that determines whether a document was created by generative AI.

[0544] Web scraping tools or APIs: Tools for copy-paste checking.

[0545] Specific examples

[0546] Specific examples of prompts include the following scenarios:

[0547] "Let's say a user uploads a file named 'invoice.pdf'. This file is uploaded from a smartphone and analyzed by an artificial intelligence (AI)-based fraud detection system. The server receives the file, converts and analyzes the text content, and then makes a judgment on the AI-generated document. The analysis result is sent to the payment provider, which then takes the necessary action."

[0548] As described above, by using this system, it is possible to prevent unauthorized use and fraudulent acts related to electronic payments.

[0549] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0550] Step 1:

[0551] A user uploads a document

[0552] Users upload relevant documents (e.g., invoices or contracts) from an application or web interface on their smartphone, tablet, or PC. The input is a document file such as PDF or DOCX, and the output is the uploaded document file. Specific operations involve selecting a document using the file selection button and clicking the upload button.

[0553] Step 2:

[0554] The server receives the document

[0555] The server receives the document sent by the user. The input is the document file sent by the user, and the output is a document file stored on the server side. Specifically, the server receives an HTTP POST request and retrieves the document file contained therein.

[0556] Step 3:

[0557] The server converts the document into text

[0558] The server uses a text conversion tool (e.g., textract) to convert the document file into a text format. The input is a document file in PDF or DOCX format, and the output is plain text data. Specifically, the text extraction library parses the document file and generates plain text.

[0559] Step 4:

[0560] The server parses the text

[0561] The server uses a natural language processing (NLP) model (e.g., spaCy) to analyze the converted text. The input is plain text data, and the output is an analysis result including grammatical elements and features. Specifically, the NLP model tokenizes the text and extracts grammatical elements such as subject, predicate, and object.

[0562] Step 5:

[0563] The server compares the information on the Internet and calculates the copy-paste rate

[0564] The server uses copy-paste checking tools and APIs to match the text with online information and calculate the copy-paste rate. The input is the analyzed text and multiple online sources, and the output is a number indicating the copy-paste rate. Specifically, the server sends the text to the web scraping tool, finds matches, and calculates the match rate.

[0565] Step 6:

[0566] The server uses the generative AI judgment model to determine the characteristics of the generative AI.

[0567] The server uses a generative AI judgment model (e.g., a BERT-based model) to determine whether a document was created by a generative AI. The input is the parsed text, and the output is a score indicating the likelihood of the generative AI. Specifically, the server inputs the text into the judgment model and obtains the scored result.

[0568] Step 7:

[0569] The server calculates the AI ​​generation rate

[0570] The server calculates the AI ​​generation rate based on the copy-paste rate and the results of the AI ​​generation judgment model. The input is the copy-paste rate and a score indicating the AI ​​generation potential, and the output is a number indicating the AI ​​generation rate. Specifically, the server applies an algorithm that combines the copy-paste rate and the score to calculate the AI ​​generation rate.

[0571] Step 8:

[0572] The server provides the results to the user

[0573] The server provides users with the results of the AI ​​generation rate and detailed analysis results. The input is a numerical value indicating the AI ​​generation rate and detailed analysis data, and the output is a report format that can be viewed by users. Specifically, the server formats the analysis results into a report in HTML or PDF format and displays it on the user's terminal interface.

[0574] The above is a specific description of the processing steps in the system of the invention.

[0575] 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.

[0576] This invention relates to a system that determines whether documents such as resumes and school reports were created by generative AI (generative artificial intelligence), and by combining it with an emotion engine that recognizes the user's emotions, it provides more advanced judgment and feedback. This system improves the user experience by adding analysis using the user's emotion data in addition to the series of processes from uploading the document to analyzing it and providing the results.

[0577] System Overview

[0578] 1. User uploads a document

[0579] Users upload their resumes and reports on the system's web interface (client terminal). Upload methods include a file selection button and meta-information input fields (university name, subjects, etc.).

[0580] Specific examples

[0581] A user accesses the web page, selects "Report.docx" in the upload form, enters the university name "General University" and the subject "Information Technology," and presses the upload button to send the document and information to the server.

[0582] 2. The server receives the file and converts it to text

[0583] The server receives the file sent by the user and converts the document from PDF or DOCX format to text format, for example using a PDF to text converter or a DOCX parser.

[0584] Specific examples

[0585] The server receives the "report.docx" and converts it to text using an appropriate library (e.g. Apache POI), which stores the converted text in memory.

[0586] 3. The server parses the text

[0587] The server applies natural language processing techniques to the converted text to analyze sentence structure and features, which may involve the use of NLP libraries (e.g., spaCy, NLTK).

[0588] Specific examples

[0589] The server analyzes the converted text and extracts grammatical elements such as the subject, predicate, and object of each sentence. It also evaluates the writing style and tone and calculates features (word frequency, sentence diversity, etc.).

[0590] 4. The server performs a generated AI check.

[0591] The server performs a spawned AI check by:

[0592] Copy-paste check: Use web scraping tools and APIs to collect information from the internet and compare it with text. Find matches and calculate the copy-paste rate.

[0593] AI document judgment: Using a machine learning model, we judge the possibility of generative AI based on writing style and characteristics. We score the possibility of generative AI for each sentence and calculate the overall generative AI rate.

[0594] Specific examples

[0595] The server uses a non-destructive scraping tool to collect information from the internet, calculates the match rate with the text, and uses a BERT-based machine learning model to analyze the style of the text and evaluate the likelihood that it is generative AI.

[0596] 5. The server performs analysis using the emotion engine

[0597] When a user uploads a document, the emotion engine recognizes the user's emotions from their facial expressions and input. This emotion data is then incorporated into the analysis and used to calculate the AI ​​generation rate and provide feedback.

[0598] Specific examples

[0599] When a user uploads a document, their facial expressions are analyzed via a webcam to obtain emotional data. If the user shows signs of anxiety, the server uses this information to adjust the threshold for AI generation rates and provide accurate feedback.

[0600] 6. The server provides the results to the user

[0601] The server generates a detailed report based on the analysis results and provides it to the user. If the AI ​​generation rate exceeds a certain threshold (e.g., 70%), information including instructions for resubmission and a warning is provided. The server can also adjust the way results are presented based on emotion data.

[0602] Specific examples

[0603] The server creates a report summarizing the results of 70% AI generation rates and their details (such as high match rates and characteristics of the AI-generated writing style) and displays it on the user's web interface. If the user shows any anxiety, it provides a gentle explanation and offers additional support options.

[0604] This concludes the description of a specific embodiment of the system according to the present invention. This system uses generative AI to evaluate the validity of documents with high accuracy, and also takes into account the user's emotions to provide more personalized feedback.

[0605] The processing flow will be explained below.

[0606] Step 1:

[0607] A user accesses the upload form on the web page, selects "Report.docx", enters meta information such as the university name and subject (e.g., university name "General University", subject "Information Technology"), and presses the submit button.

[0608] Step 2:

[0609] The server saves the "report.docx" file and meta information received from the user to the file system, and then converts the saved file to a text format using an appropriate tool (e.g., Apache POI).

[0610] Step 3:

[0611] The server stores the converted text in memory and begins analyzing it using an NLP (Natural Language Processing) library (e.g., spaCy, NLTK). For each sentence, grammatical elements such as subject, predicate, and object are extracted, and features for evaluating style and tone are calculated.

[0612] Step 4:

[0613] When a user uploads a document, the emotion engine recognizes the user's emotion from the webcam, keyboard typing speed, etc. This emotion data is analyzed using specific algorithms (e.g., facial expression recognition algorithms and typing speed analysis).

[0614] Step 5:

[0615] To perform the copy-paste check, the server collects information from the Internet using web scraping tools or APIs (e.g., Google Custom Search API). The collected information is compared with the analyzed text to find matching parts, and the matching rate is calculated as the copy-paste rate.

[0616] Step 6:

[0617] The server uses a machine learning model (e.g., a BERT-based model) to perform AI document judgment based on the analyzed text. It scores the likelihood of generative AI based on features such as consistency of writing style and excessive use of specific keywords, and makes a judgment for each sentence.

[0618] Step 7:

[0619] The server combines the copy-paste rate and the AI ​​document judgment score to calculate the overall generative AI rate, where it adjusts the threshold based on the recognized user emotion data to enable personalized feedback.

[0620] Step 8:

[0621] The server generates the AI ​​generation rate results and a detailed report (e.g., high match rate, stylistic features of the AI-generated writing) and sends it to the user's web interface. If the user shows signs of nervousness or anxiety, the report includes additional support options. The user is then instructed to review the results and resubmit the document if necessary.

[0622] This is the specific processing flow of the generative AI checker system that combines an emotion engine. This system can evaluate the validity and originality of documents with high accuracy, and provide personalized feedback that takes the user's emotions into account.

[0623] Example 2

[0624] 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."

[0625] Conventional document assessment systems have difficulty accurately determining whether a document was created by a generative AI. Furthermore, they lack a mechanism for taking the user's feelings into consideration when providing the user with a result of whether a document was AI-generated. Therefore, there is a need to improve the reliability of assessment results and the user experience.

[0626] 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.

[0627] In this invention, the server includes means for users to upload their own documents, means for converting the documents into text, means for analyzing the text using natural language processing technology, means for comparing the text with information on the Internet to calculate the copy-paste rate, means for determining the possibility of a generative AI from writing style and characteristics using a machine learning model, means for acquiring user emotion data and incorporating it into the analysis, means for calculating a generative AI rate based on the copy-paste rate, the generative AI determination result, and the emotion data, and means for providing the result of the generative AI rate to the user. This increases the accuracy of determining documents created by the generative AI and makes it possible to provide feedback that takes into account the user's emotions.

[0628] "Document" refers to a formal document such as a resume or report uploaded by a User.

[0629] "Text" refers to character string data converted from document formats such as PDF and DOCX.

[0630] "Natural language processing technology" refers to the technology of analyzing text and evaluating grammatical elements and writing style, and uses NLP libraries, etc.

[0631] The "copy-paste rate" refers to the rate at which the text matches when compared with information on the Internet.

[0632] "Generative AI" refers to text or documents created using a generative AI model.

[0633] A "machine learning model" refers to a technology that uses algorithms to learn patterns from large amounts of data and determine the characteristics of the generated AI.

[0634] "Emotion data" refers to data that indicates the emotional state of the user that can be obtained from their facial expressions and input content.

[0635] "Generative AI Rate" refers to the percentage that indicates the likelihood that the text was created by a generative AI.

[0636] "Feedback" refers to information that provides analysis results and advice to users.

[0637] This invention is a system that determines whether documents such as resumes and school reports were created by generative AI, and by combining it with an emotion engine that recognizes the user's emotions, it provides more advanced judgment and feedback. This system improves the user experience by adding analysis using the user's emotion data in addition to the series of processes from uploading the document to analyzing it and providing the results.

[0638] First, the user uploads their document through the device's web interface. The uploading method includes a file selection button and meta-information input fields (university name, subject, etc.). Specifically, the user accesses the web page, selects "Report.docx" in the upload form, and enters the university name "General University" and the subject "Information Technology." Then, by pressing the upload button, the document and information are sent to the server.

[0639] The server receives the file sent by the user and converts it from file format (PDF or DOCX) to text format using the appropriate library. For example, the server receives a "report.docx" and converts it to text using the Apache POI library. This converted text is stored in memory.

[0640] The server then applies natural language processing techniques to the converted text to analyze the structure and characteristics of the sentences. NLP libraries such as spaCy and NLTK are used. For example, the server analyzes the converted text and extracts grammatical elements such as the subject, predicate, and object of each sentence. It also evaluates the writing style and tone and calculates features (word frequency, sentence diversity, etc.).

[0641] In addition, the server performs copy-paste checks and AI document assessments. Copy-paste checks involve collecting information on the Internet using web scraping tools and APIs and calculating the match rate with the text. As a specific example, the server collects information on the Internet using a non-destructive scraping tool and calculates the match rate with the text. AI document assessment uses a BERT-based machine learning model to analyze the writing style and evaluate the possibility that the text is generative.

[0642] The system also performs analysis using an emotion engine. When a user uploads a document, their facial expressions are analyzed via a webcam to capture emotional data. This emotional data is incorporated into the analysis and reflected in the calculation of the AI ​​generation rate and the feedback content. For example, if the user expresses anxiety, the server can use this information to adjust the AI ​​generation rate threshold and provide more accurate feedback.

[0643] Finally, the server generates a detailed report based on the analysis results and provides it to the user. If the AI ​​generation rate exceeds a certain threshold (e.g., 70%), information including instructions for resubmission and a warning is provided. The server can also adjust the way results are presented based on emotional data. For example, the server could create a report summarizing the results of a 70% AI generation rate and its details (high match rate, characteristics of the AI-generated writing style, etc.) and display it on the user's web interface. Users who express anxiety could be offered gentle explanations and additional support options.

[0644] The system uses generative AI to accurately assess the legitimacy of documents and also takes user sentiment into account to provide more personalized feedback.

[0645] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0646] Step 1:

[0647] The user inputs the document file and meta information and presses the upload button on the device's web interface. This is the process of inputting a file called "report.docx" and meta information such as the university name "general university" and the subject "information technology," and then clicking the upload button.

[0648] Input: Document file (e.g., report.docx), meta information (university name, subject)

[0649] Output: The file and meta information are sent to the server.

