A wave of court decisions expected in 2026 could redefine how U.S. copyright law treats artificial intelligence, as judges confront whether tech companies can lawfully use copyrighted material to train AI systems. After years of legal uncertainty, recent lawsuits and settlements have pushed the issue toward a moment of reckoning for both creators and the technology industry.
At the heart of the disputes is the fair use doctrine, which allows limited use of copyrighted works without permission under certain conditions. Companies developing generative AI argue that training models on large datasets transforms existing works into something new. Copyright holders counter that the practice amounts to mass copying that threatens creative markets and demands compensation.
The legal battle intensified throughout 2025. Major media and entertainment companies, including The New York Times and Disney, filed new lawsuits. At the same time, authors reached a $1.5 billion class action settlement with Anthropic, the largest known copyright payout in U.S. history. Those developments signaled that courts would soon have to confront the issue directly.
Judges offer conflicting views on fair use and AI
Federal judges have now begun issuing substantive rulings, and their views diverge sharply. In June, U.S. District Judge William Alsup in San Francisco ruled that AI training can qualify as transformative, a key factor in fair use analysis. He wrote that copyright law exists to promote new creative output, not to shield authors from competition.
.6
However, Alsup also found that Anthropic violated copyright law by storing millions of pirated books in a separate internal library unrelated to training. That finding exposed the company to enormous potential liability before it reached a settlement later in the year.
Only days later, another San Francisco judge, Vince Chhabria, ruled in favor of Meta but warned that fair use would not apply in all AI training scenarios. He raised concerns that generative AI could overwhelm creative markets with new content, undermining the incentives copyright law seeks to protect.
The contrast between the rulings highlights the uncertainty ahead. Alsup dismissed fears of market harm, comparing them to objections against teaching people to write well. Chhabria, by contrast, viewed generative AI as a possible threat to creative industries.
More hearings are scheduled for 2026. Courts are set to consider disputes involving Anthropic and music publishers, Google and visual artists, Stability AI, and AI-driven music platforms. New rulings could bring clarity or deepen existing divisions, shaping whether AI companies rely on fair use or move toward widespread licensing.
Some companies have already chosen negotiation over litigation. Several copyright owners have entered licensing agreements with AI developers. Disney, for example, agreed to invest $1 billion in OpenAI and allow its characters to be used in AI-generated video. Warner Music also settled lawsuits against AI music startups and plans joint platforms in 2026.
As judges weigh these cases, their decisions could reshape the legal and economic foundations of generative AI. For both copyright holders and technology firms, the coming year may determine how innovation and intellectual property coexist in the digital age.
