Japanese Content Industry Group Accuses OpenAI's Sora 2 of Copyright Infringement
Japanese content companies and publishers, represented by CODA, accused OpenAI's Sora 2 video generation system of copyright infringement, claiming it reproduces Japanese anime and game characters without permission and requesting OpenAI stop using member content as training data.
What happened
Sora 2, a video generation AI service launched on September 30, 2025, can reportedly generate videos depicting Japanese anime and game characters, according to observations shortly after its release.
On October 27, 2025, CODA (γ³γ³γγ³γζ΅·ε€ζ΅ιδΏι²ζ©ζ§), a coalition of 36 Japanese content companies and 11 organizations including publishers, anime studios, and production companies, submitted a written request to OpenAI demanding two actions: that Sora 2 not use CODA member content as training data without permission, and that OpenAI respond sincerely to copyright infringement claims and consultations from CODA members.
Reported impact
CODA stated it confirmed that Sora 2 generates videos that are identical to or closely resemble existing Japanese content.
CODA argues that under Japanese copyright law, use of copyrighted works requires prior permission, and that learning processes involving reproduction of copyrighted works could constitute copyright infringement itself, even if generated outputs are similar to existing works.
Index analysis
OpenAI reportedly offers an opt-out mechanism allowing rights holders to request exclusion from training, but CODA contends this post-hoc approach contradicts Japanese copyright principles requiring advance permission.
The Motion Picture Association also raised concerns about Sora 2 in October 2025, stating it would take countermeasures against the system.
Governance controls
OpenAI established an opt-out submission process for rights holders regarding Sora 2 training data.
CODA stated it would work with member companies to seek sincere response from OpenAI, aiming to balance healthy AI technology development with creator and rights holder protection.
Classification
Sources and evidence
Sources not yet published
This record was researched and written by the Index from original reporting. Its sources are still being verified by an editor, and they are listed here once that check is complete. Until then the record carries no reviewer.