[0650] Step 2:

[0651] The server receives the file sent by the user and converts it from PDF or DOCX format to a text format using an appropriate library (e.g., Apache POI). This converted text data is temporarily stored in memory.

[0652] Input: Document file

[0653] Data processing: Converting from file format to text format (using Apache POI)

[0654] Output: Text format data (temporarily saved in memory)

[0655] Step 3:

[0656] The server analyzes the converted text, using natural language processing techniques (e.g., spaCy, NLTK) to extract grammatical elements (e.g., subject, predicate, object) and calculate features of style and tone (e.g., word frequency, sentence diversity).

[0657] Input: Text format data

[0658] Data processing: Extracting grammatical elements and analyzing style and tone (using spaCy and NLTK)

[0659] Output: Analysis results (grammatical elements, style and tone features)

[0660] Step 4:

[0661] The server performs copy-paste checks, collects information from the internet using web scraping tools and APIs, compares it with text, and calculates the match rate. In addition, it analyzes writing style using a BERT-based machine learning model to evaluate the potential of generative AI.

[0662] Input: Text format data

[0663] Data Computing: Web scraping and text matching, matching rate calculation, stylistic analysis with the BERT model

[0664] Output: Match rate, possibility evaluation score of the generated AI

[0665] Step 5:

[0666] When a user uploads a document, their facial expression is captured using the device's webcam. The server uses an emotion engine to analyze the facial expression data and obtain emotional data, which is then used for analysis.

[0667] Input: Facial expression capture data

[0668] Data calculation: Analysis of facial expression data (using emotion engine)

[0669] Output: Emotion data

[0670] Step 6:

[0671] The server calculates the AI ​​generation rate by integrating and evaluating the copy-paste rate, stylistic analysis results, and emotional data. A specific algorithm is used to calculate the AI ​​generation rate.

[0672] Input: Match rate, generated AI evaluation score, emotion data

[0673] Data calculation: Algorithm-based data integration and generation AI rate calculation

[0674] Output: Generation AI rate

[0675] Step 7:

[0676] The server generates a detailed report based on the analysis results and provides it to the user. If the AI ​​generation rate exceeds a certain threshold (e.g., 70%), the report will include information such as instructions for resubmission and warnings. The server will also adjust the presentation of results based on emotion data.

[0677] Input: AI generation rate, emotion data

[0678] Output: Detailed report, feedback (resubmission instructions, warnings, etc.)

[0679] A feature of this system is that it takes user emotions into consideration while performing highly accurate document judgment using generative AI and providing personalized feedback to users.

[0680] (Application example 2)

[0681] 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."

[0682] In modern logistics centers, verifying the authenticity of package documents is important, but there is a risk of documents created by generative AI being mixed in. Furthermore, the emotions and stress of inspectors can affect work efficiency and accuracy, but there is no system that effectively monitors and provides feedback. Therefore, a system that combines package document authenticity verification and worker emotion recognition is needed.

[0683] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 2 is realized by the following means.

[0684] In this invention, the server includes means for users to upload their own documents, means for converting the documents into text, means for analyzing the text using natural language processing technology, means for calculating the copy-paste rate by comparing it with information on the Internet, means for determining the characteristics of the generating AI, means for calculating the generating AI rate based on the copy-paste rate and the determination result, means for providing the generating AI rate result to the user, means for recognizing the user's emotions, and means for adjusting the method for presenting the generating AI rate result based on the emotion data. This makes it possible to evaluate the legitimacy of package documents with high accuracy and provide feedback that takes into account the emotions of inspection workers.

[0685] "User" means a person or institution that uses the System to upload documents and obtain analysis results thereof.

[0686] "Document" means any electronic or paper information that can be converted into text format, including, for example, a resume, report, or shipping instructions.

[0687] The "means for converting to text" is a mechanism for converting an image document into text data using optical character recognition (OCR) technology.

[0688] "Natural language processing technology" is a technology for analyzing text data and extracting and analyzing grammatical structure and meaning, and specifically includes morphological analysis, grammatical analysis, and semantic analysis.

[0689] The "copy-paste rate" is the percentage of a document that has been copied in whole or in part from another existing document, and is calculated by comparing it with sources on the Internet.

[0690] "Means for determining characteristics of generative artificial intelligence" refers to techniques that use machine learning models to analyze the style and structure of documents in order to identify whether they were created by generative AI.

[0691] The "generated AI rate" is an indicator of the likelihood that a document was created by a generative AI, and is calculated based on the copy-paste rate and the AI ​​judgment results.

[0692] "Means for recognizing the user's emotions" refers to technology for extracting and analyzing emotions from the user's facial expressions and voice data, and typically involves using a camera or microphone.

[0693] The "means for adjusting the method of presenting results" is a mechanism that changes the way the analysis results are displayed based on the user's emotional data and provides optimal feedback.

[0694] A head-mounted display is a device worn on the head that displays information directly into the field of view. It also has built-in cameras and sensors that can acquire data on the external environment in real time.

[0695] The system for implementing this invention aims to control quality in logistics centers. Specifically, it determines whether shipping instructions and content descriptions included in packages were created using generative AI, and further improves work efficiency by providing feedback using the emotion data of inspection workers.

[0696] 1. System Configuration

[0697] The system includes the following main components:

[0698] 1.1 Head-mounted display (HMD)

[0699] An inspection worker wearing an HMD (e.g., a typical smart HMD device) scans the shipping instructions and contents description inside each package.

[0700] 1.2 Camera and microphone

[0701] The HMD's built-in camera scans documents and converts the images into text using OCR technology (e.g., Tesseract OCR).The camera and microphone also capture the inspector's facial expressions and voice data.

[0702] 1.3 Server

[0703] The server converts the received image data into text, analyzes grammatical structure and writing style using natural language processing (NLP) technology (e.g., spaCy), and uses a generative AI model (BERT-based model) to determine whether the document was created by generative AI and calculates the generative AI rate.

[0704] 1.4 Emotion Engine

[0705] An emotion engine running on the server analyzes the emotional data of inspection workers and adjusts the feedback content.

[0706] 2. Explanation of the processing flow

[0707] The system process flow is as follows:

[0708] 2.1 Image to Text Conversion

[0709] An inspector wearing an HMD scans the package documents with a camera and sends the image data to a server, which then uses OCR technology to convert the image data into text.

[0710] 2.2 Analysis using natural language processing

[0711] The server applies NLP techniques to the converted text to extract grammatical structures and features, including analyzing the subject, predicate, and object of sentences, and assessing sentence variety and tone.

[0712] 2.3 Generation AI judgment

[0713] The server uses the generative AI model to determine whether a document was created by generative AI, performs copy-paste checks and AI document determination functions, and calculates the generative AI rate.

[0714] 2.4 Sentiment Analysis and Feedback

[0715] The server uses an emotion engine to analyze the inspector's emotional data, and if the inspector shows signs of stress or fatigue, it adjusts the feedback method and provides appropriate alerts and suggestions (e.g., to take a break).

[0716] 3. Specific Examples

[0717] As a concrete example, an inspection worker scans a shipping instruction sheet with the file name "Doc123.jpg" using an HMD and sends it to the server. The server then uses OCR technology to convert the image into text and calculates the AI ​​generation rate. At the same time, emotional data displayed by the worker is also collected and analyzed, and feedback based on the results is displayed on the HMD display.

[0718] "Please determine if this report was created by a generative AI." is an example of a prompt sentence, which is used as input to a generative AI model.

[0719] Names of specific hardware and software used

[0720] HMD: General smart HMD devices (e.g. Microsoft HoloLens)

[0721] Camera and microphone: Camera and microphone built into the HMD

[0722] OCR technology: Tesseract OCR

[0723] NLP technology: spaCy

[0724] Generative AI model: BERT-based generative AI decision model

[0725] Emotion Engine: Emotion Recognition Library (OpenCV and Dlib)

[0726] In this way, the present invention provides an efficient and highly accurate system for evaluating package documents through quality control and worker emotion recognition in logistics centers.

[0727] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[0728] Step 1:

[0729] The user wears a head-mounted display (HMD) and scans shipping instructions and content descriptions.

[0730] Input: Image of shipping instructions or contents description

[0731] Output: Image data

[0732] Specific operation: The camera built into the HMD is used to take a photo of the document inside the package and obtain image data.

[0733] Step 2:

[0734] The server receives the image data and converts the image into text using optical character recognition (OCR) technology.

[0735] Input: Image data

[0736] Output: Text data

[0737] What happens: The server uses an OCR library (e.g., Tesseract) to analyze the image data and extract the text within the document.

[0738] Step 3:

[0739] The server analyzes the extracted text using natural language processing (NLP) techniques.

[0740] Input: Text data

[0741] Output: Grammatical structure and features

[0742] What it does: The server applies NLP libraries (e.g., spaCy) to extract grammatical elements (e.g., subject, predicate, object) in the text and evaluate sentence variety and tone.

[0743] Step 4:

[0744] The server checks for copy-paste and determines the AI ​​that generated the text data.

[0745] Input: Text data

[0746] Output: Generated AI rate

[0747] How it works: The server uses a copy-paste check tool to check against information on the internet and calculates the percentage of matches. It also uses a generative AI model (based on BERT) to analyze the style of the text and calculates the generative AI rate.

[0748] Step 5:

[0749] The server acquires the emotion data of the inspector using the camera and microphone built into the HMD.

[0750] Input: Facial expression and voice data

[0751] Output: Emotion data

[0752] How it works: The server uses the Emotion Recognition Library (OpenCV and Dlib) to analyze facial expressions and voice captured by the camera and microphone to identify the worker's emotional state.

[0753] Step 6:

[0754] The server adjusts the way results are presented based on the emotional data and provides optimal feedback to the user.

[0755] Input: Generated AI rate and emotion data

[0756] Output: Feedback results

[0757] Specific operation: The server integrates the generated AI rate and emotion data and generates a report based on the analysis results. Based on the emotion data, the format and tone of the results displayed on the HMD are adjusted to provide optimal information to the user. For example, if a worker is feeling stressed, a gentle message or a suggestion to take a break is displayed.

[0758] Through these steps, the system is able to evaluate the validity of package documents and provide feedback to improve work efficiency.

[0759] 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.

[0760] 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.

[0761] 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.

[0762] [Third embodiment]

[0763] FIG. 5 shows an example of the configuration of a data processing system 310 according to the third embodiment.

[0764] 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.

[0765] 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).

[0766] 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.

[0767] 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.

[0768] 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).

[0769] 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.

[0770] 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.

[0771] 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.

[0772] 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.

[0773] 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.

[0774] 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."

[0775] This invention relates to a system that determines whether a document, such as a resume or a school report, was created by generative AI (generative artificial intelligence). The system performs a series of steps to determine whether the document was created by a user who uploads the document and analyzes it.

[0776] System Overview

[0777] 1. User uploads a document

[0778] Users upload their resumes and reports on the system's web interface (client terminal). Upload methods include a file selection button and meta-information input fields (university name, subjects, etc.).

[0779] Specific examples

[0780] A user accesses the web page, selects "Report.docx" in the upload form, enters the university name "General University" and the subject "Information Technology," and presses the upload button to send the document and information to the server.

[0781] 2. The server receives the file and converts it to text

[0782] The server receives the file sent by the user and converts the document from PDF or DOCX format to text format, for example using a PDF to text converter or a DOCX parser.

[0783] Specific examples

[0784] The server receives the "report.docx" and converts it to text using an appropriate library (e.g. Apache POI), which stores the converted text in memory.

[0785] 3. The server parses the text

[0786] The server applies natural language processing techniques to the converted text to analyze sentence structure and features, which may involve the use of NLP libraries (e.g., spaCy, NLTK).

[0787] Specific examples

[0788] The server analyzes the converted text and extracts grammatical elements such as the subject, predicate, and object of each sentence. It also evaluates the writing style and tone and calculates features (word frequency, sentence diversity, etc.).

[0789] 4. The server performs a generated AI check.

[0790] The server performs a spawned AI check by:

[0791] Copy-paste check: Use web scraping tools and APIs to collect information from the internet and compare it with text. Find matches and calculate the copy-paste rate.

[0792] AI document judgment: Using a machine learning model, we judge the possibility of generative AI based on writing style and characteristics. We score the possibility of generative AI for each sentence and calculate the overall generative AI rate.

[0793] Specific examples

[0794] The server uses a non-destructive scraping tool to collect information from the internet, calculates the match rate with the text, and uses a BERT-based machine learning model to analyze the style of the text and evaluate the likelihood that it is generative AI.

[0795] 5. The server provides the results to the user

[0796] The server generates a detailed report based on the analysis results and provides it to the user. If the generated AI rate exceeds a certain threshold (e.g., 70%), information including instructions for resubmission and a warning will be provided.

[0797] Specific examples

[0798] The server creates a report summarizing the results of the 70% AI generation rate and details (such as high match rate and characteristics of the AI-generated writing style) and displays it on the user's web interface, allowing the user to check the results and resubmit the document if necessary.

[0799] This concludes the description of a specific embodiment of the system according to the present invention, which is a powerful tool for evaluating the legitimacy of documents with high accuracy using generative AI.

[0800] The processing flow will be explained below.

[0801] Step 1:

[0802] A user accesses the upload form on the web page, selects "Report.docx", enters meta information such as the university name and subject (e.g., university name "General University", subject "Information Technology"), and presses the submit button.

[0803] Step 2:

[0804] The server saves the "report.docx" file and meta information received from the user to the file system, and then converts the saved file to a text format using an appropriate tool (e.g., Apache POI).

