Creative thinking support generative AI system

Abstract

This invention provides a web application system that allows instructors to effectively use generative AI in classes to match learning tasks, reducing the risk of students asking questions that are inappropriate for the task or trusting false information from the generative AI, and encouraging students' creative thinking through dialogue using generative AI, with the aim of supporting creative activities. SOLUTION: The web application system of this invention creates prompts by integrating questions from learners with prompts preset by the instructor, and provides automatically generated answers through API connection. The app is built using HTML code with a website creation tool, so it can be managed without specialized knowledge. Prompts are edited using three items: Theme, Conditions, and Additional Instructions, to control the answers of the generating AI. In addition, restrictions on questions outside the topic can be set to prevent misuse.

Description

[Technical field]

[0001] This proposal is a web application system that works in conjunction with a generative AI API to broaden the thinking of system users (learners) through advice and counter-questions in response to their questions, thereby helping them generate creative ideas that are aligned with their learning objectives. [Background technology]

[0002] When using general-purpose generative AI services in elementary, junior high, and high school classes, there was a concern that students would be tempted to blindly accept the answers given by the generative AI due to the high degree of freedom in their conversations with the AI, and thus stop thinking for themselves. [Prior art documents] [Patent documents]

[0003] JP 2024-148117 Published Patent Application [Non-patent literature]

[0004] none Summary of the Invention [Problem to be solved by the invention]

[0005] In classes, it is difficult for teachers to use general-purpose generative AI services to manage and operate the dialogue between learners and generative AI. There is also a risk that learners will ask questions that are not in line with the learning task, or that incorrect information will be generated as the answer from the generative AI. Furthermore, dialogue that encourages creative thinking to deepen learning, such as questions and advice from the generative AI, requires ingenuity in prompts. Therefore, this system aims to solve the above-mentioned shortcomings by allowing the system administrator (instructor) to set learning tasks and dialogue conditions, etc. as prompts in advance within the system, and for learners to simply input the dialogue text on the display screen of the web application to generate a prompt by integrating the advance prompt and the dialogue text. [Means for solving the problem]

[0006] This proposal solves the problem by having the system, when a learner inputs a question, create a prompt by integrating the learner's "question" with a "prompt" pre-set within the system, communicate data with the generative AI model via an API, obtain a response from the generative AI to the prompt, and output the answer to the learner.

[0007] The system is implemented using HTML code, and a website can be created by embedding the implemented code using a general website creation tool that is available free of charge. Because it is available on the web, learners can easily use it by sharing the URL. Instructors can edit the system's top screen using the website creation tool without needing specialist knowledge, and add supplementary information and explanations about the learning tasks.

[0008] The preset prompts are structured by incorporating the variables Theme, Conditions, and Additional Instructions into the system settings. These settings are written in HTML code, and are intended to allow instructors using the system to set the conditions required for the generation AI in advance according to the purpose of the lesson.

[0009] The Theme allows students to input learning tasks such as, "Let's think of ideas for getting along with transfer students from overseas through activities at school and in the community."

[0010] The conditions clearly state the position of the questioner, stating that "the questioner is a sixth grade elementary school student." It also stipulates the intention to use the system and the conditions for generating answers required of the AI, stating that "the system will answer questions from the questioner, ask the questioner questions in return about points that the questioner does not understand, and support the questioner in coming up with ideas related to the topic. The system will provide answers that understand the topic and give advice and questions, rather than giving 'answers' that stop the student from thinking."

[0011] Additional instructions can include additional conditions for the generation AI, such as "The generation AI should be aware of the questioner's grade and situation, and should answer using clear words and kanji without using technical terms. The AI ​​should be aware that the questioner is trying to generate ideas, and when the questioner gives an ambiguous opinion, it should include a question related to that opinion so that the questioner can generate another idea." In addition, it includes answering constraints such as "For questions that are not related to the topic, please reply, 'I will not answer the question because it is off topic.'" In addition, the number of output characters is specified as the number of tokens for the answer text generated by the generation AI to improve readability for the children and students who will be the learners. Effect of the Invention

[0012] The instructor can assign a role (position) to the generating AI and specify the desired answer and the way of answering. By changing the Theme and Conditions, the instructor can adjust the content of the generated answers, making it easy to customize them to suit the purpose of the lesson.

