Copyright and Attribution When Treating AI as a Co-Author: This Blog's Policy and Current Understanding
About This Article
While writing this blog, there were two questions I wanted to address properly at some point: “Who holds the copyright when AI generates the text?” and “Should I disclose to readers that AI was involved in writing?”
The more I researched, the clearer it became that the law has not yet caught up in many areas, and that practical decisions must be made by each person based on the information currently available. This article explains what I found and the policy I have adopted for this blog.
Japan’s Current Copyright Framework
Japan’s Copyright Act defines a copyrighted work as something that “creatively expresses thoughts or feelings,” and copyright belongs in principle to a natural person or a legal entity. Because AI does not have legal personhood, the generally accepted understanding at this point is that copyright does not arise in text that AI generates autonomously [1].
However, this does not mean “text generated by AI is entirely in the public domain.” According to the Agency for Cultural Affairs, when a human uses AI as a tool and the resulting work reflects creative direction, selection, or editing by that human, copyright for the human’s contribution may be recognized [1].
What counts as “creative human involvement” varies by situation. Case law in this area is still limited, and there is considerable room for interpretation.
How I Use AI on This Blog
The main situations in which I use AI on this blog are as follows.
- Giving instructions about article structure and having a draft generated
- Asking for tone and phrasing adjustments
- Having English translations generated
- Reading through what was generated and revising and adding to it myself
The final article is not simply published AI output. All direction, editing, and fact-checking is handled by me.
This Blog’s Disclosure Policy
I have adopted a policy of not concealing from readers that AI was involved in writing the articles. There are two reasons for this.
The first is a matter of trust with readers. Whether a reader knows that AI contributed to an article affects how they receive it. I want this blog’s use of AI to be something readers can find out about.
The second is a matter of my own integrity. “Presenting text that AI wrote as something I wrote” is not accurate as a statement of fact. I believe that being explicit about the stance — AI is used as a tool, but the final judgment and responsibility rest with a human — is more honest over the long term.
On Future Changes
The question of AI and copyright is one where the law has not kept pace with technological development. The Agency for Cultural Affairs continues to refine its guidance[1], and it is not certain that today’s framework will remain valid in the future.
The content of this article represents my understanding as of June 2026. If the law is revised or new precedents emerge, I will revisit the policy of this blog accordingly.
Summary
- In Japan at this point, the generally accepted view is that copyright does not arise in text that AI generates autonomously.
- When a human is creatively involved, copyright for that human’s contribution may be recognized.
- On this blog, I use AI as a tool, with all direction, editing, and fact-checking handled by me.
- I follow a policy of disclosing to readers that AI is used.
Both the law and practical norms in this area continue to evolve, and I intend to revisit this periodically.
References
- Agency for Cultural Affairs (Japan), 考え方について AIと著作権に関する, March 2024