[0805] Step 3:

[0806] The server stores the converted text in memory and begins analyzing it using an NLP (Natural Language Processing) library (e.g., spaCy, NLTK). For each sentence, grammatical elements such as subject, predicate, and object are extracted, and features for evaluating style and tone are calculated.

[0807] Step 4:

[0808] To perform the copy-paste check, the server collects information from the Internet using web scraping tools or APIs (e.g., Google Custom Search API). The collected information is compared with the analyzed text to find matching parts, and the matching rate is calculated as the copy-paste rate.

[0809] Step 5:

[0810] The server uses a machine learning model (e.g., a BERT-based model) to perform AI document judgment based on the analyzed text. It scores the likelihood of generative AI based on features such as consistency of writing style and excessive use of specific keywords, and makes a judgment for each sentence.

[0811] Step 6:

[0812] The server combines the copy-paste rate and the AI ​​document judgment score to calculate the overall AI generation rate, and determines whether the AI ​​generation rate exceeds a certain threshold (e.g., 70%).

[0813] Step 7:

[0814] The server generates the AI ​​generation rate results and a detailed report (e.g., high match rate, stylistic features of the AI-generated text) and sends them to the user's web interface. The user can review the results and receive instructions to resubmit the document if necessary.

[0815] This is the specific processing flow of the generative AI checker system, which makes it possible to evaluate the legitimacy and originality of documents with high accuracy.

[0816] Example 1

[0817] 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."

[0818] In recent years, advances in generative AI technology have led to an increase in the number of documents, such as resumes and reports, automatically generated by AI. These documents may lack originality and reliability, posing a major problem for educational institutions and businesses. However, existing methods lack effective methods for identifying and detecting such AI-generated documents, which poses the risk of incorrect evaluations and judgments.

[0819] 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.

[0820] In this invention, the server includes a means for users to upload their own documents, a means for converting the documents into text, a means for analyzing the text using natural language processing technology, a means for collecting information on the Internet and comparing it with the text to calculate a match rate, a means for determining the characteristics of the generative AI using a machine learning model, a means for calculating the generative AI rate based on the match rate and the determination result, and a means for providing the result of the generative AI rate to the user, thereby enabling highly accurate evaluation and determination of whether a document was created by a generative AI.

[0821] "User" means a person or entity that accesses the System and uploads their own documents for analysis.

[0822] "Document" refers to a file containing text data, such as a resume or report, that a user uploads to the system.

[0823] "Means of converting to text" refers to software or tools used to convert document formats such as PDF or DOCX into text format.

[0824] "Natural language processing technology" refers to technology for analyzing text data and extracting its grammatical structure and characteristics.

[0825] "Information on the Internet" refers to publicly available information accessible via the Internet, such as web pages and online databases.

[0826] "Match rate" refers to an indicator of the degree to which the text matches the content of the uploaded document and information on the Internet.

[0827] A "machine learning model" refers to an algorithm or mathematical model that learns from data and automatically performs specific tasks.

[0828] "Generative AI characteristics" refers to the stylistic and structural features that are unique to documents created by generative AI.

[0829] "Generative AI Rate" refers to a score or percentage that indicates the likelihood that a document was created by a generative AI.

[0830] "Warning" refers to a message or information that notifies users when a document is likely to have been created by generative AI.

[0831] "Resubmission" refers to a request by a user to submit a different document if the document is likely to have been created by generative AI.

[0832] This invention relates to a system that determines whether a document, such as a resume or a school report, was created by generative AI. The system performs this determination through a series of steps in which a user uploads the document and analyzes it.

[0833] System configuration

[0834] The system includes the following major components:

[0835] 1. User's device

[0836] 2. Server

[0837] 3. Natural Language Processing Libraries

[0838] 4. Machine Learning Models

[0839] System Details

[0840] Document upload

[0841] Users upload documents using a web interface that includes a file selection button and fields for entering meta information such as university name and subject name.

[0842] As a concrete example, a user accesses a web page, selects "Report.docx" in the upload form, enters the university name "General University" and the subject "Information Technology," and then presses the upload button, which sends the document and this information to the server.

[0843] Text conversion

[0844] The server converts the received file to text format using a PDF to text converter or a DOCX parser (e.g., Apache POI).

[0845] For example, the server receives a "report.docx" and uses the Apache POI library to parse the document into text format, which is then stored in memory.

[0846] Text Analysis

[0847] The server analyzes the text using natural language processing techniques, using natural language processing libraries (e.g., spaCy, NLTK).

[0848] For example, the server analyzes text data and extracts grammatical elements such as subject, predicate, and object of each sentence. It also evaluates writing style and tone and calculates features such as word frequency and sentence diversity.

[0849] Generation AI Check

[0850] The server performs a generated AI check using the following means:

[0851] 1. Copy-paste check: Use web scraping tools and APIs to collect information from the internet and compare it with text.

[0852] 2. AI document assessment: Using a machine learning model (e.g., a BERT-based model), we assess the potential for generative AI based on writing style and features.

[0853] For example, the server uses a non-destructive scraping tool to collect information from the internet, calculates the match rate with the text, and uses a BERT-based machine learning model to analyze the style of the text and score the potential of the generative AI.

[0854] Providing results

[0855] The server generates a detailed report based on the analysis results and provides it to the user. If the generated AI rate exceeds a certain threshold (e.g., 70%), the server provides the user with information including instructions for resubmission and a warning.

[0856] For example, the server generates a report summarizing the results of a 70% AI generation rate and details (such as high match rates and characteristics of the AI-generated writing style) and displays it on the user's web interface, allowing the user to check the results and resubmit the document if necessary.

[0857] This system is a powerful tool for evaluating the legitimacy of documents with high accuracy using generative AI, and will contribute to ensuring the reliability of documents in educational institutions and companies.

[0858] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0859] Step 1:

[0860] The user uploads their own document. The user accesses the web interface, selects the document file (e.g., report.docx) using the file selection button, and enters meta information such as the university name and subject in the input fields. Specifically, the user uploads "report.docx," enters "general university" and "information technology," and presses the upload button. This sends the file and meta information as input to the server.

[0861] Step 2:

[0862] The server converts the received file into text format. The server uses a PDF to text conversion tool or a DOCX parser (e.g., Apache POI) to convert the received document file from PDF or DOCX format to text format. Specifically, the server converts "report.docx" to text using the conversion tool and saves it in memory. The input of this step is the file data, and the output is the converted text data.

[0863] Step 3:

[0864] The server analyzes the converted text. It uses natural language processing techniques (e.g., spaCy, NLTK) to analyze grammatical structure and stylistic features. Specifically, the server passes the text data to an NLP library to extract grammatical elements such as the subject, predicate, and object of each sentence. It also calculates word frequency and sentence diversity. The input for this step is the text data, and the output is the analysis results (grammatical element and feature data).

[0865] Step 4:

[0866] The server performs a generation AI check. The server determines whether the document was created by a generation AI using two methods:

[0867] Copy-paste check: Collects information from the Internet using web scraping tools and APIs, compares it with the text, and calculates the match rate. Specifically, the server compares the text with multiple sources collected by the scraping tool and calculates the match rate.

[0868] AI document assessment: A BERT-based machine learning model is used to determine the likelihood of generative AI based on writing style and features. Specifically, the server uses the BERT model to score the likelihood of generative AI using the analysis results as input. The input for this step is the analysis results and information data on the Internet, and the output is the copy-paste rate and the score of the generative AI.

[0869] Step 5:

[0870] The server calculates the generation AI rate based on the analysis results and provides it to the user. If the generation AI rate exceeds a certain threshold (e.g., 70%), a detailed report is generated that includes information such as instructions for resubmission and warnings. Specifically, the server creates a report and displays the results on the user's web interface. The inputs to this step are the copy-paste rate and the generation AI score, and the output is a detailed report.

[0871] Through the above processing steps, this system can accurately evaluate the possibility that a document can be created by AI, and provide users with reliable information.

[0872] (Application example 1)

[0873] 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."

[0874] In recent years, the use of generative AI to generate documents has become commonplace, leading to an increase in fraudulent use and acts of fraud. In particular, in electronic payments, the risk of fraudulent transactions is increasing when important documents such as contracts and invoices are forged by generative AI. Effective measures to prevent such fraudulent use and acts are needed, but existing technologies cannot adequately address this issue.

[0875] 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.

[0876] In this invention, the server includes means for users to upload their own documents, means for converting the documents into text, means for analyzing the text using natural language processing technology, means for calculating the copy-paste rate by comparing it with information on the Internet, means for determining the characteristics of the generating AI, means for calculating the generating AI rate based on the copy-paste rate and the determination result, means for providing the generating AI rate result to the user, and means for determining whether documents related to payment have been created by a generating AI in order to prevent fraudulent use and misuse during electronic payments. This makes it possible to prevent fraudulent use and misuse during electronic payments.

[0877] "User" refers to any individual or entity that uses the Service or System.

[0878] "Document" refers to any paper or file containing information, including formats such as PDF and DOCX.

[0879] "Means of converting to text" refers to tools and technologies used to analyze the contents of document files and convert them to plain text. Specifically, text extraction libraries are used.

[0880] "Natural language processing technology" refers to a set of techniques and methods that enable computers to understand and analyze human language, including NLP libraries and machine learning models.

[0881] The "copy-paste rate" is a measure of how closely the text in a document matches other sources on the Internet.

[0882] "Generative AI characteristics" refers to specific patterns, style, structure, and other characteristics found in documents created by generative AI.

[0883] The "AI generation rate" is a numerical indicator that indicates the likelihood that a document was created by a generative AI.

[0884] "Unauthorized use" refers to the intentional and unfair use of a system or service.

[0885] "Fraud" is the act of deceiving others through dishonest means in order to gain profits.

[0886] "Payment" refers to the process of making monetary transactions for goods and services.

[0887] "Means of determination" refers to the technology or algorithms used to assess whether a document was generated by AI.

[0888] MODE FOR CARRYING OUT THE INVENTION

[0889] This invention provides a system that allows users to upload documents related to electronic payments and performs a series of processes to determine whether the documents were created by a generation AI. The specific configuration and processing steps of the system are described below.

[0890] System configuration and operation

[0891] The system consists of the following components:

[0892] 1. User device: Provides an interface that runs on a smartphone, tablet, or PC and allows users to upload documents.

[0893] 2. Server: The central processing unit that receives uploaded documents, converts them to text, and performs analysis.

[0894] 3. Text conversion tools: Software that converts documents into text format (e.g., textract).

[0895] 4. Natural Language Processing (NLP) models: Libraries for analyzing text and extracting grammatical structures and features (e.g., spaCy).

[0896] 5. Generative AI Verification Model: A machine learning model to determine whether a document was created by generative AI (e.g., a BERT-based model).

[0897] 6. Copy-paste check tools: Web scraping tools and APIs that compare information on the internet to calculate copy-paste rates.

[0898] Processing flow

[0899] 1. The user uploads relevant documents (e.g., invoices, contracts) through a web interface or application provided on the user's device.

[0900] 2. The server receives the document sent by the user and converts the document into text format using a text conversion tool (e.g., textract).

[0901] 3. The server analyzes the converted text using a natural language processing (NLP) model (e.g., spaCy) to extract grammatical structures and features.

[0902] 4. The server uses a copy-paste check tool to compare the text with information on the Internet and calculate the copy-paste rate.

[0903] 5. The server uses a generative AI decision model (e.g., a BERT-based model) to determine whether the text was created by a generative AI and scores the likelihood.

[0904] 6. The server calculates the generative AI rate based on the copy-paste rate and the results of the generative AI judgment model.

[0905] 7. The server provides the user with the results of the generated AI rating and, if necessary, provides additional authentication measures and instructions for resubmission.

[0906] Hardware and software used

[0907] Server: A computer system that performs major processing, often deployed on a cloud service.

[0908] User devices: smartphones, tablets, computers, etc.

[0909] software:

[0910] Textract: A tool that converts documents into text.

[0911] SpaCy: A library for natural language processing.

[0912] BERT: A machine learning model that determines whether a document was created by generative AI.

[0913] Web scraping tools or APIs: Tools for copy-paste checking.

[0914] Specific examples

[0915] Specific examples of prompts include the following scenarios:

[0916] "Let's say a user uploads a file named 'invoice.pdf'. This file is uploaded from a smartphone and analyzed by an artificial intelligence (AI)-based fraud detection system. The server receives the file, converts and analyzes the text content, and then makes a judgment on the AI-generated document. The analysis result is sent to the payment provider, which then takes the necessary action."

[0917] As described above, by using this system, it is possible to prevent unauthorized use and fraudulent acts related to electronic payments.

[0918] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0919] Step 1:

[0920] A user uploads a document

[0921] Users upload relevant documents (e.g., invoices or contracts) from an application or web interface on their smartphone, tablet, or PC. The input is a document file such as PDF or DOCX, and the output is the uploaded document file. Specific operations involve selecting a document using the file selection button and clicking the upload button.

[0922] Step 2:

[0923] The server receives the document

[0924] The server receives the document sent by the user. The input is the document file sent by the user, and the output is a document file stored on the server side. Specifically, the server receives an HTTP POST request and retrieves the document file contained therein.

[0925] Step 3:

[0926] The server converts the document into text

[0927] The server uses a text conversion tool (e.g., textract) to convert the document file into a text format. The input is a document file in PDF or DOCX format, and the output is plain text data. Specifically, the text extraction library parses the document file and generates plain text.

