We would like to share an important update from Japan that may significantly affect companies involved in generative AI development, content creation, and data analytics.
Revisions to Japan’s Copyright Law in Response to the Rise of Generative AI
In recent years, generative AI—such as ChatGPT—has rapidly expanded across industries worldwide, supporting content creation, creative assistance, and data analysis. These technologies rely heavily on large volumes of existing content used as training data for machine learning processes.
Japan’s Legal Framework: More Permissive Than Many Others
Under Japan’s Copyright Act (Article 30-4), amended in 2018, both commercial and non-commercial entities are allowed to use copyrighted works for data analysis—including machine learning—without obtaining prior permission from copyright holders. This legal environment is considered among the most flexible in the world.
Here is a brief comparison with other major jurisdictions:
| Criteria | Japan | European Union (EU) | United States (U.S.) |
|---|---|---|---|
| Use of copyrighted works for AI training | Explicitly permitted (Article 30-4) | Limited; opt-out by rightsholder required | Relies on “Fair Use” doctrine; case-by-case |
| Use for commercial purposes | Permitted | Generally not permitted unless licensed | Often litigated; no clear statutory rule |
| Compensation or licensing requirements | None (currently) | Opt-out system in place | None; litigation is primary enforcement route |
As a result, Japan currently provides a relatively risk-free legal environment for AI training, but this may soon change.
Rights Holders’ Pushback and Policy Response
In December 2024, the Japan Newspaper Publishers & Editors Association submitted a formal request to the Japanese government to reconsider this approach. The Association raised concerns that current laws allow tech companies to use copyrighted content freely without fair compensation, potentially threatening the sustainability of Japan’s cultural and creative industries.
Their demands include:
-
Transparency on the origin of training data
-
A framework for licensing or compensation
-
Clarification on the ownership and use of AI-generated works
The Japanese government is now expected to address these concerns in its upcoming Intellectual Property Strategic Program 2025. Legislative changes—either amending Article 30-4 or introducing new provisions—may follow shortly thereafter.
Looking Ahead: What Should Foreign Companies Be Aware of?
As Japan moves toward potential legal reform, foreign companies engaging with the Japanese market should prepare for regulatory changes in the following areas:
-
Training Data Compliance:
You may be required to disclose data sources or obtain permission when training on Japanese content, particularly from major publishers or media companies. -
Licensing and Royalty Systems:
New licensing frameworks or royalty schemes may be introduced, increasing operational complexity and cost. -
AI-Generated Outputs:
Clarification may be provided regarding whether outputs derived from Japanese works can be freely used or monetized. -
Legal and Reputational Risk:
The current permissive regime may tighten, increasing exposure for companies relying on broad data usage without localization or rights clearance. -
Cross-border Strategy:
Diverging international rules may require region-specific AI data governance and IP strategies.
Final Thoughts
While Japan has long provided a welcoming legal environment for AI development, it is increasingly aligning with international trends emphasizing transparency, fair use, and creator compensation. Companies that proactively review their data policies and IP frameworks now will be best positioned to adapt swiftly to the coming changes.
