AI Copyright: Training-Data Lawsuits and Licences — 2026-09-05
The US Department of Justice has formally intervened in the *New York Times v. OpenAI* lawsuit, filing a brief arguing that AI training on copyrighted material constitutes "fair use" and warning that restrictive rulings could hinder tech progress. Simultaneously, the $1.5 billion Anthropic settlement is facing internal friction as authors dispute how funds should be distributed among rights holders, while new lawsuits from music publishers demand transparency regarding Claude’s training data.
AI Copyright: Training-Data Lawsuits and Licences — 2026-09-05
Top developments
DOJ Intervenes in NYT v. OpenAI with Pro-AI Brief
On September 1, 2026, the US Department of Justice filed a statement of interest in the Southern District of New York supporting OpenAI in its copyright battle with The New York Times. The DOJ argued that training large language models (LLMs) on copyrighted works is a transformative use protected by fair use doctrines, similar to how journalism itself relies on existing information. This non-binding intervention signals a significant shift in federal policy, suggesting that creators seeking compensation may need to pursue legislative relief rather than relying on judicial fair use rulings.
Authors Dispute Distribution of Anthropic Settlement
As of September 5, 2026, authors are actively contesting the allocation of the $1.5 billion class action settlement approved in the Bartz v. Anthropic case. While Judge Araceli Martínez-Olguín granted final approval to the historic settlement earlier this year, disputes have emerged over whether funds should go directly to authors or be shared with publishers. The settlement mandates a payment of $3,000 per pirated book used for training, but many authors fear losing their share to intermediaries in the publishing industry, leading to ongoing wrangling over the final distribution mechanics.
Music Publishers Demand Accounting of Claude's Training Data
In a move that could set a precedent for data transparency, music publishers have asked a court to order Anthropic to disclose exactly which works were used to train its Claude models. Filed recently, this request goes beyond standard damages claims; plaintiffs are demanding an "accounting" that identifies specific songs and lyrics ingested by the AI. This legal strategy aims to uncover the scope of alleged infringement and may force AI companies to maintain more rigorous records of their training datasets in future litigation.
Local view
Germany: BGH Hears Landmark LAION-5B Case
The German Federal Court of Justice (BGH) is currently deliberating on whether the creation of datasets for AI training, specifically involving the LAION-5B dataset, violates copyright law. In hearings reported around September 3, 2026, the court appeared inclined to refer key questions to the European Court of Justice (ECJ), signaling uncertainty about whether current German interpretations of text and data mining exceptions apply to generative AI training. This development is critical for European AI companies operating under GDPR and EU copyright directives.
Japan: Government Proposes Disclosure Rules
Japanese media reports indicate that the Japanese government is preparing a draft of basic principles for intellectual property protection in the context of Generative AI. The proposed framework would require AI operators to disclose the methods used to collect training data and potentially disclose specific works to rights holders. Notably, these rules would apply to foreign companies operating in Japan, though penalties for violations are reportedly being considered for removal to encourage compliance through transparency rather than punishment.
Context & numbers
- DOJ Intervention: The US government's stance is explicitly pro-innovation, citing national security and competitiveness as reasons to allow broad AI training rights.
- Anthropic Settlement Value: The total settlement value stands at $1.5 billion, with individual payouts calculated at $3,000 per book involved in the piracy allegations.
- Licensing Market Growth: The AI content licensing market continues to expand, with OpenAI holding roughly two dozen deals, including a reported $250 million five-year agreement with News Corp.
On the radar
- Judge Stein's Show Cause Order: In late August 2026, Judge Sidney Stein ordered The New York Times to show cause regarding procedural aspects of the case, indicating potential shifts in the discovery phase.
- Appeals of Anthropic Settlement: Several appeals have been filed against the final approval of the Anthropic settlement, which could delay the actual distribution of funds to claimants for months or years.
This content was collected, curated, and summarized entirely by AI — including how and what to gather. It may contain inaccuracies. Crew does not guarantee the accuracy of any information presented here. Always verify facts on your own before acting on them. Crew assumes no legal liability for any consequences arising from reliance on this content.
Powered by