[0928] Step 4:

[0929] The server parses the text

[0930] The server uses a natural language processing (NLP) model (e.g., spaCy) to analyze the converted text. The input is plain text data, and the output is an analysis result including grammatical elements and features. Specifically, the NLP model tokenizes the text and extracts grammatical elements such as subject, predicate, and object.

[0931] Step 5:

[0932] The server compares the information on the Internet and calculates the copy-paste rate

[0933] The server uses copy-paste checking tools and APIs to match the text with online information and calculate the copy-paste rate. The input is the analyzed text and multiple online sources, and the output is a number indicating the copy-paste rate. Specifically, the server sends the text to the web scraping tool, finds matches, and calculates the match rate.

[0934] Step 6:

[0935] The server uses the generative AI judgment model to determine the characteristics of the generative AI.

[0936] The server uses a generative AI judgment model (e.g., a BERT-based model) to determine whether a document was created by a generative AI. The input is the parsed text, and the output is a score indicating the likelihood of the generative AI. Specifically, the server inputs the text into the judgment model and obtains the scored result.

[0937] Step 7:

[0938] The server calculates the AI ​​generation rate

[0939] The server calculates the AI ​​generation rate based on the copy-paste rate and the results of the AI ​​generation judgment model. The input is the copy-paste rate and a score indicating the AI ​​generation potential, and the output is a number indicating the AI ​​generation rate. Specifically, the server applies an algorithm that combines the copy-paste rate and the score to calculate the AI ​​generation rate.

[0940] Step 8:

[0941] The server provides the results to the user

[0942] The server provides users with the results of the AI ​​generation rate and detailed analysis results. The input is a numerical value indicating the AI ​​generation rate and detailed analysis data, and the output is a report format that can be viewed by users. Specifically, the server formats the analysis results into a report in HTML or PDF format and displays it on the user's terminal interface.

[0943] The above is a specific description of the processing steps in the system of the invention.

[0944] 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.

[0945] This invention relates to a system that determines whether documents such as resumes and school reports were created by generative AI (generative artificial intelligence), and by combining it with an emotion engine that recognizes the user's emotions, it provides more advanced judgment and feedback. This system improves the user experience by adding analysis using the user's emotion data in addition to the series of processes from uploading the document to analyzing it and providing the results.

[0946] System Overview

[0947] 1. User uploads a document

[0948] Users upload their resumes and reports on the system's web interface (client terminal). Upload methods include a file selection button and meta-information input fields (university name, subjects, etc.).

[0949] Specific examples

[0950] A user accesses the web page, selects "Report.docx" in the upload form, enters the university name "General University" and the subject "Information Technology," and presses the upload button to send the document and information to the server.

[0951] 2. The server receives the file and converts it to text

[0952] The server receives the file sent by the user and converts the document from PDF or DOCX format to text format, for example using a PDF to text converter or a DOCX parser.

[0953] Specific examples

[0954] The server receives the "report.docx" and converts it to text using an appropriate library (e.g. Apache POI), which stores the converted text in memory.

[0955] 3. The server parses the text

[0956] The server applies natural language processing techniques to the converted text to analyze sentence structure and features, which may involve the use of NLP libraries (e.g., spaCy, NLTK).

[0957] Specific examples

[0958] The server analyzes the converted text and extracts grammatical elements such as the subject, predicate, and object of each sentence. It also evaluates the writing style and tone and calculates features (word frequency, sentence diversity, etc.).

[0959] 4. The server performs a generated AI check.

[0960] The server performs a spawned AI check by:

[0961] Copy-paste check: Use web scraping tools and APIs to collect information from the internet and compare it with text. Find matches and calculate the copy-paste rate.

[0962] AI document judgment: Using a machine learning model, we judge the possibility of generative AI based on writing style and characteristics. We score the possibility of generative AI for each sentence and calculate the overall generative AI rate.

[0963] Specific examples

[0964] The server uses a non-destructive scraping tool to collect information from the internet, calculates the match rate with the text, and uses a BERT-based machine learning model to analyze the style of the text and evaluate the likelihood that it is generative AI.

[0965] 5. The server performs analysis using the emotion engine

[0966] When a user uploads a document, the emotion engine recognizes the user's emotions from their facial expressions and input. This emotion data is then incorporated into the analysis and used to calculate the AI ​​generation rate and provide feedback.

[0967] Specific examples

[0968] When a user uploads a document, their facial expressions are analyzed via a webcam to obtain emotional data. If the user shows signs of anxiety, the server uses this information to adjust the threshold for AI generation rates and provide accurate feedback.

[0969] 6. The server provides the results to the user

[0970] The server generates a detailed report based on the analysis results and provides it to the user. If the AI ​​generation rate exceeds a certain threshold (e.g., 70%), information including instructions for resubmission and a warning is provided. The server can also adjust the way results are presented based on emotion data.

[0971] Specific examples

[0972] The server creates a report summarizing the results of 70% AI generation rates and their details (such as high match rates and characteristics of the AI-generated writing style) and displays it on the user's web interface. If the user shows any anxiety, it provides a gentle explanation and offers additional support options.

[0973] This concludes the description of a specific embodiment of the system according to the present invention. This system uses generative AI to evaluate the validity of documents with high accuracy, and also takes into account the user's emotions to provide more personalized feedback.

[0974] The processing flow will be explained below.

[0975] Step 1:

[0976] A user accesses the upload form on the web page, selects "Report.docx", enters meta information such as the university name and subject (e.g., university name "General University", subject "Information Technology"), and presses the submit button.

[0977] Step 2:

[0978] The server saves the "report.docx" file and meta information received from the user to the file system, and then converts the saved file to a text format using an appropriate tool (e.g., Apache POI).

[0979] Step 3:

[0980] The server stores the converted text in memory and begins analyzing it using an NLP (Natural Language Processing) library (e.g., spaCy, NLTK). For each sentence, grammatical elements such as subject, predicate, and object are extracted, and features for evaluating style and tone are calculated.

[0981] Step 4:

[0982] When a user uploads a document, the emotion engine recognizes the user's emotion from the webcam, keyboard typing speed, etc. This emotion data is analyzed using specific algorithms (e.g., facial expression recognition algorithms and typing speed analysis).

[0983] Step 5:

[0984] To perform the copy-paste check, the server collects information from the Internet using web scraping tools or APIs (e.g., Google Custom Search API). The collected information is compared with the analyzed text to find matching parts, and the matching rate is calculated as the copy-paste rate.

[0985] Step 6:

[0986] The server uses a machine learning model (e.g., a BERT-based model) to perform AI document judgment based on the analyzed text. It scores the likelihood of generative AI based on features such as consistency of writing style and excessive use of specific keywords, and makes a judgment for each sentence.

[0987] Step 7:

[0988] The server combines the copy-paste rate and the AI ​​document judgment score to calculate the overall generative AI rate, where it adjusts the threshold based on the recognized user emotion data to enable personalized feedback.

[0989] Step 8:

[0990] The server generates the AI ​​generation rate results and a detailed report (e.g., high match rate, stylistic features of the AI-generated writing) and sends it to the user's web interface. If the user shows signs of nervousness or anxiety, the report includes additional support options. The user is then instructed to review the results and resubmit the document if necessary.

[0991] This is the specific processing flow of the generative AI checker system that combines an emotion engine. This system can evaluate the validity and originality of documents with high accuracy, and provide personalized feedback that takes the user's emotions into account.

[0992] Example 2

[0993] 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."

[0994] Conventional document assessment systems have difficulty accurately determining whether a document was created by a generative AI. Furthermore, they lack a mechanism for taking the user's feelings into consideration when providing the user with a result of whether a document was AI-generated. Therefore, there is a need to improve the reliability of assessment results and the user experience.

[0995] 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.

[0996] In this invention, the server includes means for users to upload their own documents, means for converting the documents into text, means for analyzing the text using natural language processing technology, means for comparing the text with information on the Internet to calculate the copy-paste rate, means for determining the possibility of a generative AI from writing style and characteristics using a machine learning model, means for acquiring user emotion data and incorporating it into the analysis, means for calculating a generative AI rate based on the copy-paste rate, the generative AI determination result, and the emotion data, and means for providing the result of the generative AI rate to the user. This increases the accuracy of determining documents created by the generative AI and makes it possible to provide feedback that takes into account the user's emotions.

[0997] "Document" refers to a formal document such as a resume or report uploaded by a User.

[0998] "Text" refers to character string data converted from document formats such as PDF and DOCX.

[0999] "Natural language processing technology" refers to the technology of analyzing text and evaluating grammatical elements and writing style, and uses NLP libraries, etc.

[1000] The "copy-paste rate" refers to the rate at which the text matches when compared with information on the Internet.

[1001] "Generative AI" refers to text or documents created using a generative AI model.

[1002] A "machine learning model" refers to a technology that uses algorithms to learn patterns from large amounts of data and determine the characteristics of the generated AI.

[1003] "Emotion data" refers to data that indicates the emotional state of the user that can be obtained from their facial expressions and input content.

[1004] "Generative AI Rate" refers to the percentage that indicates the likelihood that the text was created by a generative AI.

[1005] "Feedback" refers to information that provides analysis results and advice to users.

[1006] This invention is a system that determines whether documents such as resumes and school reports were created by generative AI, and by combining it with an emotion engine that recognizes the user's emotions, it provides more advanced judgment and feedback. This system improves the user experience by adding analysis using the user's emotion data in addition to the series of processes from uploading the document to analyzing it and providing the results.

[1007] First, the user uploads their document through the device's web interface. The uploading method includes a file selection button and meta-information input fields (university name, subject, etc.). Specifically, the user accesses the web page, selects "Report.docx" in the upload form, and enters the university name "General University" and the subject "Information Technology." Then, by pressing the upload button, the document and information are sent to the server.

[1008] The server receives the file sent by the user and converts it from file format (PDF or DOCX) to text format using the appropriate library. For example, the server receives a "report.docx" and converts it to text using the Apache POI library. This converted text is stored in memory.

[1009] The server then applies natural language processing techniques to the converted text to analyze the structure and characteristics of the sentences. NLP libraries such as spaCy and NLTK are used. For example, the server analyzes the converted text and extracts grammatical elements such as the subject, predicate, and object of each sentence. It also evaluates the writing style and tone and calculates features (word frequency, sentence diversity, etc.).

[1010] In addition, the server performs copy-paste checks and AI document assessments. Copy-paste checks involve collecting information on the Internet using web scraping tools and APIs and calculating the match rate with the text. As a specific example, the server collects information on the Internet using a non-destructive scraping tool and calculates the match rate with the text. AI document assessment uses a BERT-based machine learning model to analyze the writing style and evaluate the possibility that the text is generative.

[1011] The system also performs analysis using an emotion engine. When a user uploads a document, their facial expressions are analyzed via a webcam to capture emotional data. This emotional data is incorporated into the analysis and reflected in the calculation of the AI ​​generation rate and the feedback content. For example, if the user expresses anxiety, the server can use this information to adjust the AI ​​generation rate threshold and provide more accurate feedback.

[1012] Finally, the server generates a detailed report based on the analysis results and provides it to the user. If the AI ​​generation rate exceeds a certain threshold (e.g., 70%), information including instructions for resubmission and a warning is provided. The server can also adjust the way results are presented based on emotional data. For example, the server could create a report summarizing the results of a 70% AI generation rate and its details (high match rate, characteristics of the AI-generated writing style, etc.) and display it on the user's web interface. Users who express anxiety could be offered gentle explanations and additional support options.

[1013] The system uses generative AI to accurately assess the legitimacy of documents and also takes user sentiment into account to provide more personalized feedback.

[1014] The flow of the identification process in the second embodiment will be described with reference to FIG.

[1015] Step 1:

[1016] The user inputs the document file and meta information and presses the upload button on the device's web interface. This is the process of inputting a file called "report.docx" and meta information such as the university name "general university" and the subject "information technology," and then clicking the upload button.

[1017] Input: Document file (e.g., report.docx), meta information (university name, subject)

[1018] Output: The file and meta information are sent to the server.

[1019] Step 2:

[1020] The server receives the file sent by the user and converts it from PDF or DOCX format to a text format using an appropriate library (e.g., Apache POI). This converted text data is temporarily stored in memory.

[1021] Input: Document file

[1022] Data processing: Converting from file format to text format (using Apache POI)

[1023] Output: Text format data (temporarily saved in memory)

[1024] Step 3:

[1025] The server analyzes the converted text, using natural language processing techniques (e.g., spaCy, NLTK) to extract grammatical elements (e.g., subject, predicate, object) and calculate features of style and tone (e.g., word frequency, sentence diversity).

[1026] Input: Text format data

[1027] Data processing: Extracting grammatical elements and analyzing style and tone (using spaCy and NLTK)

[1028] Output: Analysis results (grammatical elements, style and tone features)

[1029] Step 4:

[1030] The server performs copy-paste checks, collects information from the internet using web scraping tools and APIs, compares it with text, and calculates the match rate. In addition, it analyzes writing style using a BERT-based machine learning model to evaluate the potential of generative AI.

[1031] Input: Text format data

[1032] Data Computing: Web scraping and text matching, matching rate calculation, stylistic analysis with the BERT model

[1033] Output: Match rate, possibility evaluation score of the generated AI

[1034] Step 5:

[1035] When a user uploads a document, their facial expression is captured using the device's webcam. The server uses an emotion engine to analyze the facial expression data and obtain emotional data, which is then used for analysis.

