Artificial intelligence personification data privacy protection platform based on psychological linguistics

Abstract

The invention relates to the technical field of data privacy protection, and particularly provides a psychological linguistics-based artificial intelligence anthropomorphic data privacy protection platform, which comprises a data processing module used for acquiring original text data generated when students use an online education platform, then desensitizing the original text data, and storing the desensitized original text data in a database; desensitizing the text data to obtain desensitized text data; the feature acquisition module is used for analyzing the desensitized text data to obtain psychological linguistic features for representing student data privacy psychological states; the strategy generation module is used for obtaining a data privacy psychological state evaluation result of the student according to psychological linguistics feature analysis, and then generating a privacy setting guide strategy according to the data privacy psychological state evaluation result; the interaction module is used for carrying out privacy interaction with the students according to the privacy setting guide strategy; the system can effectively solve the problem that students are forced to accept default options of the platform during privacy setting due to the fact that the students cannot fully understand the text protocol.

Description

Technical Field

[0001] The present application relates to the technical field of data privacy protection, and more specifically, to an artificial intelligence anthropomorphic data privacy protection platform based on psycholinguistics. Background Art

[0002] Online education platforms are becoming increasingly popular, with students engaging in a variety of learning activities, including taking courses, completing exercises, and participating in discussions. To improve teaching effectiveness and provide a personalized learning experience, platforms often need to collect and analyze student learning behavior data. By analyzing students' viewing time and practice patterns, they can assess their mastery of key knowledge points. By analyzing discussion forums and notes, they can understand students' confusions and thought processes. And by analyzing conversations with assistants, they can understand students' personalized learning needs and challenges.

[0003] However, as data collection deepens, students are increasingly concerned about how their personal data is collected, used, and shared. Different students have vastly different levels of awareness and sensitivity regarding data privacy. Existing online education platforms typically employ traditional privacy policy disclosure methods, providing students with lengthy, highly technical text agreements. Students generally struggle to fully understand the specific terms of these agreements, the data processing process, and the potential impact of these processes on them. Consequently, existing technologies pose a problem: students are forced to accept the platform's default options when setting their privacy settings due to their inability to fully understand the text agreements. This compromises students' right to informed consent and prevents the data collected by the platforms from fully reflecting their true wishes.

[0004] There is no effective technical solution to the above problems. It should be noted that the above information disclosed in this section is only used to understand the background of the present invention, and therefore may contain information that does not constitute prior art. Summary of the Invention

[0005] The purpose of this application is to provide an artificial intelligence anthropomorphic data privacy protection platform based on psycholinguistics, which can effectively solve the problem that students are forced to accept the platform's default options when setting privacy because they cannot fully understand the text agreement.

[0006] This application provides an artificial intelligence anthropomorphic data privacy protection platform based on psycholinguistics, which includes: The data processing module is used to obtain the original text data generated by students when using the online education platform, and then desensitize the original text data to obtain desensitized text data; The feature acquisition module is used to analyze the desensitized text data to obtain psycholinguistic features used to characterize the psychological state of students' data privacy; A policy generation module is used to obtain the student's data privacy psychological state assessment results based on psycholinguistic feature analysis, and then generate a privacy setting guidance policy based on the data privacy psychological state assessment results; The interaction module is used to conduct private interactions with students based on the privacy setting guidance strategy.

[0007] From the above, it can be seen that the present application provides an artificial intelligence anthropomorphic data privacy protection platform based on psycholinguistics, which can first analyze the original text data generated by students in the process of using the platform to extract psycholinguistic features, and then obtain the students' data privacy psychological state assessment results based on the psycholinguistic feature analysis, and finally generate a privacy setting guidance strategy based on the data privacy psychological state assessment results to achieve personalized and anthropomorphic privacy setting guidance strategy based on the students' individual data privacy psychological state. That is, the present application can enable the privacy policy to be fully understood by students. Therefore, the present application can effectively solve the problem that students are forced to accept the platform's default options when making privacy settings because they cannot fully understand the text agreement, thereby effectively avoiding the situation where students' right to informed consent is damaged and the data collected by the platform fails to fully reflect the students' true intentions due to being forced to accept the platform's default options. That is, the present application can effectively improve the accuracy and reliability of students' data privacy settings. BRIEF DESCRIPTION OF THE DRAWINGS

[0008] Figure 1 A schematic diagram of the structure of an artificial intelligence anthropomorphic data privacy protection platform based on psycholinguistics provided in an embodiment of the present application.

[0009] Figure 2 A flowchart of an artificial intelligence anthropomorphic data privacy protection platform based on psycholinguistics is provided in an embodiment of the present application.

[0010] Figure 3 This is a flowchart of generating a privacy setting guidance strategy based on the data privacy psychological state assessment results provided in an embodiment of the present application.

[0011] Figure numerals: 1. Data processing module; 2. Feature acquisition module; 3. Strategy generation module; 4. Interaction module. DETAILED DESCRIPTION

[0012] The technical solutions in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all of the embodiments. The components of the embodiments of the present application generally described and shown in the drawings here can be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present application provided in the drawings is not intended to limit the scope of the application for protection, but merely represents the selected embodiments of the present application. Based on the embodiments of the present application, all other embodiments obtained by those skilled in the art without making creative work fall within the scope of protection of the present application.

[0013] It should be noted that similar reference numerals and letters represent similar items in the following drawings. Therefore, once an item is defined in one drawing, it does not need to be further defined or explained in subsequent drawings. At the same time, in the description of this application, the terms "first", "second", etc. are only used to distinguish the description and should not be understood as indicating or implying relative importance.

[0014] Traditional online education platforms require the collection and analysis of user learning behavior data to improve teaching effectiveness and provide learning experiences tailored to user characteristics. Due to varying levels of user awareness and sensitivity regarding data privacy, existing online education platforms utilize text-based privacy policies. This makes it difficult for users to understand the specific terms of these agreements, the data processing process, and the impact of these processes on them. Consequently, users may be forced to accept the platform's default settings when configuring their privacy settings, which in turn impacts their right to informed consent and may result in the data collected by the platform failing to reflect their actual wishes.

