Anthropic Hits $1.5B Copyright Deal to Compensate 500,000 Works
On Monday, a U.S. federal court granted final approval for a $1.5 billion class-action settlement between Anthropic and a coalition of authors and publishers, as reported by Reuters. The deal resolves ongoing litigation regarding Anthropic’s alleged copyright infringement during the training of its AI models, with the company now obligated to begin disbursing compensation to the affected copyright holders.
U.S. District Judge Araceli Martinez-Olguin of the Northern District of California signed the final order ratifying the agreement. This follows a previous provisional approval by retired Judge William Alsup, who had previously determined that Anthropic had unlawfully downloaded and stored millions of copyrighted books.
Under the approved compensation framework, each affected work will receive $3,000 in damages. This figure applies to approximately 500,000 works, with the payouts distributed among the authors and publishers holding the relevant copyrights. The total sum ranks among the largest settlements in the history of U.S. copyright law.

Despite the settlement, legal uncertainties surrounding AI training data persist. Judge Alsup had previously sided with Anthropic on the core legal question, suggesting that using copyrighted texts to train AI models might qualify as "fair use"—a significant potential precedent in AI copyright disputes. However, he also noted that certain methods Anthropic used to acquire this data were illegal.
Anthropic acquired its training data through two primary channels: legally purchased and scanned books, and content scraped from pirate sites like Library Genesis and Pirate Library Mirror. The court ruled the latter approach illegal. To avoid prolonged litigation and the risk of substantial penalties, Anthropic opted for a settlement.
Industry analysts observe that while this agreement closes this specific case, it fails to establish a unified legal standard for the broader AI sector. Because Anthropic settled rather than appealing, Judge Alsup’s prior comments on "fair use" in AI training will not become binding judicial precedent.
Lawsuits concerning the copyright of AI training data remain active, with cases pending against major tech firms including Google, Meta, Midjourney, and OpenAI. Recently, publishers such as Hachette, Cengage, and Elsevier, along with author Scott Turow and the organization SCRIBE, filed a joint lawsuit against Google. They allege that Google used copyrighted works without permission to train its Gemini AI platform.
As generative AI technology advances rapidly, balancing model training requirements with copyright protection has emerged as a critical legal and business challenge for the global AI industry. Although the Anthropic settlement offers one resolution path, long-term guidelines for AI data compliance remain unclear and require further clarification.
Related article
OpenAI Launches Open-Source Codex Security CLI to Uncover Vulnerabilities in CI
On July 29, OpenAI introduced a specialized security scanning tool for code repositories. Named Codex Security CLI, this open-source command-line utility is designed to help developers scan codebases, detect and verify security vulnerabilities, and s
WeChat WeMM-Embedding Powers Moments Search and Video Highlights With 1 Billion Daily Calls
On September 4, WeChat AI announced the open-source release of its general multimodal embedding model, WeMM-Embedding. Available in 2B, 4B, and 9B versions, the model supports text, images, videos, visual documents, and any interleaved multimodal inp
Xiaohongshu, Peking University, Shanghai Jiao Tong University Unveil HYPIC to Slash First Token Latency by 3.25x in Hybrid Attention Models
The primary focus of large model services is transitioning from single-turn conversations to retrieval-augmented question answering, multi-document summarization, and long-range agents. These emerging workloads share a key trait: a single request's p
Related Special Topic Recommendations
Comments (0)
0/500
On Monday, a U.S. federal court granted final approval for a $1.5 billion class-action settlement between Anthropic and a coalition of authors and publishers, as reported by Reuters. The deal resolves ongoing litigation regarding Anthropic’s alleged copyright infringement during the training of its AI models, with the company now obligated to begin disbursing compensation to the affected copyright holders.
U.S. District Judge Araceli Martinez-Olguin of the Northern District of California signed the final order ratifying the agreement. This follows a previous provisional approval by retired Judge William Alsup, who had previously determined that Anthropic had unlawfully downloaded and stored millions of copyrighted books.
Under the approved compensation framework, each affected work will receive $3,000 in damages. This figure applies to approximately 500,000 works, with the payouts distributed among the authors and publishers holding the relevant copyrights. The total sum ranks among the largest settlements in the history of U.S. copyright law.

Despite the settlement, legal uncertainties surrounding AI training data persist. Judge Alsup had previously sided with Anthropic on the core legal question, suggesting that using copyrighted texts to train AI models might qualify as "fair use"—a significant potential precedent in AI copyright disputes. However, he also noted that certain methods Anthropic used to acquire this data were illegal.
Anthropic acquired its training data through two primary channels: legally purchased and scanned books, and content scraped from pirate sites like Library Genesis and Pirate Library Mirror. The court ruled the latter approach illegal. To avoid prolonged litigation and the risk of substantial penalties, Anthropic opted for a settlement.
Industry analysts observe that while this agreement closes this specific case, it fails to establish a unified legal standard for the broader AI sector. Because Anthropic settled rather than appealing, Judge Alsup’s prior comments on "fair use" in AI training will not become binding judicial precedent.
Lawsuits concerning the copyright of AI training data remain active, with cases pending against major tech firms including Google, Meta, Midjourney, and OpenAI. Recently, publishers such as Hachette, Cengage, and Elsevier, along with author Scott Turow and the organization SCRIBE, filed a joint lawsuit against Google. They allege that Google used copyrighted works without permission to train its Gemini AI platform.
As generative AI technology advances rapidly, balancing model training requirements with copyright protection has emerged as a critical legal and business challenge for the global AI industry. Although the Anthropic settlement offers one resolution path, long-term guidelines for AI data compliance remain unclear and require further clarification.
OpenAI Launches Open-Source Codex Security CLI to Uncover Vulnerabilities in CI
On July 29, OpenAI introduced a specialized security scanning tool for code repositories. Named Codex Security CLI, this open-source command-line utility is designed to help developers scan codebases, detect and verify security vulnerabilities, and s
WeChat WeMM-Embedding Powers Moments Search and Video Highlights With 1 Billion Daily Calls
On September 4, WeChat AI announced the open-source release of its general multimodal embedding model, WeMM-Embedding. Available in 2B, 4B, and 9B versions, the model supports text, images, videos, visual documents, and any interleaved multimodal inp
Xiaohongshu, Peking University, Shanghai Jiao Tong University Unveil HYPIC to Slash First Token Latency by 3.25x in Hybrid Attention Models
The primary focus of large model services is transitioning from single-turn conversations to retrieval-augmented question answering, multi-document summarization, and long-range agents. These emerging workloads share a key trait: a single request's p





Home