[1036] Input: Facial expression capture data

[1037] Data calculation: Analysis of facial expression data (using emotion engine)

[1038] Output: Emotion data

[1039] Step 6:

[1040] The server calculates the AI ​​generation rate by integrating and evaluating the copy-paste rate, stylistic analysis results, and emotional data. A specific algorithm is used to calculate the AI ​​generation rate.

[1041] Input: Match rate, generated AI evaluation score, emotion data

[1042] Data calculation: Algorithm-based data integration and generation AI rate calculation

[1043] Output: Generation AI rate

[1044] Step 7:

[1045] The server generates a detailed report based on the analysis results and provides it to the user. If the AI ​​generation rate exceeds a certain threshold (e.g., 70%), the report will include information such as instructions for resubmission and warnings. The server will also adjust the presentation of results based on emotion data.

[1046] Input: AI generation rate, emotion data

[1047] Output: Detailed report, feedback (resubmission instructions, warnings, etc.)

[1048] A feature of this system is that it takes user emotions into consideration while performing highly accurate document judgment using generative AI and providing personalized feedback to users.

[1049] (Application example 2)

[1050] 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."

[1051] In modern logistics centers, verifying the authenticity of package documents is important, but there is a risk of documents created by generative AI being mixed in. Furthermore, the emotions and stress of inspectors can affect work efficiency and accuracy, but there is no system that effectively monitors and provides feedback. Therefore, a system that combines package document authenticity verification and worker emotion recognition is needed.

[1052] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 2 is realized by the following means.

[1053] In this invention, the server includes means for users to upload their own documents, means for converting the documents into text, means for analyzing the text using natural language processing technology, means for calculating the copy-paste rate by comparing it with information on the Internet, means for determining the characteristics of the generating AI, means for calculating the generating AI rate based on the copy-paste rate and the determination result, means for providing the generating AI rate result to the user, means for recognizing the user's emotions, and means for adjusting the method for presenting the generating AI rate result based on the emotion data. This makes it possible to evaluate the legitimacy of package documents with high accuracy and provide feedback that takes into account the emotions of inspection workers.

[1054] "User" means a person or institution that uses the System to upload documents and obtain analysis results thereof.

[1055] "Document" means any electronic or paper information that can be converted into text format, including, for example, a resume, report, or shipping instructions.

[1056] The "means for converting to text" is a mechanism for converting an image document into text data using optical character recognition (OCR) technology.

[1057] "Natural language processing technology" is a technology for analyzing text data and extracting and analyzing grammatical structure and meaning, and specifically includes morphological analysis, grammatical analysis, and semantic analysis.

[1058] The "copy-paste rate" is the percentage of a document that has been copied in whole or in part from another existing document, and is calculated by comparing it with sources on the Internet.

[1059] "Means for determining characteristics of generative artificial intelligence" refers to techniques that use machine learning models to analyze the style and structure of documents in order to identify whether they were created by generative AI.

[1060] The "generated AI rate" is an indicator of the likelihood that a document was created by a generative AI, and is calculated based on the copy-paste rate and the AI ​​judgment results.

[1061] "Means for recognizing the user's emotions" refers to technology for extracting and analyzing emotions from the user's facial expressions and voice data, and typically involves using a camera or microphone.

[1062] The "means for adjusting the method of presenting results" is a mechanism that changes the way the analysis results are displayed based on the user's emotional data and provides optimal feedback.

[1063] A head-mounted display is a device worn on the head that displays information directly into the field of view. It also has built-in cameras and sensors that can acquire data on the external environment in real time.

[1064] The system for implementing this invention aims to control quality in logistics centers. Specifically, it determines whether shipping instructions and content descriptions included in packages were created using generative AI, and further improves work efficiency by providing feedback using the emotion data of inspection workers.

[1065] 1. System Configuration

[1066] The system includes the following main components:

[1067] 1.1 Head-mounted display (HMD)

[1068] An inspection worker wearing an HMD (e.g., a typical smart HMD device) scans the shipping instructions and contents description inside each package.

[1069] 1.2 Camera and microphone

[1070] The HMD's built-in camera scans documents and converts the images into text using OCR technology (e.g., Tesseract OCR).The camera and microphone also capture the inspector's facial expressions and voice data.

[1071] 1.3 Server

[1072] The server converts the received image data into text, analyzes grammatical structure and writing style using natural language processing (NLP) technology (e.g., spaCy), and uses a generative AI model (BERT-based model) to determine whether the document was created by generative AI and calculates the generative AI rate.

[1073] 1.4 Emotion Engine

[1074] An emotion engine running on the server analyzes the emotional data of inspection workers and adjusts the feedback content.

[1075] 2. Explanation of the processing flow

[1076] The system process flow is as follows:

[1077] 2.1 Image to Text Conversion

[1078] An inspector wearing an HMD scans the package documents with a camera and sends the image data to a server, which then uses OCR technology to convert the image data into text.

[1079] 2.2 Analysis using natural language processing

[1080] The server applies NLP techniques to the converted text to extract grammatical structures and features, including analyzing the subject, predicate, and object of sentences, and assessing sentence variety and tone.

[1081] 2.3 Generation AI judgment

[1082] The server uses the generative AI model to determine whether a document was created by generative AI, performs copy-paste checks and AI document determination functions, and calculates the generative AI rate.

[1083] 2.4 Sentiment Analysis and Feedback

[1084] The server uses an emotion engine to analyze the inspector's emotional data, and if the inspector shows signs of stress or fatigue, it adjusts the feedback method and provides appropriate alerts and suggestions (e.g., to take a break).

[1085] 3. Specific Examples

[1086] As a concrete example, an inspection worker scans a shipping instruction sheet with the file name "Doc123.jpg" using an HMD and sends it to the server. The server then uses OCR technology to convert the image into text and calculates the AI ​​generation rate. At the same time, emotional data displayed by the worker is also collected and analyzed, and feedback based on the results is displayed on the HMD display.

[1087] "Please determine if this report was created by a generative AI." is an example of a prompt sentence, which is used as input to a generative AI model.

[1088] Names of specific hardware and software used

[1089] HMD: General smart HMD devices (e.g. Microsoft HoloLens)

[1090] Camera and microphone: Camera and microphone built into the HMD

[1091] OCR technology: Tesseract OCR

[1092] NLP technology: spaCy

[1093] Generative AI model: BERT-based generative AI decision model

[1094] Emotion Engine: Emotion Recognition Library (OpenCV and Dlib)

[1095] In this way, the present invention provides an efficient and highly accurate system for evaluating package documents through quality control and worker emotion recognition in logistics centers.

[1096] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[1097] Step 1:

[1098] The user wears a head-mounted display (HMD) and scans shipping instructions and content descriptions.

[1099] Input: Image of shipping instructions or contents description

[1100] Output: Image data

[1101] Specific operation: The camera built into the HMD is used to take a photo of the document inside the package and obtain image data.

[1102] Step 2:

[1103] The server receives the image data and converts the image into text using optical character recognition (OCR) technology.

[1104] Input: Image data

[1105] Output: Text data

[1106] What happens: The server uses an OCR library (e.g., Tesseract) to analyze the image data and extract the text within the document.

[1107] Step 3:

[1108] The server analyzes the extracted text using natural language processing (NLP) techniques.

[1109] Input: Text data

[1110] Output: Grammatical structure and features

[1111] What it does: The server applies NLP libraries (e.g., spaCy) to extract grammatical elements (e.g., subject, predicate, object) in the text and evaluate sentence variety and tone.

[1112] Step 4:

[1113] The server checks for copy-paste and determines the AI ​​that generated the text data.

[1114] Input: Text data

[1115] Output: Generated AI rate

[1116] How it works: The server uses a copy-paste check tool to check against information on the internet and calculates the percentage of matches. It also uses a generative AI model (based on BERT) to analyze the style of the text and calculates the generative AI rate.

[1117] Step 5:

[1118] The server acquires the emotion data of the inspector using the camera and microphone built into the HMD.

[1119] Input: Facial expression and voice data

[1120] Output: Emotion data

[1121] How it works: The server uses the Emotion Recognition Library (OpenCV and Dlib) to analyze facial expressions and voice captured by the camera and microphone to identify the worker's emotional state.

[1122] Step 6:

[1123] The server adjusts the way results are presented based on the emotional data and provides optimal feedback to the user.

[1124] Input: Generated AI rate and emotion data

[1125] Output: Feedback results

[1126] Specific operation: The server integrates the generated AI rate and emotion data and generates a report based on the analysis results. Based on the emotion data, the format and tone of the results displayed on the HMD are adjusted to provide optimal information to the user. For example, if a worker is feeling stressed, a gentle message or a suggestion to take a break is displayed.

[1127] Through these steps, the system is able to evaluate the validity of package documents and provide feedback to improve work efficiency.

[1128] 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.

[1129] 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.

[1130] 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.

[1131] [Fourth embodiment]

[1132] FIG. 7 shows an example of the configuration of a data processing system 410 according to the fourth embodiment.

[1133] 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.

[1134] 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).

[1135] 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.

[1136] 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.

[1137] 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).

[1138] 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.

[1139] 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.

[1140] 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.

[1141] 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.

[1142] 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.

[1143] 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.

[1144] 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."

[1145] This invention relates to a system that determines whether a document, such as a resume or a school report, was created by generative AI (generative artificial intelligence). The system performs a series of steps to determine whether the document was created by a user who uploads the document and analyzes it.

[1146] System Overview

[1147] 1. User uploads a document

[1148] Users upload their resumes and reports on the system's web interface (client terminal). Upload methods include a file selection button and meta-information input fields (university name, subjects, etc.).

[1149] Specific examples

[1150] A user accesses the web page, selects "Report.docx" in the upload form, enters the university name "General University" and the subject "Information Technology," and presses the upload button to send the document and information to the server.

[1151] 2. The server receives the file and converts it to text

[1152] The server receives the file sent by the user and converts the document from PDF or DOCX format to text format, for example using a PDF to text converter or a DOCX parser.

[1153] Specific examples

[1154] The server receives the "report.docx" and converts it to text using an appropriate library (e.g. Apache POI), which stores the converted text in memory.

[1155] 3. The server parses the text

[1156] The server applies natural language processing techniques to the converted text to analyze sentence structure and features, which may involve the use of NLP libraries (e.g., spaCy, NLTK).

[1157] Specific examples

[1158] The server analyzes the converted text and extracts grammatical elements such as the subject, predicate, and object of each sentence. It also evaluates the writing style and tone and calculates features (word frequency, sentence diversity, etc.).

[1159] 4. The server performs a generated AI check.

[1160] The server performs a spawned AI check by:

[1161] Copy-paste check: Use web scraping tools and APIs to collect information from the internet and compare it with text. Find matches and calculate the copy-paste rate.

[1162] AI document judgment: Using a machine learning model, we judge the possibility of generative AI based on writing style and characteristics. We score the possibility of generative AI for each sentence and calculate the overall generative AI rate.

[1163] Specific examples

[1164] The server uses a non-destructive scraping tool to collect information from the internet, calculates the match rate with the text, and uses a BERT-based machine learning model to analyze the style of the text and evaluate the likelihood that it is generative AI.

[1165] 5. The server provides the results to the user

[1166] The server generates a detailed report based on the analysis results and provides it to the user. If the generated AI rate exceeds a certain threshold (e.g., 70%), information including instructions for resubmission and a warning will be provided.

[1167] Specific examples

[1168] The server creates a report summarizing the results of the 70% AI generation rate and details (such as high match rate and characteristics of the AI-generated writing style) and displays it on the user's web interface, allowing the user to check the results and resubmit the document if necessary.

[1169] This concludes the description of a specific embodiment of the system according to the present invention, which is a powerful tool for evaluating the legitimacy of documents with high accuracy using generative AI.

[1170] The processing flow will be explained below.

[1171] Step 1:

[1172] A user accesses the upload form on the web page, selects "Report.docx", enters meta information such as the university name and subject (e.g., university name "General University", subject "Information Technology"), and presses the submit button.

[1173] Step 2:

[1174] The server saves the "report.docx" file and meta information received from the user to the file system, and then converts the saved file to a text format using an appropriate tool (e.g., Apache POI).

[1175] Step 3:

[1176] The server stores the converted text in memory and begins analyzing it using an NLP (Natural Language Processing) library (e.g., spaCy, NLTK). For each sentence, grammatical elements such as subject, predicate, and object are extracted, and features for evaluating style and tone are calculated.

[1177] Step 4:

[1178] To perform the copy-paste check, the server collects information from the Internet using web scraping tools or APIs (e.g., Google Custom Search API). The collected information is compared with the analyzed text to find matching parts, and the matching rate is calculated as the copy-paste rate.

[1179] Step 5:

[1180] The server uses a machine learning model (e.g., a BERT-based model) to perform AI document judgment based on the analyzed text. It scores the likelihood of generative AI based on features such as consistency of writing style and excessive use of specific keywords, and makes a judgment for each sentence.

[1181] Step 6:

[1182] The server combines the copy-paste rate and the AI ​​document judgment score to calculate the overall AI generation rate, and determines whether the AI ​​generation rate exceeds a certain threshold (e.g., 70%).

[1183] Step 7:

[1184] The server generates the AI ​​generation rate results and a detailed report (e.g., high match rate, stylistic features of the AI-generated text) and sends them to the user's web interface. The user can review the results and receive instructions to resubmit the document if necessary.

[1185] This is the specific processing flow of the generative AI checker system, which makes it possible to evaluate the legitimacy and originality of documents with high accuracy.

[1186] Example 1

[1187] 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."