[0015] For example, suppose a user uses an online education platform for the first time. The platform prompts the user to configure privacy settings and displays a text agreement containing terms for data collection, processing, and sharing, using both legal and technical terms. The settings interface provides a series of options, such as whether to allow the platform to analyze note content and whether to allow learning progress to be shared with third parties. When users fail to understand the text agreement, they make choices based on intuition or the platform's default settings. For example, users may not understand how analyzing note content will affect their privacy or are unclear about the specific scope and purpose of data sharing. In this case, they allow the platform's default settings of allowing note content analysis and data sharing, resulting in privacy settings choices made by users that may not be consistent with their actual expectations for data privacy.

[0016] If these issues are not addressed, platforms may collect and use data without users' full understanding and consent, potentially posing compliance risks. The data collected by platforms based on these settings may not accurately reflect users' actual privacy preferences, impacting preference-based data processing and user-specific services. This could lead to reduced user activity on the platform, hindering data accumulation and functionality optimization.

[0017] In this regard, Figure 1-Figure 3 As shown, this application provides an artificial intelligence anthropomorphic data privacy protection platform based on psycholinguistics, which includes: Data processing module 1 is used to obtain the original text data generated by students when using the online education platform, and then desensitize the original text data to obtain desensitized text data; Feature acquisition module 2 is used to analyze the desensitized text data to obtain psycholinguistic features for characterizing the psychological state of students' data privacy; Strategy generation module 3 is used to obtain the student's data privacy psychological state assessment results based on psycholinguistic feature analysis, and then generate a privacy setting guidance strategy based on the data privacy psychological state assessment results; The interactive module 4 is used to conduct privacy interactions with students according to the privacy setting guidance strategy.

[0018] The data processing module 1 is used to obtain the original text data generated by students when using the online education platform, and then desensitize the original text data to obtain desensitized text data. The original text data may include text content generated by students in scenarios such as forums, discussion areas, notes, private messages, and conversations with platform assistants. Desensitization refers to processing the original text data to remove or replace sensitive information. This embodiment can use replacement processing to achieve desensitization of the original text data, such as replacing the student's real name, student number and other personal identity information with an anonymous identifier. This embodiment can also use deletion processing to achieve desensitization of the original text data, such as deleting sentences or paragraphs containing sensitive topics. It should be understood that the original text data is a carrier of students' natural expression, and the original text data can reflect their true thoughts and emotions. This embodiment can desensitize the original text data so that subsequent analysis will not directly expose sensitive information, so as to protect the student's personal identity and sensitive information while obtaining the data source required for analysis.

[0019] Feature acquisition module 2 is used to analyze the desensitized text data to obtain psycholinguistic features that characterize students' data privacy psychology. Psycholinguistic features refer to linguistic attributes in text that can reflect an individual's psychological state. This embodiment can obtain psycholinguistic features that characterize students' data privacy psychology by analyzing the desensitized text data using natural language processing techniques. For example, this can involve analyzing the frequency of use of specific vocabulary (such as security-related terms) in the text. In this embodiment, the student data privacy psychology refers to the cognition, emotions, and attitudes expressed by students regarding the collection, use, storage, and sharing of personal data while using the online education platform. This student data privacy psychology can reflect the student's understanding and sensitivity regarding data privacy. It should be understood that because this embodiment indirectly and effectively obtains psycholinguistic features that characterize students' data privacy psychology by analyzing vocabulary, syntax, and sentiment in the desensitized text data, it can effectively avoid the subjective bias or comprehension barriers that can arise from direct questioning. In other words, compared to direct questioning, the psycholinguistic features obtained in this embodiment can more objectively reflect students' understanding and sensitivity regarding data privacy.

[0020] Among them, the policy generation module 3 is used to obtain the student's data privacy psychological state assessment results based on psycholinguistic feature analysis, and then generate a privacy setting guidance policy based on the data privacy psychological state assessment results. The data privacy psychological state assessment results are a quantitative or qualitative assessment of the student's cognition and sensitivity with respect to data privacy. This embodiment can use a pre-set scoring mechanism to obtain the student's data privacy psychological state assessment results based on psycholinguistic feature analysis. For example, by analyzing the frequency range of occurrence of each security-related word, the score of each security-related word is determined, and then all scores are summed to obtain the score of each security-related word. This embodiment can also obtain the student's data privacy psychological state assessment results based on psycholinguistic feature analysis by inputting psycholinguistic features into a pre-trained classification model. For example, the classification model is used to classify students into categories such as high sensitivity, medium sensitivity, and low sensitivity based on psycholinguistic features. The privacy setting guidance strategy refers to the content, method, and process used to guide privacy interactions with students. This embodiment can use a pre-set rule engine to generate a privacy setting guidance strategy based on the data privacy psychological state assessment results. For example, the pre-set rule engine is: if the student's sensitivity is high and the understanding score is low, then a guidance strategy with detailed explanations and clear steps is generated. This embodiment can also be implemented by inputting the data privacy psychological state assessment results into a pre-trained method. Specifically, a machine learning model can be used to generate a privacy setting guidance strategy based on the data privacy psychological state assessment results. The machine learning model can directly output the optimal guidance strategy template based on the data privacy psychological state assessment results. This embodiment can also generate a privacy setting guidance strategy based on the data privacy psychological state assessment results by querying a pre-built mapping relationship table between psychological state assessment results and guidance strategies based on the data privacy psychological state assessment results. It should be understood that since the policy generation module 3 of this embodiment first obtains the student's data privacy psychological state assessment results based on psycholinguistic feature analysis, and then generates a privacy setting guidance policy based on the data privacy psychological state assessment results, this embodiment is equivalent to generating a set of personalized privacy setting guidance policies based on the data privacy psychological state assessment results. This means that the guidance content, methods and processes are no longer stereotyped, but are tailored to the student's specific psychological state, thereby improving the student's understanding and acceptance of privacy settings.

