
AI and Copyright, What Is Being Debated Now: Key Issues for Creators in 2026
This article was translated from Japanese and was originally written for readers in Japan.
AI and copyright is a topic where opinions split easily, that gets technical, and that is tiring to keep up with. Even so, people who create and publish cannot avoid it. This article narrows things down to two points, what is currently at issue in Japan and how you should act as an individual who creates, and lays them out as plainly as possible (the content is based on publicly available information as of June 2026 and is not legal advice).
What Is Being Debated Right Now?#
The Agency for Cultural Affairs has presented its General Understanding on AI and Copyright in Japan and has been sorting out the rights issues around generative AI. The issues that keep coming up in the news can be boiled down to three broad ones.
- How training data is handled: How far to permit the relationship between copyright and the data AI uses for training
- The problem of outputs resembling someone's work: The criteria for judging whether it is infringement when an output is similar to an existing copyrighted work
- Compensation for creators and transparency: How to reward the creators whose work was used for training, and how far to disclose what data a model was trained on
In recent years, "compensation" and "transparency" have become a particular focus. In the intellectual property field as well, work has been under way to set out principles of conduct that call for transparency from AI developers and providers.
It Is Not That Things Have Been "Settled in Black and White"#
What you should keep in mind is that many of these are still in the middle of being debated and sorted out. At this point, no simple line of "do this and you are completely safe" has been drawn. That is exactly why it helps to keep your distance from categorical claims and to take the stance of checking primary sources (materials published by the Agency for Cultural Affairs and others).
How to Act as an Individual Who Creates#
Even if the issues have not fully settled, there are things you can do within the bounds of self-protection and good sense.
- Avoid generation aimed at imitating a specific artist or work: Uses that target a proper name, such as "in the style of so-and-so," easily become a source of trouble
- Don't flatly claim an output as "my original" as is: Check with your own eyes before publishing that it does not closely resemble an existing work
- Don't hide the fact that you used AI where it matters: This is a sincere stance in line with the direction of the transparency debate
- Keep the perspective of protecting your own work, too: Ways to state your wishes if you do not want your work used for training are gradually being put in place
An Era in Which You Are Both "the One Who Uses" and "the One Who Is Used"#
Most publishers are on the side that uses AI and, at the same time, on the side whose publishing is used to train AI. That is why this topic is never someone else's problem. You don't need to be so afraid that you stop creating, but do hold on to the minimum good sense above. Striking that balance is the realistic middle ground for now.
Summary#
Around AI and copyright, the main issues are the handling of training data, the similarity of outputs, and compensation and transparency, and many are still being sorted out. As an individual who creates, the realistic stance is to stay within good sense: don't aim to imitate a specific artist, check outputs before putting them out, and don't hide AI use where it matters. Don't be overly afraid, and keep engaging with it while checking primary sources.
This is for you if you want to use AI as a "drafting partner" while finishing things in your own words. Toukoukun is a tool where, once you register what your product stands for, AI writes tailored drafts for each social network, and you add your own touch to complete them. As a way of publishing that deals with AI sincerely, try it from the dashboard.

