AI Copyright Battles Enter Pivotal Year as US Courts Weigh ‘Fair Use’

The sprawling legal battle over technology companies’ large-scale use of copyrighted material to train artificial intelligence systems could be entering a decisive phase in 2026.

After a series of new lawsuits and a landmark settlement in 2025, the year ahead is expected to bring a wave of court rulings that may define how US copyright law applies to generative AI. At stake is whether companies such as OpenAI, Google and Meta can rely on the legal doctrine of fair use to shield themselves from liability, or whether they will be required to compensate copyright holders -- potentially at a cost running into billions of dollars.

The conflict intensified sharply last year. The New York Times, Disney and other major copyright owners filed fresh lawsuits, while authors secured a $1.5 billion class action settlement with Anthropic, the largest known copyright payout in US history.

For the first time, federal judges also began ruling on whether training generative AI models qualifies as fair use, which permits limited, unauthorised use of copyrighted material in certain circumstances. Early decisions were mixed, highlighting the uncertainty facing both rights holders and the technology industry.

Divided rulings

Defendants in most cases have argued that their AI systems make fair use of copyrighted material by transforming it into something new.

In June, US District Judge William Alsup in San Francisco described AI training as “quintessentially transformative”, siding with the company on a key fair-use factor. Copyright law, he wrote, “seeks to advance original works of authorship, not to protect authors against competition”.

However, Alsup also held the company liable for storing millions of pirated books in a “central library” unrelated to training -- a finding that exposed Anthropic to potential liability of up to a trillion dollars before the case was settled in December.

Two days later, Judge Vince Chhabria, also sitting in San Francisco, ruled in favour of Meta in a similar case but warned that AI training would “in many circumstances” not qualify as fair use. He expressed concern that generative AI could “flood the market” with content, undermining incentives for human creators -- a core objective of copyright law.

While Alsup dismissed fears of market harm, likening them to complaints that “training schoolchildren to write well” creates competition, Chhabria viewed generative AI as a potential existential threat to creative markets.

What’s next

 

More hearings are scheduled, or are expected to take place, in 2026 in disputes involving Anthropic and music publishers, Google and visual artists, Stability AI, and AI music generators. New rulings could clarify how fair use applies to AI or further deepen the uncertainty, while helping to determine whether AI companies will enjoy broad fair-use protections or face a licensing regime that reshapes the industry’s economics.

In parallel, some major copyright owners have entered into licensing agreements with AI-focused technology firms, signalling a more collaborative approach. Alongside Anthropic’s class action settlement, Disney in December agreed to invest $1 billion in OpenAI and to allow the use of Disney characters in its Sora AI video generator. Warner Music also settled lawsuits against AI music companies Suno and Udio, agreeing to launch joint music-creation platforms with them in 2026.

Thomson Reuters, the parent company of Reuters News, licensed Reuters content to Meta for use in its AI systems in 2024. It is also involved in an ongoing copyright dispute with former legal research competitor Ross Intelligence over the alleged misuse of Westlaw material in AI training.

 

For any enquiries or information, contact ask@tlr.ae or call us on +971 52 644 3004. Follow The Law Reporters on WhatsApp Channels.