[0021] The interactive module 4 is used to conduct privacy interactions with students based on the privacy setting guidance policy. Privacy interaction refers to the communication process between the platform and students regarding data privacy settings. This embodiment can use a chat robot to implement privacy interactions with students based on the privacy setting guidance policy, such as communicating with students through text or voice through a dialog box. This embodiment can also use personalized pop-ups to implement privacy interactions with students based on the privacy setting guidance policy, such as popping up customized privacy setting prompts when students perform specific operations. This embodiment can also use a guided interface to implement privacy interactions with students based on the privacy setting guidance policy, such as adjusting the layout, prompts and operation procedures of the privacy setting page according to the policy. The interactive module 4 of this embodiment can execute a privacy interaction process with students based on the received privacy setting guidance policy to guide students to understand and make privacy settings. It should be understood that since the interactive module 4 of this embodiment is used to conduct privacy interactions with students based on the privacy setting guidance strategy, this interactive method is more friendly and easy to understand. Therefore, this embodiment can effectively explain to students the purpose, scope and risks of data processing, so as to guide students to set privacy according to their true wishes and help students overcome understanding barriers, thereby truly achieving informed consent and ensuring that their data privacy settings can accurately reflect their psychological state and needs, thereby solving the problem of insufficient or inaccurate student privacy settings under traditional methods.

[0022] Specifically, this platform aims to solve the problem that students on online education platforms are forced to accept the platform's default options when setting their privacy because they find it difficult to fully understand the privacy policy. The platform obtains the original text data generated by students when using the platform through data processing module 1. This data is a carrier of students' natural expression and can reflect their true thoughts and emotions. Then, these original text data are desensitized, which is the first step to protect students' privacy and ensure that subsequent analysis does not directly expose sensitive information. Feature acquisition module 2 receives the desensitized text data and uses psycholinguistic methods to conduct in-depth analysis. By analyzing psycholinguistic features such as vocabulary, syntax, and emotion in the text, it can indirectly and effectively obtain information used to characterize students' data privacy psychological state. This method avoids the subjective bias or comprehension barriers that may be caused by direct questioning, and can more objectively reflect students' cognition and sensitivity in terms of data privacy. Policy Generation Module 3 is the core intelligent component of the platform. It assesses students' privacy psychology based on the psycholinguistic features obtained by Feature Acquisition Module 2 to obtain a data privacy psychology assessment result. Based on this data privacy psychology assessment result, Policy Generation Module 3 generates a personalized privacy setting guidance policy. This means that the guidance content, method, and process are no longer one-size-fits-all, but are tailored to the student's specific psychology, thereby improving students' understanding and acceptance of privacy settings. Finally, Interaction Module 4 conducts personalized privacy interactions with students based on the personalized privacy setting guidance policy output by Policy Generation Module 3. This interactive approach is more user-friendly and easy to understand, effectively explaining the purpose, scope, and risks of data processing to students and guiding them to make privacy settings based on their true wishes. Through this personalized and intelligent interaction, the platform helps students overcome comprehension barriers, truly achieve informed consent, and ensure that their data privacy settings accurately reflect their psychology and needs, thereby resolving the problem of insufficient or inaccurate student privacy settings under traditional methods.

[0023] As a preferred embodiment, the solution of the present application is specifically implemented as follows: Suppose a student posts a message in a course discussion area of ​​an online education platform, expressing concern about the course data being used for "analysis" and using words such as "worry", "don't know", and "will it be". The data processing module 1 obtains the original text data, desensitizes the text, replaces the student's username with an anonymous ID, and replaces the course name with the course category ID to obtain desensitized text data. The feature acquisition module 2 receives the desensitized text data and uses a psycholinguistic analysis tool (such as the LIWC dictionary) to analyze the text to identify high-frequency negative emotion words (such as "worry"), cognitive process words (such as "don't know"), and interrogative sentences (such as "will it be"), and then obtains psycholinguistic features by calculating the frequency of these words and sentences. The strategy generation module 3 evaluates the student's privacy psychology state based on the received psycholinguistic features. For example, the evaluation result is: the user's understanding of privacy settings is insufficient and the user is less sensitive to privacy settings. Then, based on this evaluation result, a privacy setting guidance strategy is generated. For example, the privacy setting guidance strategy is: using simple and easy-to-understand language to explain the purpose of data analysis (for example, to improve teaching, provide personalized recommendations), explain which data will be used and how to protect privacy, provide setting options for data usage preferences, and design a step-by-step guidance process. Interaction module 4 initiates privacy interaction through the chat window within the platform based on the generated privacy setting guidance strategy. Taking the data privacy psychological state assessment results as showing that students have a high level of concern about data sharing and a low score for understanding the privacy policy, the privacy setting guidance strategy instructs the interaction module to adopt a patient, detailed and clear-step dialogue method, and focuses on emphasizing the security of data use and students' control over data sharing. For example, the chatbot starts the conversation with the user in a friendly and direct manner: "Hello! We want to talk to you about your learning data privacy settings to help you better understand and manage your personal information." Then, the interaction module will explain the purpose of the platform collecting and using student data in simple and easy-to-understand language: "In order to recommend more suitable course content and learning resources for you, the platform will analyze your learning records. These analyses will only be used to improve your learning outcomes and will not be used for other purposes." At the same time, the data protection measures taken by the platform will be explained to alleviate students' concerns. The interactive module then guides students step by step to review and set specific privacy options. For example, it may direct students to the "Data Sharing Settings" page and explain the meaning of each option: "Here, you can choose whether to allow us to use your learning progress data for personalized recommendations. If you choose not to allow, we will not use this data for recommendations." During the explanation process, the interactive module will use non-technical language and may provide specific examples to illustrate the impact of different settings.

[0024] The present application provides an artificial intelligence anthropomorphic data privacy protection platform based on psycholinguistics, which can first analyze the original text data generated by students in the process of using the platform to extract psycholinguistic features, then obtain the students' data privacy psychological state assessment results based on the psycholinguistic feature analysis, and finally generate a privacy setting guidance strategy based on the data privacy psychological state assessment results to achieve personalized and anthropomorphic privacy setting guidance strategy based on the students' individual data privacy psychological state. That is, the present application can enable the privacy policy to be fully understood by students. Therefore, the present application can effectively solve the problem that students are forced to accept the platform's default options when making privacy settings because they cannot fully understand the text agreement, thereby effectively avoiding the situation where students' right to informed consent is damaged and the data collected by the platform fails to fully reflect the students' true intentions due to being forced to accept the platform's default options. That is, the present application can effectively improve the accuracy and effectiveness of students' data privacy settings.