[1188] In recent years, advances in generative AI technology have led to an increase in the number of documents, such as resumes and reports, automatically generated by AI. These documents may lack originality and reliability, posing a major problem for educational institutions and businesses. However, existing methods lack effective methods for identifying and detecting such AI-generated documents, which poses the risk of incorrect evaluations and judgments.

[1189] 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.

[1190] In this invention, the server includes a means for users to upload their own documents, a means for converting the documents into text, a means for analyzing the text using natural language processing technology, a means for collecting information on the Internet and comparing it with the text to calculate a match rate, a means for determining the characteristics of the generative AI using a machine learning model, a means for calculating the generative AI rate based on the match rate and the determination result, and a means for providing the result of the generative AI rate to the user, thereby enabling highly accurate evaluation and determination of whether a document was created by a generative AI.

[1191] "User" means a person or entity that accesses the System and uploads their own documents for analysis.

[1192] "Document" refers to a file containing text data, such as a resume or report, that a user uploads to the system.

[1193] "Means of converting to text" refers to software or tools used to convert document formats such as PDF or DOCX into text format.

[1194] "Natural language processing technology" refers to technology for analyzing text data and extracting its grammatical structure and characteristics.

[1195] "Information on the Internet" refers to publicly available information accessible via the Internet, such as web pages and online databases.

[1196] "Match rate" refers to an indicator of the degree to which the text matches the content of the uploaded document and information on the Internet.

[1197] A "machine learning model" refers to an algorithm or mathematical model that learns from data and automatically performs specific tasks.

[1198] "Generative AI characteristics" refers to the stylistic and structural features that are unique to documents created by generative AI.

[1199] "Generative AI Rate" refers to a score or percentage that indicates the likelihood that a document was created by a generative AI.

[1200] "Warning" refers to a message or information that notifies users when a document is likely to have been created by generative AI.

[1201] "Resubmission" refers to a request by a user to submit a different document if the document is likely to have been created by generative AI.

[1202] This invention relates to a system that determines whether a document, such as a resume or a school report, was created by generative AI. The system performs this determination through a series of steps in which a user uploads the document and analyzes it.

[1203] System configuration

[1204] The system includes the following major components:

[1205] 1. User's device

[1206] 2. Server

[1207] 3. Natural Language Processing Libraries

[1208] 4. Machine Learning Models

[1209] System Details

[1210] Document upload

[1211] Users upload documents using a web interface that includes a file selection button and fields for entering meta information such as university name and subject name.

[1212] As a concrete example, a user accesses a web page, selects "Report.docx" in the upload form, enters the university name "General University" and the subject "Information Technology," and then presses the upload button, which sends the document and this information to the server.

[1213] Text conversion

[1214] The server converts the received file to text format using a PDF to text converter or a DOCX parser (e.g., Apache POI).

[1215] For example, the server receives a "report.docx" and uses the Apache POI library to parse the document into text format, which is then stored in memory.

[1216] Text Analysis

[1217] The server analyzes the text using natural language processing techniques, using natural language processing libraries (e.g., spaCy, NLTK).

[1218] For example, the server analyzes text data and extracts grammatical elements such as subject, predicate, and object of each sentence. It also evaluates writing style and tone and calculates features such as word frequency and sentence diversity.

[1219] Generation AI Check

[1220] The server performs a generated AI check using the following means:

[1221] 1. Copy-paste check: Use web scraping tools and APIs to collect information from the internet and compare it with text.

[1222] 2. AI document assessment: Using a machine learning model (e.g., a BERT-based model), we assess the potential for generative AI based on writing style and features.

[1223] For example, the server uses a non-destructive scraping tool to collect information from the internet, calculates the match rate with the text, and uses a BERT-based machine learning model to analyze the style of the text and score the potential of the generative AI.

[1224] Providing results

[1225] The server generates a detailed report based on the analysis results and provides it to the user. If the generated AI rate exceeds a certain threshold (e.g., 70%), the server provides the user with information including instructions for resubmission and a warning.

[1226] For example, the server generates a report summarizing the results of a 70% AI generation rate and details (such as high match rates and characteristics of the AI-generated writing style) and displays it on the user's web interface, allowing the user to check the results and resubmit the document if necessary.

[1227] This system is a powerful tool for evaluating the legitimacy of documents with high accuracy using generative AI, and will contribute to ensuring the reliability of documents in educational institutions and companies.

[1228] The flow of the identification process in the first embodiment will be described with reference to FIG.

[1229] Step 1:

[1230] The user uploads their own document. The user accesses the web interface, selects the document file (e.g., report.docx) using the file selection button, and enters meta information such as the university name and subject in the input fields. Specifically, the user uploads "report.docx," enters "general university" and "information technology," and presses the upload button. This sends the file and meta information as input to the server.

[1231] Step 2:

[1232] The server converts the received file into text format. The server uses a PDF to text conversion tool or a DOCX parser (e.g., Apache POI) to convert the received document file from PDF or DOCX format to text format. Specifically, the server converts "report.docx" to text using the conversion tool and saves it in memory. The input of this step is the file data, and the output is the converted text data.

[1233] Step 3:

[1234] The server analyzes the converted text. It uses natural language processing techniques (e.g., spaCy, NLTK) to analyze grammatical structure and stylistic features. Specifically, the server passes the text data to an NLP library to extract grammatical elements such as the subject, predicate, and object of each sentence. It also calculates word frequency and sentence diversity. The input for this step is the text data, and the output is the analysis results (grammatical element and feature data).

[1235] Step 4:

[1236] The server performs a generation AI check. The server determines whether the document was created by a generation AI using two methods:

[1237] Copy-paste check: Collects information from the Internet using web scraping tools and APIs, compares it with the text, and calculates the match rate. Specifically, the server compares the text with multiple sources collected by the scraping tool and calculates the match rate.

[1238] AI document assessment: A BERT-based machine learning model is used to determine the likelihood of generative AI based on writing style and features. Specifically, the server uses the BERT model to score the likelihood of generative AI using the analysis results as input. The input for this step is the analysis results and information data on the Internet, and the output is the copy-paste rate and the score of the generative AI.

[1239] Step 5:

[1240] The server calculates the generation AI rate based on the analysis results and provides it to the user. If the generation AI rate exceeds a certain threshold (e.g., 70%), a detailed report is generated that includes information such as instructions for resubmission and warnings. Specifically, the server creates a report and displays the results on the user's web interface. The inputs to this step are the copy-paste rate and the generation AI score, and the output is a detailed report.

[1241] Through the above processing steps, this system can accurately evaluate the possibility that a document can be created by AI, and provide users with reliable information.

[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 robot 414 will be referred to as a "terminal."

[1244] In recent years, the use of generative AI to generate documents has become commonplace, leading to an increase in fraudulent use and acts of fraud. In particular, in electronic payments, the risk of fraudulent transactions is increasing when important documents such as contracts and invoices are forged by generative AI. Effective measures to prevent such fraudulent use and acts are needed, but existing technologies cannot adequately address this issue.

[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 users to upload their own documents, means for converting the documents into text, means for analyzing the text using natural language processing technology, means for calculating the copy-paste rate by comparing it with information on the Internet, means for determining the characteristics of the generating AI, means for calculating the generating AI rate based on the copy-paste rate and the determination result, means for providing the generating AI rate result to the user, and means for determining whether documents related to payment have been created by a generating AI in order to prevent fraudulent use and misuse during electronic payments. This makes it possible to prevent fraudulent use and misuse during electronic payments.

[1247] "User" refers to any individual or entity that uses the Service or System.

[1248] "Document" refers to any paper or file containing information, including formats such as PDF and DOCX.

[1249] "Means of converting to text" refers to tools and technologies used to analyze the contents of document files and convert them to plain text. Specifically, text extraction libraries are used.

[1250] "Natural language processing technology" refers to a set of techniques and methods that enable computers to understand and analyze human language, including NLP libraries and machine learning models.

[1251] The "copy-paste rate" is a measure of how closely the text in a document matches other sources on the Internet.

[1252] "Generative AI characteristics" refers to specific patterns, style, structure, and other characteristics found in documents created by generative AI.

[1253] The "AI generation rate" is a numerical indicator that indicates the likelihood that a document was created by a generative AI.

[1254] "Unauthorized use" refers to the intentional and unfair use of a system or service.

[1255] "Fraud" is the act of deceiving others through dishonest means in order to gain profits.

[1256] "Payment" refers to the process of making monetary transactions for goods and services.

[1257] "Means of determination" refers to the technology or algorithms used to assess whether a document was generated by AI.

[1258] MODE FOR CARRYING OUT THE INVENTION

[1259] This invention provides a system that allows users to upload documents related to electronic payments and performs a series of processes to determine whether the documents were created by a generation AI. The specific configuration and processing steps of the system are described below.

[1260] System configuration and operation

[1261] The system consists of the following components:

[1262] 1. User device: Provides an interface that runs on a smartphone, tablet, or PC and allows users to upload documents.

[1263] 2. Server: The central processing unit that receives uploaded documents, converts them to text, and performs analysis.

[1264] 3. Text conversion tools: Software that converts documents into text format (e.g., textract).

[1265] 4. Natural Language Processing (NLP) models: Libraries for analyzing text and extracting grammatical structures and features (e.g., spaCy).

[1266] 5. Generative AI Verification Model: A machine learning model to determine whether a document was created by generative AI (e.g., a BERT-based model).

[1267] 6. Copy-paste check tools: Web scraping tools and APIs that compare information on the internet to calculate copy-paste rates.

[1268] Processing flow

[1269] 1. The user uploads relevant documents (e.g., invoices, contracts) through a web interface or application provided on the user's device.

[1270] 2. The server receives the document sent by the user and converts the document into text format using a text conversion tool (e.g., textract).

[1271] 3. The server analyzes the converted text using a natural language processing (NLP) model (e.g., spaCy) to extract grammatical structures and features.

[1272] 4. The server uses a copy-paste check tool to compare the text with information on the Internet and calculate the copy-paste rate.

[1273] 5. The server uses a generative AI decision model (e.g., a BERT-based model) to determine whether the text was created by a generative AI and scores the likelihood.

[1274] 6. The server calculates the generative AI rate based on the copy-paste rate and the results of the generative AI judgment model.

[1275] 7. The server provides the user with the results of the generated AI rating and, if necessary, provides additional authentication measures and instructions for resubmission.

[1276] Hardware and software used

[1277] Server: A computer system that performs major processing, often deployed on a cloud service.

[1278] User devices: smartphones, tablets, computers, etc.

[1279] software:

[1280] Textract: A tool that converts documents into text.

[1281] SpaCy: A library for natural language processing.

[1282] BERT: A machine learning model that determines whether a document was created by generative AI.

[1283] Web scraping tools or APIs: Tools for copy-paste checking.

[1284] Specific examples

[1285] Specific examples of prompts include the following scenarios:

[1286] "Let's say a user uploads a file named 'invoice.pdf'. This file is uploaded from a smartphone and analyzed by an artificial intelligence (AI)-based fraud detection system. The server receives the file, converts and analyzes the text content, and then makes a judgment on the AI-generated document. The analysis result is sent to the payment provider, which then takes the necessary action."

[1287] As described above, by using this system, it is possible to prevent unauthorized use and fraudulent acts related to electronic payments.

[1288] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[1289] Step 1:

[1290] A user uploads a document

[1291] Users upload relevant documents (e.g., invoices or contracts) from an application or web interface on their smartphone, tablet, or PC. The input is a document file such as PDF or DOCX, and the output is the uploaded document file. Specific operations involve selecting a document using the file selection button and clicking the upload button.

[1292] Step 2:

[1293] The server receives the document

[1294] The server receives the document sent by the user. The input is the document file sent by the user, and the output is a document file stored on the server side. Specifically, the server receives an HTTP POST request and retrieves the document file contained therein.

[1295] Step 3:

[1296] The server converts the document into text

[1297] The server uses a text conversion tool (e.g., textract) to convert the document file into a text format. The input is a document file in PDF or DOCX format, and the output is plain text data. Specifically, the text extraction library parses the document file and generates plain text.

[1298] Step 4:

[1299] The server parses the text

[1300] The server uses a natural language processing (NLP) model (e.g., spaCy) to analyze the converted text. The input is plain text data, and the output is an analysis result including grammatical elements and features. Specifically, the NLP model tokenizes the text and extracts grammatical elements such as subject, predicate, and object.

[1301] Step 5:

[1302] The server compares the information on the Internet and calculates the copy-paste rate

[1303] The server uses copy-paste checking tools and APIs to match the text with online information and calculate the copy-paste rate. The input is the analyzed text and multiple online sources, and the output is a number indicating the copy-paste rate. Specifically, the server sends the text to the web scraping tool, finds matches, and calculates the match rate.

[1304] Step 6:

[1305] The server uses the generative AI judgment model to determine the characteristics of the generative AI.

[1306] The server uses a generative AI judgment model (e.g., a BERT-based model) to determine whether a document was created by a generative AI. The input is the parsed text, and the output is a score indicating the likelihood of the generative AI. Specifically, the server inputs the text into the judgment model and obtains the scored result.

[1307] Step 7:

[1308] The server calculates the AI ​​generation rate

[1309] The server calculates the AI ​​generation rate based on the copy-paste rate and the results of the AI ​​generation judgment model. The input is the copy-paste rate and a score indicating the AI ​​generation potential, and the output is a number indicating the AI ​​generation rate. Specifically, the server applies an algorithm that combines the copy-paste rate and the score to calculate the AI ​​generation rate.