[0013] The feature is that it is possible to pre-set prompts so that the generative AI does not give an "answer" but asks questions in return, allowing it to come up with ideas related to the theme.

[0014] Prompts are preset so that the AI ​​does not answer questions that are not related to the lesson. This prevents students from blindly accepting the AI's answers, and allows instructors to use the AI ​​with confidence. [Brief description of the drawings]

[0015] [Figure 1] This is a conceptual diagram of our creative thinking support generative AI system. [Diagram 2] FIG. 1 is a diagram showing the display screen of the creative thinking support generation AI system. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0016] Obtain the API of the generation AI and enable sending prompts and receiving answers from the generation AI using HTML code. To make the system easily configurable, use a general website creation tool that can embed HTML code to create a website.

[0017] The HTML code consists of 1) setting the basic HTML structure, 2) metadata and style settings, 3) page heading, input form, and button element settings, 4) output area settings, 5) pre-prompt settings, 6) integration of questions from the input form with pre-prompts and past dialogue history, 7) sending prompts integrated with API connection to the generation AI, and 8) obtaining and displaying output from the generation AI. The system is configured by embedding the HTML code, transferring the created website to the server, and publishing the website, linking with the user's terminal, web server, and API.

[0018] By sharing the website URL with students, instructors can use the system without restrictions on location, time, or number of students. In addition, by editing the pre-prompt portion of the HTML code, instructors can control the answer conditions and themes of the AI ​​generation to suit the lesson. EXAMPLES

[0019] Figure 1 is a conceptual diagram of this creative thinking support generation AI system, in which when a question is entered on the user (learner) terminal (1), the data is sent to a web application (2). In the web application, the system administrator (instructor) imports learning tasks and constraints into the application as pre-set prompts, and the questions and prompts are integrated in the web application (2).

[0020] The integrated prompts are sent to the generative AI model (3) via an API connection, and the generative AI model generates answer sentences, which are received by the web application (2). The received data is displayed on the learner's device (1).

[0021] Figure 2 shows the display screen of this creative thinking support generation AI system, with 4 showing the entire screen. 5 is an input form where learners input ideas and questions according to their learning tasks.

[0022] 6 is the send button for input content. By pressing the button, the input idea or question is sent to the system and integrated with the pre-prompt in the system. After that, it automatically connects to the generation AI via API, and after sending data to the generation AI, the generation AI generates an answer, and the data is received in the system.

[0023] Numeral 7 is an output field that outputs the answer from the generation AI, and displays the data received from the generation AI. Because the number of characters is limited by a pre-prompt, the generated answer is displayed in a number of characters that the learner can read in a short time. Numeral 8 is an answer history display field, and automatically displays the dialogue history between the learner and the generation AI. Specifically, it displays the timestamp, the learner's question, and the answer from the generation AI, and the answer history is displayed in order of most recent. [Explanation of symbols]

[0024] 1 User (student) terminal 2. Web Applications 3. API-connected generative AI model 4. Entire display screen of student's device 5. Input Form 6. Send button 7 Output display field 8 History display field

Claims

[Claim 1] A web application system that works in conjunction with a generative AI API to broaden the thinking of system users (learners) by providing advice in response to their questions and asking questions in return, helping them generate creative ideas that align with their learning objectives. The system administrator (instructor) sets up pre-prompts in the system, such as learning tasks and dialogue conditions, and when a learner enters a conversation sentence (question) on the system's display screen, the system automatically integrates the pre-prompt with the learner's question, sends the data to the generative AI model via API communication, receives the response by the generative AI as data, and outputs and displays the answer generated by the generative AI on the screen.

Images