[0025] In some preferred embodiments, the process of obtaining a student's data privacy psychological state assessment result based on psycholinguistic feature analysis and then generating a privacy setting guidance strategy based on the data privacy psychological state assessment result includes: S11. Obtain students' data privacy psychological status assessment results based on psycholinguistic feature analysis. The data privacy psychological status assessment results include privacy understanding scores, privacy sensitivity index, and degree of concern; S12. Determine the privacy interaction content and privacy interaction process based on the privacy understanding score and privacy sensitivity index; S13. Determine privacy interaction style based on privacy sensitivity index and degree of concern; S14. Integrate privacy interaction content, privacy interaction process, and privacy interaction style to form a privacy setting guidance strategy.

[0026] The privacy understanding score is an indicator that assesses students' understanding of data privacy-related concepts, policies, or settings. It can be implemented as a score or a grade. The privacy sensitivity index is an indicator that assesses students' sensitivity to the collection, use, or sharing of personal data. It can be implemented as a numerical value or a grade. The level of concern is an indicator that assesses students' level of concern about data privacy risks or potential consequences. It can be implemented as a numerical value or a grade. Privacy interaction content refers to the information carriers (such as text, images, audio, or video) used to guide students in setting privacy settings. It can be implemented as explanatory text, question-and-answer dialogues, or graphic illustrations. This embodiment can be implemented by querying a pre-built mapping table of understanding scores, sensitivity indices, interaction content, and interaction processes based on the privacy understanding score and privacy sensitivity index. The privacy interaction process refers to the steps, sequence, or interaction path followed when guiding students in setting privacy settings. It can be implemented as a linear step-by-step guide, a branching decision tree, or an interactive dialogue process. The privacy interaction style refers to the tone, wording, or presentation used when guiding students in setting privacy settings. It can be formal, friendly, soothing, or encouraging. Specifically, this embodiment can determine the content and process of private interaction based on the privacy understanding score and privacy sensitivity index by querying a pre-built mapping relationship table of understanding score, sensitivity index, interaction content, and interaction process based on the privacy understanding score and privacy sensitivity index. This embodiment can also determine the style of private interaction based on the privacy sensitivity index and level of concern by querying a pre-built mapping relationship table of sensitivity index, level of concern, and interaction style based on the privacy sensitivity index and level of concern. A privacy setting guidance strategy refers to a specific plan for guiding private interactions with students, generated based on the results of a student's psychological state assessment. It can be implemented using structured text instructions, dialogue scripts, or interaction flowcharts.

[0027] Specifically, this solution improves the relevance and effectiveness of privacy setting guidance strategies by refining the process of generating them based on students' psycholinguistic characteristics. First, step S11 obtains specific data privacy psychological state assessment results based on psycholinguistic characteristics analysis, quantifying the abstract psychological state into three specific dimensions: privacy understanding score, privacy sensitivity index, and level of concern. This provides a clear input basis for subsequent strategy generation. Next, step S12 determines the content and process of privacy interaction based on the student's privacy understanding score and privacy sensitivity index. This means that the guidance content is adjusted according to the student's comprehension ability, and the steps and complexity of the interaction are also adapted to the student's sensitivity, ensuring that students can understand and effectively participate in the interaction. For example, for students with low comprehension scores, the interaction content can be illustrated with text and pictures, with a step-by-step process design and simple confirmation at each step. For students with high sensitivity indexes, the interaction content can include more detailed data usage instructions and risk warnings, and more options can be added to the process for students to choose from. Subsequently, step S13 determines the style of privacy interaction based on the student's privacy sensitivity index and level of concern. This allows the tone and manner of interaction to align with the student's emotional state, helping to build trust and reduce resistance. For example, for students with high levels of concern, the interaction style can adopt a gentle, explanatory tone, emphasizing the platform's data protection measures; for students with low levels of concern, the style can be more direct and efficient. Finally, step S14 integrates the privacy interaction content, process, and style determined based on the above assessment results to form a unified and personalized privacy setting guidance strategy and ensure that the entire interaction process matches the student's specific psychological state in content, structure, and emotion, thereby more effectively guiding students to make informed and desired privacy settings. This process assesses students' privacy psychological state based on psycholinguistic features extracted from student text data and generates a customized guidance strategy based on this. Unlike traditional methods that rely solely on general rules or preset processes, this process can more deeply reflect students' individual differences and provide more targeted privacy setting guidance. Through the above scheme, this application can provide more targeted, easier to understand and accept privacy setting guidance based on the specific psychological state of students, effectively solving the problem that students cannot effectively set privacy due to difficulty in understanding in the existing technology, and improving students' awareness and autonomy in setting privacy.

[0028] In some preferred embodiments, step S14 includes: S141. Integrate the privacy interaction content, privacy interaction process, and privacy interaction style to form a preliminary setting guidance strategy; S142, obtain students’ cognitive preferences and learning styles; S143. Determine the guidance adjustment strategy corresponding to the initial guidance strategy based on cognitive preferences and learning styles; S144: Adjust the preliminary setting guidance policy according to the guidance adjustment policy to obtain a privacy setting guidance policy.

