More than a year after authors first sued AI company Anthropic over its use of copyrighted books, one of the biggest copyright disputes in U.S. history has officially come to a close.
On July 20, 2026, a federal judge granted final approval for a historic $1.5 billion class action settlement against Anthropic, the company behind the Claude AI models. This landmark case brought the growing legal friction between generative AI development and creative rights to a head, forcing the courts to answer a critical question: Can AI companies legally use copyrighted books to train their models without permission?
Case Background
In August 2024, authors Andrea Bartz, Charles Graeber, and Kirk Wallace Johnson filed a class action lawsuit on behalf of writers nationwide. They alleged that Anthropic downloaded millions of copyrighted books from illicit online shadow libraries like Library Genesis (LibGen) and Pirate Library Mirror (PiLiMi) to build its AI training datasets rather than licensing or purchasing the works legitimately.
Before the case could reach a jury, Judge William Alsup issued a decisive summary judgment ruling that fundamentally redefined the playing field:
- Training on Lawful Data: Using lawfully acquired books to train an AI model is generally protected as fair use under U.S. copyright law due to its transformative nature.
- Pirated Data Acquisition: Downloading, storing, and maintaining a permanent internal library of pirated books is not fair use.
Anthropic’s reliance on pirated shadow libraries crossed a clear legal line, creating staggering liability under federal copyright law.
The Settlement and Immediate Impact
Faced with potential statutory damages that could have spiraled past $70 billion, Anthropic agreed to a $1.5 billion settlement, which has now received final judicial approval.
The agreement covers nearly 500,000 copyrighted works, yielding an estimated payout of $3,000 to $3,100 per qualifying book. By default, settlement proceeds are split 50/50 between authors and publishers, unless their specific publishing contracts dictate a different breakdown. Beyond financial compensation, Anthropic is required to destroy its illicitly downloaded digital files. Crucially, the settlement does not grant Anthropic a permanent, forward-looking license to continue training on those works.
The Unresolved Future: Looking to the Music Industry Model
While a $1.5 billion recovery is a major victory, a one-time monetary settlement leaves the bigger structural problem unsolved. A lump-sum payout resolves past illegal harvesting, but it does not account for the ongoing competition AI models generate against human creators once trained.
To create a sustainable future, the creative community is increasingly looking to proven frameworks in other sectors, particularly music licensing.
For decades, the music industry has relied on Performing Rights Organizations (PROs) like ASCAP and BMI, as well as statutory entities like SoundExchange. These organizations track when music is played publicly or streamed digital radio, collect blanket licensing fees from commercial users, and distribute ongoing performance royalties directly to song creators and copyright holders.
Applying a similar blanket collective licensing or statutory royalty mechanism to generative AI could transform how creators are compensated:
- Ongoing Revenue Streams: Instead of chasing individual lawsuits or relying on rare class action settlements, creators would receive recurring micro-royalties based on model training and generation outputs.
- Collective Management: A centralized licensing body could streamline permissioning for tech companies while protecting individual artists from unfair bargaining leverage.
- Fair Commercial Standard: Tech platforms would gain certainty and legal compliance, while creatives would participate directly in the long-term revenue generated by AI tools built on their work.
What This Means for Creatives
Bartz v. Anthropic is a watershed moment, but it is only the first act. This ruling confirms that technology companies cannot bypass traditional intellectual property rules through illegal shortcuts. However, as AI tools mature, the focus will inevitably shift from compensating past data harvesting to establishing fair, standardized royalty models for the future.
Copyright law remains the most powerful tool creative entrepreneurs have to protect their work. As these legal and commercial frameworks evolve, staying informed about your copyright registrations and contractual rights is more important than ever.
Recent Comments