Usage Guide for Generative AI Image Tools in Educational Institutions
[Disclaimer] This guide was created personally and does not represent the views of any affiliated organizations. Please use it as a reference material for voluntary use.
ChatGPT's New Image Generation Capabilities
The image generation capabilities in ChatGPT reached a major turning point in March 2025. OpenAI replaced the traditional DALL·E 3 system with native implementation of image generation capabilities within the GPT-4o model itself [1].
Notably, the rendering accuracy of images containing Japanese text has improved significantly, making it easier to create more practical and precise images, such as diagrams, presentation visuals, manga-style illustrations, and logo designs.
In the context of university education, it is expected that faculty will use these tools to quickly prepare diagrams tailored to lecture content, and students will generate visual elements for reports and presentation materials themselves. However, the fact that anyone can now easily generate high-quality images also suggests a need to raise awareness of copyright and ethical issues once again.
Distinction Between "Expression" and "Idea"
Under Japanese copyright law, the subject of protection is limited to creative expressions of thoughts or emotions—that is, specific expressions. Ideas themselves are not protected (Article 2, Paragraph 1, Item 1 of the Copyright Act).
Based on this principle, styles such as artistic style or technique are considered to be in the realm of ideas, and it is interpreted that imitating them does not immediately constitute copyright infringement. The Agency for Cultural Affairs has also expressed the view that images generated using prompts such as "in the style of a specific artist" do not necessarily constitute copyright infringement. Images obtained by instructing generative AI to use a specific style are generally considered expressions of new ideas rather than reproductions of existing copyrighted works. [2]
However, not all output from generative AI can be used freely. It is necessary to pay attention to whether the specific expression is similar to an existing copyrighted work. If an AI-generated image is very similar to a specific existing copyrighted work, even if the instruction was to use a "style of X," the act of using that generated product may be judged as infringing on the reproduction or adaptation rights of the original work.
In Japanese law, "similarity to an existing work" and "reliance" (accessibility) on that work are important factors in determining copyright infringement. Since generative AI learns from vast amounts of existing work data, it is difficult to deny the possibility that the generated results rely on the source copyrighted works, and if the similarity is high, there is a risk that it will not be considered an independent creation. However, even if it is highly likely that the generative AI tool used did not learn from existing work data, the current legal interpretation is that if a high degree of similarity is recognized, reliance is presumed (Reference: The Boiled Bean Seller Case). Therefore, when publishing or using generated images, it is required to carefully determine whether they can be called unique creative expressions that are clearly different from any existing works. [3]
Furthermore, in extreme situations that unfairly harm the interests of the rights holder of the work in question, even if similarity and reliance are not significant, the exception clauses of the copyright limitation provisions may be applied, and it may be judged as illegal. [4]
When students use generative AI in an educational setting, they need to fully understand this point. It is important for instructors to discuss with students and provide appropriate guidance on what kind of social reactions and legal evaluations might result from easily publishing works that simply imitate the style of famous artists.
(※ Exception provisions for educational purposes)
When using generative AI in educational institutions such as universities, it is necessary to pay attention to both the management of legal risks and ethical perspectives. First, from the perspective of copyright compliance, it is required to accurately understand the exception provisions for educational purposes (Article 35 of the Copyright Act). Article 35 is a general exception provision regarding the use of copyrighted works in educational institutions. This is not a provision specific to generative AI, but it stipulates that teachers and students may reproduce or publicly transmit copyrighted works for use in the course of lessons at schools and other educational institutions. However, this is limited to use in the limited situation of a lesson.
Therefore, even if an image created by a student or teacher using generative AI happens to be similar to an existing copyrighted work, if its use is within the scope of the lesson, it is permitted under Article 35. However, if the generated product is distributed outside of the lesson or published on the Internet, this educational use exception does not apply. Therefore, it is wise to check in advance whether there is any similarity to specific works when using AI-generated products as lesson materials and avoid use that exceeds the scope of the lesson.
Ethical Considerations
All creative and expressive activities, not limited to generative AI, are accompanied by socially required ethical responsibilities. Ethical considerations include ensuring respect for creators and transparency. Even if there are no legal issues, when generating and publishing images that imitate the style of a specific individual artist, it is desirable to be considerate of the feelings of that artist and related communities. For example, disclosing sources or reference information, such as clearly stating in the description of the work that it is "a unique work inspired by the style of Mr./Ms. X" or specifically indicating the names of the works referenced when citation is possible, is a practice that also aligns with research ethics.
Furthermore, it is important to ensure that the content of the work does not promote discrimination or prejudice regarding race, gender, etc., does not violate public order and morals, and does not unfairly infringe on the personality rights or privacy of others. Understanding that freedom of creation comes with responsibility, and eliminating discriminatory expressions and maintaining appropriate information ethics are attitudes always required, regardless of whether generative AI is used or not. Although these are not direct copyright issues, they are essential perspectives for educational institutions to cultivate human resources with the ability to use AI responsibly in society. [5]
In summary, in the context of university education, it is necessary to leverage the benefits of generative AI while correctly understanding the scope and limitations permitted by copyright law, and for stakeholders to cooperate in developing and revising guidelines. As legal interpretations and social consensus may change with technological progress, it is required to keep a close watch on the latest trends and practice the use of AI while respecting copyright and ethics.
References
Introducing 4o Image Generation. Accessed March 30, 2025. https://openai.com/index/introducing-4o-image-generation/
[AI Copyright Issues: Is the World Created by AI an 'Expression'? An 'Idea'? Or...] | shiba@Creative Management. note. September 8, 2023. Accessed March 30, 2025. https://note.com/48art/n/nbea41b5f0972
Do text and images created with generative AI constitute copyright infringement? Points to keep in mind when using them. Accessed March 30, 2025. https://www.docusign.com/ja-jp/blog/AI-Engineer-Explains-Copyright-for-Generative-AI-Content
Taichi Kakinuma. [Series] Generative AI and Copyright: Based on the 'Concept' of the Copyright Subcommittee of the Council for Cultural Affairs, Part 2 | For legal consultations on intellectual property, IT, artificial intelligence, and venture business [STORIA Law Office]. STORIA Law Office. April 11, 2024. Accessed March 30, 2025. https://storialaw.jp/blog/10782
Handling of Generative AI in Teaching at Universities and Colleges: MEXT. Ministry of Education, Culture, Sports, Science and Technology Website. Accessed March 30, 2025. https://www.mext.go.jp/b_menu/houdou/2023/mext_01260.html
(*Regarding the images in this article)
The images in this article were generated by 4o-Image-Generation based on the article content, and were partially modified and used with the graphic design tool Canva.