[0029] Determining the guidance adjustment strategy corresponding to the preliminary setting guidance strategy based on cognitive preferences and learning styles refers to determining how to modify the presentation method, language style, information organization structure, etc. of the preliminary setting guidance strategy based on the individual student's habitual way of processing information, understanding concepts, and accepting guidance. This embodiment can obtain students' cognitive preferences and learning styles by analyzing students' historical learning behavior data, conducting questionnaires on students, or using a pre-trained cognitive preference and learning style assessment model to assess cognitive preferences and learning styles based on historical learning behavior data. The cognitive preferences and learning styles reflect the individual student's habitual way of processing information and accepting guidance. This embodiment can determine the guidance adjustment strategy corresponding to the preliminary setting guidance strategy based on cognitive preferences and learning styles by querying a pre-built mapping relationship table of cognitive preferences, learning styles, setting guidance strategies, and adjustment strategies based on cognitive preferences, learning styles, and preliminary setting guidance strategies. For example, diagrams or flowcharts can be added to explain privacy setting options, complex setting steps can be broken down into clearly numbered sub-steps, and information can be presented in lists rather than paragraphs. Since this embodiment first determines the guidance adjustment strategy corresponding to the preliminary setting guidance strategy based on cognitive preferences and learning styles, and then adjusts the preliminary setting guidance strategy based on the guidance adjustment strategy, that is, this embodiment is equivalent to deciding how to adjust the presentation method, language style and information organization structure of the preliminary strategy based on the students' cognitive and learning characteristics, this embodiment can make the privacy setting guidance strategy more in line with the students' cognitive preferences and learning styles in terms of content, process or style. This application solves the problem that the formed privacy setting guidance strategy fails to further consider the differences in individual students' cognitive preferences and learning styles, thereby effectively improving the presentation effect of the generated guidance strategy when actually interacting with students, enhancing students' understanding and acceptance of privacy setting guidance, and improving the accuracy and effectiveness of students' data privacy settings.

[0030] In some preferred embodiments, step S143 includes: A1. Query a pre-built mapping table of cognitive preferences, learning styles, and adjustment strategies based on cognitive preferences and learning styles to obtain a preliminary adjustment strategy. A2. Obtain the student's current emotional state and then determine a correction strategy for the initial adjustment strategy based on the current emotional state; A3. Adjust and calibrate the preliminary adjustment strategy according to the correction strategy to obtain the guided adjustment strategy.

[0031] A pre-built mapping table of cognitive preferences, learning styles, and adjustment strategies refers to structured data that stores the correspondence between different cognitive preference types, learning style types, and recommended preliminary adjustment strategies. A preliminary adjustment strategy refers to a basic adjustment plan generated for adjusting the preliminary guidance strategy based on the student's cognitive preferences and learning style. It can be implemented using a set of parameters, a set of rules, or a policy template. A student's current emotional state refers to the student's real-time emotional expression or psychological feelings during private interactions. It can be implemented using an emotion classification label (e.g., positive, negative, neutral) or an emotional intensity value. This embodiment can analyze the student's emotional state using data collected by a wearable device. A correction strategy for the preliminary adjustment strategy refers to a solution for modifying or optimizing the preliminary adjustment strategy based on the student's current emotional state. This correction strategy can be an adjustment rule, a weighting factor, or a policy modification instruction. This embodiment can determine the correction strategy for the preliminary adjustment strategy based on the current emotional state by querying a pre-built mapping table of emotional state, adjustment strategy, and correction strategy based on the current emotional state and the preliminary adjustment strategy. Guidance adjustment strategy refers to the specific plan used to adjust the initially set guidance strategy after comprehensively considering students' cognitive preferences, learning styles and current emotional states.

[0032] Specifically, this solution aims to address the issue of failing to fully consider students' real-time emotional states when determining guidance and adjustment strategies based on their cognitive preferences and learning styles, thereby improving the accuracy and effectiveness of guidance and adjustment strategies. Step A1 quickly and specifically generates a basic adjustment direction by querying a pre-built mapping table based on the student's cognitive preferences and learning styles, ensuring that the guidance strategy is generally consistent with the student's cognitive habits and learning style. Step A2 involves obtaining the student's current emotional state and determining a correction strategy based on this emotional state. This embodiment dynamically adjusts the initial adjustment strategy by obtaining the current emotional state and determining the corresponding correction strategy to make it more adaptable to the student's current emotional state. For example, when a student is depressed or irritable, a gentler or more concise guidance approach may be adopted, thereby reducing interaction resistance and improving interaction effectiveness. Step A3 combines the initial adjustment strategy obtained in step A1 with the correction strategy obtained in step A2 to ultimately obtain the guidance and adjustment strategy. This step embodies the core concept of this solution: combining initial adjustments based on individual stable characteristics with dynamic corrections based on real-time emotional states. Through this combination, the generated guidance adjustment strategy takes into account both the student's long-term preferences and their current specific emotions, making the final privacy setting guidance strategy more personalized, flexible, and adaptable in real time, thereby better guiding students in privacy settings. Through the above solution, this application solves the problem that only considering students' cognitive preferences and learning styles is not enough to cope with students' real-time status changes, so that the determined guidance adjustment strategy can fully adapt to the student's specific situation at the time, improving the effectiveness of the final privacy setting guidance strategy.

[0033] In some preferred embodiments, the desensitized text data includes desensitized data of students in different scenarios, and the desensitized text data is analyzed to obtain psycholinguistic features for characterizing the psychological state of student data privacy: S21. Analyze the corresponding desensitized data for each scenario to obtain the psycholinguistic features corresponding to the scenario for characterizing the student's data privacy psychological state; Step S11 includes: S111. For each psycholinguistic feature, obtain the student's preliminary understanding score, preliminary sensitivity index, and preliminary worry level based on the psycholinguistic feature analysis; S112. Determine, based on all scenarios, a first weighted weight corresponding to each preliminary understanding score, a second weighted weight corresponding to each preliminary sensitivity index, and a third weighted weight corresponding to each preliminary worry level; S113. Calculate the privacy understanding score based on all preliminary understanding scores and their corresponding first weighted weights, calculate the privacy sensitivity index based on all preliminary sensitivity indexes and their corresponding second weighted weights, and calculate the degree of concern based on all preliminary degree of concern and their corresponding third weighted weights.