[1310] Step 8:

[1311] The server provides the results to the user

[1312] The server provides users with the results of the AI ​​generation rate and detailed analysis results. The input is a numerical value indicating the AI ​​generation rate and detailed analysis data, and the output is a report format that can be viewed by users. Specifically, the server formats the analysis results into a report in HTML or PDF format and displays it on the user's terminal interface.

[1313] The above is a specific description of the processing steps in the system of the invention.

[1314] 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.

[1315] This invention relates to a system that determines whether documents such as resumes and school reports were created by generative AI (generative artificial intelligence), and by combining it with an emotion engine that recognizes the user's emotions, it provides more advanced judgment and feedback. This system improves the user experience by adding analysis using the user's emotion data in addition to the series of processes from uploading the document to analyzing it and providing the results.

[1316] System Overview

[1317] 1. User uploads a document

[1318] Users upload their resumes and reports on the system's web interface (client terminal). Upload methods include a file selection button and meta-information input fields (university name, subjects, etc.).

[1319] Specific examples

[1320] A user accesses the web page, selects "Report.docx" in the upload form, enters the university name "General University" and the subject "Information Technology," and presses the upload button to send the document and information to the server.

[1321] 2. The server receives the file and converts it to text

[1322] The server receives the file sent by the user and converts the document from PDF or DOCX format to text format, for example using a PDF to text converter or a DOCX parser.

[1323] Specific examples

[1324] The server receives the "report.docx" and converts it to text using an appropriate library (e.g. Apache POI), which stores the converted text in memory.

[1325] 3. The server parses the text

[1326] The server applies natural language processing techniques to the converted text to analyze sentence structure and features, which may involve the use of NLP libraries (e.g., spaCy, NLTK).

[1327] Specific examples

[1328] The server analyzes the converted text and extracts grammatical elements such as the subject, predicate, and object of each sentence. It also evaluates the writing style and tone and calculates features (word frequency, sentence diversity, etc.).

[1329] 4. The server performs a generated AI check.

[1330] The server performs a spawned AI check by:

[1331] Copy-paste check: Use web scraping tools and APIs to collect information from the internet and compare it with text. Find matches and calculate the copy-paste rate.

[1332] AI document judgment: Using a machine learning model, we judge the possibility of generative AI based on writing style and characteristics. We score the possibility of generative AI for each sentence and calculate the overall generative AI rate.

[1333] Specific examples

[1334] The server uses a non-destructive scraping tool to collect information from the internet, calculates the match rate with the text, and uses a BERT-based machine learning model to analyze the style of the text and evaluate the likelihood that it is generative AI.

[1335] 5. The server performs analysis using the emotion engine

[1336] When a user uploads a document, the emotion engine recognizes the user's emotions from their facial expressions and input. This emotion data is then incorporated into the analysis and used to calculate the AI ​​generation rate and provide feedback.

[1337] Specific examples

[1338] When a user uploads a document, their facial expressions are analyzed via a webcam to obtain emotional data. If the user shows signs of anxiety, the server uses this information to adjust the threshold for AI generation rates and provide accurate feedback.

[1339] 6. The server provides the results to the user

[1340] The server generates a detailed report based on the analysis results and provides it to the user. If the AI ​​generation rate exceeds a certain threshold (e.g., 70%), information including instructions for resubmission and a warning is provided. The server can also adjust the way results are presented based on emotion data.

[1341] Specific examples

[1342] The server creates a report summarizing the results of 70% AI generation rates and their details (such as high match rates and characteristics of the AI-generated writing style) and displays it on the user's web interface. If the user shows any anxiety, it provides a gentle explanation and offers additional support options.

[1343] This concludes the description of a specific embodiment of the system according to the present invention. This system uses generative AI to evaluate the validity of documents with high accuracy, and also takes into account the user's emotions to provide more personalized feedback.

[1344] The processing flow will be explained below.

[1345] Step 1:

[1346] A user accesses the upload form on the web page, selects "Report.docx", enters meta information such as the university name and subject (e.g., university name "General University", subject "Information Technology"), and presses the submit button.

[1347] Step 2:

[1348] The server saves the "report.docx" file and meta information received from the user to the file system, and then converts the saved file to a text format using an appropriate tool (e.g., Apache POI).

[1349] Step 3:

[1350] The server stores the converted text in memory and begins analyzing it using an NLP (Natural Language Processing) library (e.g., spaCy, NLTK). For each sentence, grammatical elements such as subject, predicate, and object are extracted, and features for evaluating style and tone are calculated.

[1351] Step 4:

[1352] When a user uploads a document, the emotion engine recognizes the user's emotion from the webcam, keyboard typing speed, etc. This emotion data is analyzed using specific algorithms (e.g., facial expression recognition algorithms and typing speed analysis).

[1353] Step 5:

[1354] To perform the copy-paste check, the server collects information from the Internet using web scraping tools or APIs (e.g., Google Custom Search API). The collected information is compared with the analyzed text to find matching parts, and the matching rate is calculated as the copy-paste rate.

[1355] Step 6:

[1356] The server uses a machine learning model (e.g., a BERT-based model) to perform AI document judgment based on the analyzed text. It scores the likelihood of generative AI based on features such as consistency of writing style and excessive use of specific keywords, and makes a judgment for each sentence.

[1357] Step 7:

[1358] The server combines the copy-paste rate and the AI ​​document judgment score to calculate the overall generative AI rate, where it adjusts the threshold based on the recognized user emotion data to enable personalized feedback.

[1359] Step 8:

[1360] The server generates the AI ​​generation rate results and a detailed report (e.g., high match rate, stylistic features of the AI-generated writing) and sends it to the user's web interface. If the user shows signs of nervousness or anxiety, the report includes additional support options. The user is then instructed to review the results and resubmit the document if necessary.

[1361] This is the specific processing flow of the generative AI checker system that combines an emotion engine. This system can evaluate the validity and originality of documents with high accuracy, and provide personalized feedback that takes the user's emotions into account.

[1362] Example 2

[1363] 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."

[1364] Conventional document assessment systems have difficulty accurately determining whether a document was created by a generative AI. Furthermore, they lack a mechanism for taking the user's feelings into consideration when providing the user with a result of whether a document was AI-generated. Therefore, there is a need to improve the reliability of assessment results and the user experience.

[1365] 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.

[1366] In this invention, the server includes means for users to upload their own documents, means for converting the documents into text, means for analyzing the text using natural language processing technology, means for comparing the text with information on the Internet to calculate the copy-paste rate, means for determining the possibility of a generative AI from writing style and characteristics using a machine learning model, means for acquiring user emotion data and incorporating it into the analysis, means for calculating a generative AI rate based on the copy-paste rate, the generative AI determination result, and the emotion data, and means for providing the result of the generative AI rate to the user. This increases the accuracy of determining documents created by the generative AI and makes it possible to provide feedback that takes into account the user's emotions.

[1367] "Document" refers to a formal document such as a resume or report uploaded by a User.

[1368] "Text" refers to character string data converted from document formats such as PDF and DOCX.

[1369] "Natural language processing technology" refers to the technology of analyzing text and evaluating grammatical elements and writing style, and uses NLP libraries, etc.

[1370] The "copy-paste rate" refers to the rate at which the text matches when compared with information on the Internet.

[1371] "Generative AI" refers to text or documents created using a generative AI model.

[1372] A "machine learning model" refers to a technology that uses algorithms to learn patterns from large amounts of data and determine the characteristics of the generated AI.

[1373] "Emotion data" refers to data that indicates the emotional state of the user that can be obtained from their facial expressions and input content.

[1374] "Generative AI Rate" refers to the percentage that indicates the likelihood that the text was created by a generative AI.

[1375] "Feedback" refers to information that provides analysis results and advice to users.

[1376] This invention is a system that determines whether documents such as resumes and school reports were created by generative AI, and by combining it with an emotion engine that recognizes the user's emotions, it provides more advanced judgment and feedback. This system improves the user experience by adding analysis using the user's emotion data in addition to the series of processes from uploading the document to analyzing it and providing the results.

[1377] First, the user uploads their document through the device's web interface. The uploading method includes a file selection button and meta-information input fields (university name, subject, etc.). Specifically, the user accesses the web page, selects "Report.docx" in the upload form, and enters the university name "General University" and the subject "Information Technology." Then, by pressing the upload button, the document and information are sent to the server.

[1378] The server receives the file sent by the user and converts it from file format (PDF or DOCX) to text format using the appropriate library. For example, the server receives a "report.docx" and converts it to text using the Apache POI library. This converted text is stored in memory.

[1379] The server then applies natural language processing techniques to the converted text to analyze the structure and characteristics of the sentences. NLP libraries such as spaCy and NLTK are used. For example, the server analyzes the converted text and extracts grammatical elements such as the subject, predicate, and object of each sentence. It also evaluates the writing style and tone and calculates features (word frequency, sentence diversity, etc.).

[1380] In addition, the server performs copy-paste checks and AI document assessments. Copy-paste checks involve collecting information on the Internet using web scraping tools and APIs and calculating the match rate with the text. As a specific example, the server collects information on the Internet using a non-destructive scraping tool and calculates the match rate with the text. AI document assessment uses a BERT-based machine learning model to analyze the writing style and evaluate the possibility that the text is generative.

[1381] The system also performs analysis using an emotion engine. When a user uploads a document, their facial expressions are analyzed via a webcam to capture emotional data. This emotional data is incorporated into the analysis and reflected in the calculation of the AI ​​generation rate and the feedback content. For example, if the user expresses anxiety, the server can use this information to adjust the AI ​​generation rate threshold and provide more accurate feedback.

[1382] Finally, the server generates a detailed report based on the analysis results and provides it to the user. If the AI ​​generation rate exceeds a certain threshold (e.g., 70%), information including instructions for resubmission and a warning is provided. The server can also adjust the way results are presented based on emotional data. For example, the server could create a report summarizing the results of a 70% AI generation rate and its details (high match rate, characteristics of the AI-generated writing style, etc.) and display it on the user's web interface. Users who express anxiety could be offered gentle explanations and additional support options.

[1383] The system uses generative AI to accurately assess the legitimacy of documents and also takes user sentiment into account to provide more personalized feedback.

[1384] The flow of the identification process in the second embodiment will be described with reference to FIG.

[1385] Step 1:

[1386] The user inputs the document file and meta information and presses the upload button on the device's web interface. This is the process of inputting a file called "report.docx" and meta information such as the university name "general university" and the subject "information technology," and then clicking the upload button.

[1387] Input: Document file (e.g., report.docx), meta information (university name, subject)

[1388] Output: The file and meta information are sent to the server.

[1389] Step 2:

[1390] The server receives the file sent by the user and converts it from PDF or DOCX format to a text format using an appropriate library (e.g., Apache POI). This converted text data is temporarily stored in memory.

[1391] Input: Document file

[1392] Data processing: Converting from file format to text format (using Apache POI)

[1393] Output: Text format data (temporarily saved in memory)

[1394] Step 3:

[1395] The server analyzes the converted text, using natural language processing techniques (e.g., spaCy, NLTK) to extract grammatical elements (e.g., subject, predicate, object) and calculate features of style and tone (e.g., word frequency, sentence diversity).

[1396] Input: Text format data

[1397] Data processing: Extracting grammatical elements and analyzing style and tone (using spaCy and NLTK)

[1398] Output: Analysis results (grammatical elements, style and tone features)

[1399] Step 4:

[1400] The server performs copy-paste checks, collects information from the internet using web scraping tools and APIs, compares it with text, and calculates the match rate. In addition, it analyzes writing style using a BERT-based machine learning model to evaluate the potential of generative AI.

[1401] Input: Text format data

[1402] Data Computing: Web scraping and text matching, matching rate calculation, stylistic analysis with the BERT model

[1403] Output: Match rate, possibility evaluation score of the generated AI

[1404] Step 5:

[1405] When a user uploads a document, their facial expression is captured using the device's webcam. The server uses an emotion engine to analyze the facial expression data and obtain emotional data, which is then used for analysis.

[1406] Input: Facial expression capture data

[1407] Data calculation: Analysis of facial expression data (using emotion engine)

[1408] Output: Emotion data

[1409] Step 6:

[1410] The server calculates the AI ​​generation rate by integrating and evaluating the copy-paste rate, stylistic analysis results, and emotional data. A specific algorithm is used to calculate the AI ​​generation rate.

[1411] Input: Match rate, generated AI evaluation score, emotion data

[1412] Data calculation: Algorithm-based data integration and generation AI rate calculation

[1413] Output: Generation AI rate

[1414] Step 7:

[1415] The server generates a detailed report based on the analysis results and provides it to the user. If the AI ​​generation rate exceeds a certain threshold (e.g., 70%), the report will include information such as instructions for resubmission and warnings. The server will also adjust the presentation of results based on emotion data.

[1416] Input: AI generation rate, emotion data

[1417] Output: Detailed report, feedback (resubmission instructions, warnings, etc.)

[1418] A feature of this system is that it takes user emotions into consideration while performing highly accurate document judgment using generative AI and providing personalized feedback to users.

[1419] (Application example 2)

[1420] 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."

[1421] In modern logistics centers, verifying the authenticity of package documents is important, but there is a risk of documents created by generative AI being mixed in. Furthermore, the emotions and stress of inspectors can affect work efficiency and accuracy, but there is no system that effectively monitors and provides feedback. Therefore, a system that combines package document authenticity verification and worker emotion recognition is needed.