[0034] The preliminary understanding score, preliminary sensitivity index, and preliminary concern level refer to the preliminary quantitative results of a psycholinguistic feature obtained from the analysis of desensitized data in a specific scenario, obtained after evaluating the feature's impact on a student's privacy understanding, privacy sensitivity, and concern level in the current scenario. The first weighted weight, second weighted weight, and third weighted weight refer to numerical values ​​used to measure the relative importance of different scenarios or different psycholinguistic features in assessing a student's overall privacy understanding score, privacy sensitivity index, and concern level. This embodiment can determine the first weighted weight corresponding to each preliminary understanding score, the second weighted weight corresponding to each preliminary sensitivity index, and the third weighted weight corresponding to each preliminary concern level based on all scenarios by querying a pre-constructed mapping relationship table of scenario combinations, understanding score weight combinations, sensitivity index weight combinations, and concern level weight combinations based on all scenarios. The privacy understanding score, privacy sensitivity index, and concern level refer to the final quantitative assessment of a student's overall data privacy psychological state after comprehensively considering their performance in different scenarios.

[0035] Specifically, this solution addresses the issue of student data coming from different scenarios by refining the process of acquiring psycholinguistic features and calculating the results of data privacy psychological state assessments to improve the accuracy of data privacy psychological state assessments. First, when obtaining psycholinguistic features used to characterize students' data privacy psychological states, instead of performing a holistic analysis of all desensitized text data, the solution analyzes the desensitized data corresponding to each specific scenario in step S21, thereby obtaining psycholinguistic features that reflect the student's privacy psychological state in that scenario. This scenario-based analysis method can more finely capture the characteristics of students' language expression in different situations and lay the foundation for subsequent more accurate psychological state assessments. These characteristics may be related to privacy concerns in specific scenarios. Secondly, when analyzing psycholinguistic features to obtain student data privacy psychological state assessment results (including privacy understanding scores, privacy sensitivity indexes, and levels of concern), this solution is refined through steps S111 to S113. Step S111 analyzes each psycholinguistic feature obtained from different scenarios to obtain the student's preliminary understanding score, preliminary sensitivity index, and preliminary level of concern. This means that each feature in each scenario is preliminarily evaluated to obtain a more fine-grained assessment result. Step S112 determines the weighting of these preliminary assessment results based on all scenarios. Because different scenarios may have different impacts on students' overall privacy psychological state, for example, scenarios involving personal sensitive information may be more important than ordinary learning and discussion scenarios, this embodiment can reflect the relative importance of different scenarios in the assessment by determining weights. Step S113 uses these preliminary assessment results and their corresponding weighted weights to calculate the final privacy understanding score, privacy sensitivity index, and level of concern. This weighted calculation method comprehensively considers the student's performance and importance in different scenarios, avoiding the assessment bias caused by simple averaging or ignoring scenario differences, thereby obtaining a more comprehensive and accurate assessment of the student's data privacy psychological state. Through the above scheme, this application can more accurately assess the student's data privacy psychological state in different situations and overcome the imprecision caused by simple mixed analysis or non-distinguished scenario assessment, thereby providing a basis for generating more effective and more tailored privacy setting guidance strategies for students' actual needs.

[0036] In some preferred embodiments, the process of performing private interaction with students according to the privacy setting guidance policy includes: S31. Obtaining the student's current online learning activity status information; S32. Analyze whether the student is currently suitable for private interaction based on the online learning activity status information. If so, conduct private interaction with the student based on the privacy setting guidance strategy.

[0037] Acquiring information about students' current online learning activity status means the system collects data such as students' ongoing learning tasks, operational behaviors, or system status while using the online education platform. This can be achieved by monitoring students' mouse and keyboard activity, video playback status, exercise completion progress, page dwell time, and currently active windows. Analyzing whether students are currently suitable for private interaction based on online learning activity status information means the system determines whether students are currently in a learning context that should not be interrupted based on the acquired status information. This can be achieved by setting rules or models to assess students' level of engagement, task criticality, or interruption sensitivity.

[0038] Specifically, this solution solves the problem of conducting private interactions at inappropriate times by introducing a judgment of the student's current online learning activity status. Step S31 obtains the student's current online learning activity status information, which provides the basic data required for judging the timing of the interaction, so that the system can understand whether the student is currently performing an important learning task or is in a state where they do not want to be interrupted. Step S32 analyzes and judges whether the student is currently suitable for private interaction based on the obtained online learning activity status information. Only when it is judged to be appropriate will the subsequent private interaction be performed. This judgment mechanism ensures that private interactions will not be carried out when the student is highly engaged in learning, thereby avoiding interference with the student's learning process and improving the acceptance and effectiveness of the interaction, so that the privacy setting guidance strategy generated based on psycholinguistics can be more smoothly accepted and adopted by students, ultimately better achieving the purpose of data privacy protection. Through the above solution, this application avoids conducting private interactions when students are focused on learning activities, reduces interruptions to the student's learning process, improves the effectiveness of private interactions, reduces the possibility of causing student resistance, and fully considers the impact of interaction timing on student experience and privacy setting effects.

[0039] In some preferred embodiments, performing private interaction with students according to the privacy setting guidance strategy further includes the following steps performed after step S32: S33. During privacy interaction with students, obtain students' interactive feedback, and then adjust the privacy setting guidance strategy based on the interactive feedback. The interactive feedback includes students' reply content and reactions, as well as their adoption of the privacy setting guidance strategy.

[0040] Interactive feedback refers to the collection of information generated by students during the privacy setting guidance interaction with the system, reflecting their immediate status and understanding of the interaction content. Students' responses and reactions refer to the direct information (response content) expressed by students during the interaction through text, voice, or other forms (reactions), as well as their non-verbal or indirect status (reactions). These reactions can include input speed, pauses, facial expressions, and click behaviors. Response content can be analyzed using natural language processing technology, and reactions can be obtained using behavioral analysis or physiological signal analysis technology. Adoption of privacy setting guidance strategies refers to the degree to which students accept or implement the privacy setting suggestions, explanations, or operational instructions currently proposed by the system. This can be achieved by tracking students' interactive behaviors, such as whether they clicked "Agree," "Accept," or "Next," whether they modified settings, and how long they stayed on a certain step. Adjusting the privacy setting guidance strategy based on interactive feedback means that the system dynamically modifies or optimizes the currently executed privacy setting guidance strategy based on the interactive feedback information obtained. This embodiment can use rule-based reasoning, machine learning models or preset adjustment logic to adjust the privacy setting guidance strategy based on interactive feedback. For example, the level of detail of the explanation can be adjusted according to the confusion expressed by students, the language style can be adjusted according to the resistance, and the guidance process can be adjusted according to unsatisfactory adoption.