[1422] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 2 is realized by the following means.

[1423] In this invention, the server includes means for users to upload their own documents, means for converting the documents into text, means for analyzing the text using natural language processing technology, means for calculating the copy-paste rate by comparing it with information on the Internet, means for determining the characteristics of the generating AI, means for calculating the generating AI rate based on the copy-paste rate and the determination result, means for providing the generating AI rate result to the user, means for recognizing the user's emotions, and means for adjusting the method for presenting the generating AI rate result based on the emotion data. This makes it possible to evaluate the legitimacy of package documents with high accuracy and provide feedback that takes into account the emotions of inspection workers.

[1424] "User" means a person or institution that uses the System to upload documents and obtain analysis results thereof.

[1425] "Document" means any electronic or paper information that can be converted into text format, including, for example, a resume, report, or shipping instructions.

[1426] The "means for converting to text" is a mechanism for converting an image document into text data using optical character recognition (OCR) technology.

[1427] "Natural language processing technology" is a technology for analyzing text data and extracting and analyzing grammatical structure and meaning, and specifically includes morphological analysis, grammatical analysis, and semantic analysis.

[1428] The "copy-paste rate" is the percentage of a document that has been copied in whole or in part from another existing document, and is calculated by comparing it with sources on the Internet.

[1429] "Means for determining characteristics of generative artificial intelligence" refers to techniques that use machine learning models to analyze the style and structure of documents in order to identify whether they were created by generative AI.

[1430] The "generated AI rate" is an indicator of the likelihood that a document was created by a generative AI, and is calculated based on the copy-paste rate and the AI ​​judgment results.

[1431] "Means for recognizing the user's emotions" refers to technology for extracting and analyzing emotions from the user's facial expressions and voice data, and typically involves using a camera or microphone.

[1432] The "means for adjusting the method of presenting results" is a mechanism that changes the way the analysis results are displayed based on the user's emotional data and provides optimal feedback.

[1433] A head-mounted display is a device worn on the head that displays information directly into the field of view. It also has built-in cameras and sensors that can acquire data on the external environment in real time.

[1434] The system for implementing this invention aims to control quality in logistics centers. Specifically, it determines whether shipping instructions and content descriptions included in packages were created using generative AI, and further improves work efficiency by providing feedback using the emotion data of inspection workers.

[1435] 1. System Configuration

[1436] The system includes the following main components:

[1437] 1.1 Head-mounted display (HMD)

[1438] An inspection worker wearing an HMD (e.g., a typical smart HMD device) scans the shipping instructions and contents description inside each package.

[1439] 1.2 Camera and microphone

[1440] The HMD's built-in camera scans documents and converts the images into text using OCR technology (e.g., Tesseract OCR).The camera and microphone also capture the inspector's facial expressions and voice data.

[1441] 1.3 Server

[1442] The server converts the received image data into text, analyzes grammatical structure and writing style using natural language processing (NLP) technology (e.g., spaCy), and uses a generative AI model (BERT-based model) to determine whether the document was created by generative AI and calculates the generative AI rate.

[1443] 1.4 Emotion Engine

[1444] An emotion engine running on the server analyzes the emotional data of inspection workers and adjusts the feedback content.

[1445] 2. Explanation of the processing flow

[1446] The system process flow is as follows:

[1447] 2.1 Image to Text Conversion

[1448] An inspector wearing an HMD scans the package documents with a camera and sends the image data to a server, which then uses OCR technology to convert the image data into text.

[1449] 2.2 Analysis using natural language processing

[1450] The server applies NLP techniques to the converted text to extract grammatical structures and features, including analyzing the subject, predicate, and object of sentences, and assessing sentence variety and tone.

[1451] 2.3 Generation AI judgment

[1452] The server uses the generative AI model to determine whether a document was created by generative AI, performs copy-paste checks and AI document determination functions, and calculates the generative AI rate.

[1453] 2.4 Sentiment Analysis and Feedback

[1454] The server uses an emotion engine to analyze the inspector's emotional data, and if the inspector shows signs of stress or fatigue, it adjusts the feedback method and provides appropriate alerts and suggestions (e.g., to take a break).

[1455] 3. Specific Examples

[1456] As a concrete example, an inspection worker scans a shipping instruction sheet with the file name "Doc123.jpg" using an HMD and sends it to the server. The server then uses OCR technology to convert the image into text and calculates the AI ​​generation rate. At the same time, emotional data displayed by the worker is also collected and analyzed, and feedback based on the results is displayed on the HMD display.

[1457] "Please determine if this report was created by a generative AI." is an example of a prompt sentence, which is used as input to a generative AI model.

[1458] Names of specific hardware and software used

[1459] HMD: General smart HMD devices (e.g. Microsoft HoloLens)

[1460] Camera and microphone: Camera and microphone built into the HMD

[1461] OCR technology: Tesseract OCR

[1462] NLP technology: spaCy

[1463] Generative AI model: BERT-based generative AI decision model

[1464] Emotion Engine: Emotion Recognition Library (OpenCV and Dlib)

[1465] In this way, the present invention provides an efficient and highly accurate system for evaluating package documents through quality control and worker emotion recognition in logistics centers.

[1466] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[1467] Step 1:

[1468] The user wears a head-mounted display (HMD) and scans shipping instructions and content descriptions.

[1469] Input: Image of shipping instructions or contents description

[1470] Output: Image data

[1471] Specific operation: The camera built into the HMD is used to take a photo of the document inside the package and obtain image data.

[1472] Step 2:

[1473] The server receives the image data and converts the image into text using optical character recognition (OCR) technology.

[1474] Input: Image data

[1475] Output: Text data

[1476] What happens: The server uses an OCR library (e.g., Tesseract) to analyze the image data and extract the text within the document.

[1477] Step 3:

[1478] The server analyzes the extracted text using natural language processing (NLP) techniques.

[1479] Input: Text data

[1480] Output: Grammatical structure and features

[1481] What it does: The server applies NLP libraries (e.g., spaCy) to extract grammatical elements (e.g., subject, predicate, object) in the text and evaluate sentence variety and tone.

[1482] Step 4:

[1483] The server checks for copy-paste and determines the AI ​​that generated the text data.

[1484] Input: Text data

[1485] Output: Generated AI rate

[1486] How it works: The server uses a copy-paste check tool to check against information on the internet and calculates the percentage of matches. It also uses a generative AI model (based on BERT) to analyze the style of the text and calculates the generative AI rate.

[1487] Step 5:

[1488] The server acquires the emotion data of the inspector using the camera and microphone built into the HMD.

[1489] Input: Facial expression and voice data

[1490] Output: Emotion data

[1491] How it works: The server uses the Emotion Recognition Library (OpenCV and Dlib) to analyze facial expressions and voice captured by the camera and microphone to identify the worker's emotional state.

[1492] Step 6:

[1493] The server adjusts the way results are presented based on the emotional data and provides optimal feedback to the user.

[1494] Input: Generated AI rate and emotion data

[1495] Output: Feedback results

[1496] Specific operation: The server integrates the generated AI rate and emotion data and generates a report based on the analysis results. Based on the emotion data, the format and tone of the results displayed on the HMD are adjusted to provide optimal information to the user. For example, if a worker is feeling stressed, a gentle message or a suggestion to take a break is displayed.

[1497] Through these steps, the system is able to evaluate the validity of package documents and provide feedback to improve work efficiency.

[1498] 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.

[1499] 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.

[1500] 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 robot 414.

[1501] 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.

[1502] FIG. 9 illustrates an emotion map 400 on which multiple emotions are mapped. In the emotion map 400, emotions are arranged in concentric circles radiating from the center. Emotions closer to the center of the concentric circles are more primitive. Emotions representing states and behaviors arising from a state of mind are arranged on the outer edges of the concentric circles. The concept of emotion 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.

[1503] 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.

[1504] 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).

[1505] 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.

[1506] 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."

[1507] 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.

[1508] 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).

[1509] 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.

[1510] 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.

[1511] 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.

[1512] 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.

[1513] 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.

[1514] 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.

[1515] 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.

[1516] 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.

[1517] 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.

[1518] 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.

[1519] The following is further disclosed regarding the above embodiment.

[1520] (Claim 1)

[1521] a means for users to upload their documents;

[1522] means for converting said document into text;

[1523] means for analyzing the text using natural language processing techniques;

[1524] A method to calculate the copy-paste rate by comparing it with information on the Internet,

[1525] a means for determining characteristics of the generative artificial intelligence;

[1526] means for calculating a generation artificial intelligence rate based on the copy-paste rate and the determination result;

[1527] A means for providing the results of the generated AI algorithm to a user;

[1528] A system including:

[1529] (Claim 2)

[1530] 10. The system of claim 1, further comprising means for notifying resubmission or rejection if the generated AI rate exceeds a certain threshold.

[1531] (Claim 3)

[1532] 2. The system of claim 1, wherein the copy-paste rate is calculated by comparing the text with multiple sources on the Internet.

[1533] "Example 1"

[1534] (Claim 1)

[1535] a means for users to upload their documents;

[1536] means for converting said document into text;

[1537] means for analyzing the text using natural language processing techniques;

[1538] A means for collecting information on the Internet and comparing it with the text to calculate a match rate;

[1539] a means for determining characteristics of the generative artificial intelligence using a machine learning model;

[1540] means for calculating a generation artificial intelligence rate based on the match rate and the determination result;

[1541] A means for providing the results of the generated AI algorithm to a user;

[1542] A system including:

[1543] (Claim 2)

[1544] 10. The system of claim 1, further comprising means for resubmitting or issuing a warning if the generated AI rate exceeds a certain threshold.

[1545] (Claim 3)

[1546] 2. The system of claim 1, wherein the calculation of the match rate is performed by comparing the text with multiple sources on the Internet.

[1547] "Application Example 1"

[1548] (Claim 1)

[1549] a means for users to upload their documents;

[1550] means for converting said document into text;

[1551] means for analyzing the text using natural language processing techniques;

[1552] A method to calculate the copy-paste rate by comparing it with information on the Internet,

[1553] a means for determining characteristics of the generative artificial intelligence;

[1554] means for calculating a generation artificial intelligence rate based on the copy-paste rate and the determination result;

[1555] A means for providing the results of the generated AI algorithm to a user;

[1556] A means for determining whether a document related to a payment has been created by a generation AI in order to prevent fraudulent use or misuse during electronic payments;

[1557] A system including:

[1558] (Claim 2)

[1559] 10. The system of claim 1, further comprising means for notifying resubmission or rejection if the generated AI rate exceeds a certain threshold.

[1560] (Claim 3)

[1561] 2. The system of claim 1, wherein the copy-paste rate is calculated by comparing the text with multiple sources on the Internet.

[1562] "Example 2: Combining Emotion Engines"

[1563] (Claim 1)

[1564] a means for users to upload their documents;

[1565] means for converting said document into text;

[1566] means for analyzing the text using natural language processing techniques;

[1567] A method to calculate the copy-paste rate by comparing it with information on the Internet,

[1568] A means of determining the potential of generative AI based on writing style and characteristics using machine learning models, and

[1569] A means of acquiring and analyzing user emotion data;

[1570] A means for calculating a generation AI rate based on the copy-paste rate, the generation AI determination result, and emotion data;

[1571] means for providing the results of the generated AI rate to a user;

[1572] A system including:

[1573] (Claim 2)

[1574] 10. The system of claim 1, further comprising means for providing a resubmission or warning notice when the generated AI rate exceeds a certain threshold.

[1575] (Claim 3)

[1576] 2. The system of claim 1, wherein the copy-paste rate is calculated by comparing the text with sources on the Internet.

[1577] "Application example 2 when combining emotion engines"

[1578] (Claim 1)

[1579] a means for users to upload their documents;

[1580] means for converting said document into text;

[1581] means for analyzing the text using natural language processing techniques;

[1582] A method to calculate the copy-paste rate by comparing it with information on the Internet,

[1583] a means for determining characteristics of the generative artificial intelligence;

[1584] means for calculating a generation artificial intelligence rate based on the copy-paste rate and the determination result;

[1585] A means for providing the results of the generated AI algorithm to a user;

[1586] a means for recognizing a user's emotion;

[1587] a means for adjusting a method for presenting the results of the artificial intelligence generation based on the emotion data;

[1588] A system including:

[1589] (Claim 2)

[1590] 10. The system of claim 1, further comprising means for notifying resubmission or rejection if the generated AI rate exceeds a certain threshold.

[1591] (Claim 3)

[1592] 2. The system of claim 1, wherein the copy-paste rate is calculated by comparing the text with multiple sources on the Internet.

[1593] (Claim 4)

[1594] 10. The system of claim 1, further comprising means for using video and audio data for emotion recognition.

[1595] (Claim 5)

[1596] 10. The system of claim 1, wherein the user is a worker at a logistics center and includes means for scanning and uploading documents using a head-mounted display. [Explanation of symbols]

[1597] 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 for users to upload their documents; means for converting said document into text; means for analyzing the text using natural language processing techniques; A method to calculate the copy-paste rate by comparing it with information on the Internet, a means for determining characteristics of the generative artificial intelligence; means for calculating a generation artificial intelligence rate based on the copy-paste rate and the determination result; A means for providing the results of the generated AI algorithm to a user; A system including:

2. The system of claim 1 further comprising means for notifying resubmission or rejection if the generated AI rate exceeds a certain threshold.

3. 2. The system of claim 1, wherein the copy-paste rate is calculated by comparing the text with multiple sources on the Internet.

Images