[0041] Specifically, during privacy interactions with students based on the privacy guidance strategy, the system no longer relies solely on the pre-set strategy generated before the interaction begins. Instead, it obtains real-time feedback from students. This feedback includes the content of students' responses, reactions, and adoption of the current privacy guidance strategy. The system analyzes this real-time feedback to determine the students' current understanding, emotional state, level of acceptance, and effectiveness of the current strategy. Based on the analysis results, the system dynamically adjusts the current privacy guidance strategy. This adjustment can involve changing the language style of the interaction, the level of detail of the explanation, the order of the guidance steps, or reorganizing the way information is presented. Through the above scheme, students' interactive feedback is obtained in real time during the privacy interaction process, and the privacy setting guidance strategy is dynamically adjusted according to the feedback. It is possible to promptly identify students' confusion, misunderstanding, resistance or non-acceptance of the strategy during the interaction process. The strategy adjustment based on these real-time feedback makes the interaction content, style and process more closely fit the students' immediate status and needs. Therefore, this embodiment can effectively improve the pertinence and effectiveness of privacy setting guidance and enhance the personification and adaptability of the interaction. This helps students better understand complex privacy setting options and their impact and reduces the possibility of being forced to accept default settings due to misunderstanding, thereby promoting students to make privacy setting choices that are in line with their true intentions based on full understanding and improving students' willingness to adopt privacy settings.

[0042] In some preferred embodiments, the process of obtaining original text data generated by students when using the online education platform and then performing desensitization processing on the original text data to obtain desensitized text data includes: S41. Obtaining original text data generated by students when using the online education platform; S42, preprocessing the original text data to obtain preprocessed text data; S43. Desensitizing the preprocessed text data to obtain desensitized text data.

[0043] Preprocessing refers to the process of cleaning, normalizing, and structuring text data before further analysis or processing. This can be achieved through techniques such as data cleaning, word segmentation, part-of-speech tagging, stop word removal, and standardization. Desensitization refers to the process of processing text data containing sensitive information so that it no longer contains directly identifiable personal information or sensitive content. This can be achieved through techniques such as replacement, generalization, deletion, and encryption. Through this approach, noise and inconsistencies in the original text data are effectively removed, improving data quality. Personally identifiable information and sensitive content in the text are effectively identified and processed, reducing the risk of privacy leaks. This provides clean and secure input data for subsequent psycholinguistic feature analysis, effectively improving the accuracy of subsequent analysis results.

[0044] In some preferred embodiments, preprocessing includes data cleaning and word segmentation, and desensitization includes replacing or generalizing personal identity information and sensitive topic words in the text. Data cleaning refers to the process of normalizing and denoising raw text data, which can be achieved by removing special characters, correcting typos, and unifying expression formats. Word segmentation refers to the process of dividing a continuous text sequence into semantically meaningful words or phrase units, which can be achieved by using dictionary-based methods, statistical model-based methods, or deep learning model-based methods. Desensitization refers to the process of converting sensitive information in the text so that it is no longer directly associated with a specific individual or reduces its sensitivity, which can be achieved by using replacement or generalization processing techniques. Personal identity information refers to text information that can directly or indirectly identify a specific individual, which can include name, student ID number, contact information, ID number, etc. Sensitive topic words refer to words or phrases related to personal privacy, health, family, emotions, beliefs, etc. that may lead to privacy leakage or cause discomfort. Replacement refers to the process of directly replacing sensitive information with a preset placeholder or anonymous identifier, which can be achieved by using rule-based replacement or lookup table-based replacement. Generalization refers to the process of replacing sensitive information with a broader category or scope to which it belongs, which can be achieved using classification-based methods or clustering-based methods.

[0045] In some preferred embodiments, psycholinguistic features include lexical features, syntactic features, emotional features and expression features. Lexical features refer to the selection of words in a text, the frequency of use and the relationship between words, which can be achieved by using technologies such as word frequency statistics, keyword extraction and word vector representation. Syntactic features refer to the structure and composition of sentences in a text and the relationship between sentences, which can be achieved by using technologies such as syntactic analysis, dependency analysis and sentence complexity measurement. Emotional features refer to the emotions, attitudes and emotional tendencies expressed in a text, which can be achieved by using technologies such as emotional dictionary matching, emotional classification models and emotional intensity analysis. Expression features refer to the expression style, tone and rhetoric of the speaker or author in the text, which can be achieved by using technologies such as pragmatic analysis, rhetorical analysis and style transfer models.

[0046] From the above, it can be seen that the present application provides an artificial intelligence anthropomorphic data privacy protection platform based on psycholinguistics, which can first analyze the original text data generated by students in the process of using the platform to extract psycholinguistic features, and then obtain the students' data privacy psychological state assessment results based on the psycholinguistic feature analysis, and finally generate a privacy setting guidance strategy based on the data privacy psychological state assessment results to achieve personalized and anthropomorphic privacy setting guidance strategy based on the students' individual data privacy psychological state. That is, the present application can enable the privacy policy to be fully understood by students. Therefore, the present application can effectively solve the problem that students are forced to accept the platform's default options when making privacy settings because they cannot fully understand the text agreement, thereby effectively avoiding the situation where students' right to informed consent is damaged and the data collected by the platform fails to fully reflect the students' true intentions due to being forced to accept the platform's default options. That is, the present application can effectively improve the accuracy and reliability of students' data privacy settings.

[0047] In the embodiments provided in this application, it should be understood that the disclosed devices and methods can be implemented in other ways. The device embodiments described above are merely schematic. For example, the division of the above-mentioned units is only a logical function division. There may be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another robot, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some communication interface, the indirect coupling or communication connection of the device or unit can be electrical, mechanical or other forms.

[0048] In addition, the functional modules in each embodiment of the present application can be integrated together to form an independent part, or each module can exist independently, or two or more modules can be integrated to form an independent part.

[0049] In this document, relational terms such as first and second, etc. are used merely to distinguish one entity or operation from another entity or operation, but do not necessarily require or imply any actual relationship or order between these entities or operations.

[0050] The above are merely examples of the present application and are not intended to limit the scope of protection of the present application. Those skilled in the art will appreciate that various modifications and variations are possible. Any modifications, equivalent substitutions, or improvements made within the spirit and principles of the present application shall be included within the scope of protection of the present application.

Claims

1. An artificial intelligence anthropomorphic data privacy protection platform based on psycholinguistics, characterized by: The artificial intelligence anthropomorphic data privacy protection platform based on psycholinguistics includes: A data processing module is used to obtain original text data generated by students when using the online education platform, and then desensitize the original text data to obtain desensitized text data; A feature acquisition module, configured to analyze the desensitized text data to obtain psycholinguistic features for characterizing the student's data privacy psychological state; a strategy generation module, configured to obtain a data privacy psychological state assessment result of the student based on the psycholinguistic feature analysis, and then generate a privacy setting guidance strategy based on the data privacy psychological state assessment result; An interaction module is used to perform private interaction with the student according to the privacy setting guidance strategy.

2. The artificial intelligence anthropomorphic data privacy protection platform based on psycholinguistics according to claim 1 is characterized in that: The process of obtaining the student's data privacy psychological state assessment result based on the psycholinguistic feature analysis and then generating a privacy setting guidance strategy based on the data privacy psychological state assessment result includes: S11. Obtaining a data privacy psychological state assessment result of the student based on the psycholinguistic feature analysis, wherein the data privacy psychological state assessment result includes a privacy understanding score, a privacy sensitivity index, and a degree of concern; S12. Determine the privacy interaction content and privacy interaction process according to the privacy understanding score and the privacy sensitivity index; S13. Determining a privacy interaction style according to the privacy sensitivity index and the degree of concern; S14: Integrate the privacy interaction content, the privacy interaction process, and the privacy interaction style to form a privacy setting guidance strategy.

3. The psycholinguistics-based artificial intelligence anthropomorphic data privacy protection platform according to claim 2 is characterized in that: Step S14 includes: S141: Integrate the privacy interaction content, the privacy interaction process, and the privacy interaction style to form a preliminary setting guidance strategy; S142. Obtaining the student's cognitive preferences and learning style; S143, determining a guidance adjustment strategy corresponding to the initially set guidance strategy according to the cognitive preference and the learning style; S144: Adjust the preliminary setting guidance policy according to the guidance adjustment policy to obtain a privacy setting guidance policy.

4. The artificial intelligence anthropomorphic data privacy protection platform based on psycholinguistics according to claim 3 is characterized in that: Step S143 includes: A1. querying a pre-built mapping relationship table of cognitive preferences, learning styles, and adjustment strategies based on the cognitive preferences and learning styles to obtain a preliminary adjustment strategy; A2. obtaining the student's current emotional state, and then determining a correction strategy for the preliminary adjustment strategy based on the current emotional state; A3. Adjust and calibrate the preliminary adjustment strategy according to the correction strategy to obtain a guided adjustment strategy.

5. The artificial intelligence anthropomorphic data privacy protection platform based on psycholinguistics according to claim 2 is characterized in that: The desensitized text data includes desensitized data of the student in different scenarios, and the process of analyzing the desensitized text data to obtain psycholinguistic features for characterizing the student's data privacy psychological state: S21. Analyze the corresponding desensitized data for each scenario to obtain psycholinguistic features corresponding to the scenario for characterizing the student's data privacy psychological state; Step S11 includes: S111. For each of the psycholinguistic features, obtain the student's preliminary understanding score, preliminary sensitivity index, and preliminary worry level according to the psycholinguistic feature analysis; S112, determining, based on all the scenarios, a first weight corresponding to each of the preliminary understanding scores, a second weight corresponding to each of the preliminary sensitivity indexes, and a third weight corresponding to each of the preliminary worry levels; S113. Calculate a privacy understanding score based on all of the preliminary understanding scores and their corresponding first weighted weights, calculate a privacy sensitivity index based on all of the preliminary sensitivity indices and their corresponding second weighted weights, and calculate a degree of concern based on all of the preliminary degrees of concern and their corresponding third weighted weights.

6. The artificial intelligence anthropomorphic data privacy protection platform based on psycholinguistics according to claim 1 is characterized in that: The process of performing privacy interaction with the student according to the privacy setting guidance strategy includes: S31, obtaining the student's current online learning activity status information; S32: Analyze whether the student is currently suitable for private interaction based on the online learning activity status information; if so, perform private interaction with the student based on the privacy setting guidance strategy.

7. The artificial intelligence anthropomorphic data privacy protection platform based on psycholinguistics according to claim 6 is characterized in that: The performing of privacy interaction with the student according to the privacy setting guidance strategy further includes the following steps performed after step S32: S33. During the privacy interaction with the student, obtain the student's interaction feedback, and then adjust the privacy setting guidance strategy based on the interaction feedback, wherein the interaction feedback includes the student's reply content and reaction and the adoption of the privacy setting guidance strategy.

8. The artificial intelligence anthropomorphic data privacy protection platform based on psycholinguistics according to claim 1 is characterized in that: The process of obtaining original text data generated by students when using the online education platform and then performing desensitization processing on the original text data to obtain desensitized text data includes: S41. Obtaining original text data generated by students when using the online education platform; S42, preprocessing the original text data to obtain preprocessed text data; S43: Desensitizing the preprocessed text data to obtain desensitized text data.

9. The artificial intelligence anthropomorphic data privacy protection platform based on psycholinguistics according to claim 8 is characterized in that: The preprocessing includes data cleaning and word segmentation, and the desensitization processing includes replacing or generalizing personal identity information and sensitive topic words in the text.

10. The artificial intelligence anthropomorphic data privacy protection platform based on psycholinguistics according to claim 1 is characterized in that: The psycholinguistic features include lexical features, syntactic features, emotional features and expression features.

Images