According to Reuters, the pivotal legal fight over AI companies using copyrighted material to train their systems is entering a decisive phase in 2026. This follows a major escalation in 2025, which saw new lawsuits from The New York Times and Disney, and a landmark $1.5 billion class action settlement between authors and Anthropic. Federal judges have begun issuing mixed rulings on whether this training qualifies as “fair use,” with one judge calling it “quintessentially transformative” while another warned it often would not qualify. The stakes are enormous, potentially costing companies like OpenAI, Google, and Meta billions if they lose fair use protection. In the meantime, some major players like Disney are cutting deals, investing $1 billion in OpenAI and licensing characters for its Sora video generator.
The Fair Use Fight Is All Over The Place
Here’s the thing: the courts are completely split, and that’s creating chaos. Judge Alsup in San Francisco basically said training AI is the perfect example of “transformative” fair use, which is a huge win for the tech side. He dismissed fears of market harm with a wild analogy about training schoolchildren to write. But then, just two days later, Judge Chhabria in the same city ruled for Meta but dropped a massive warning. He’s worried AI could “flood the market” and destroy the economic incentive for human creators—which is the whole point of copyright law! So which is it? Is AI training a brilliant new use of information, or an existential threat to creative industries? The judges can’t even agree, and that uncertainty is a nightmare for everyone trying to build a business here.
Billions Are On The Line, And Paths Are Diverging
Look, the financial reality is stark. If fair use doesn’t shield them, tech companies face a licensing regime that could reshape the entire economics of AI. We’re talking about retraining massive models from scratch with properly licensed data—a cost that could run into the trillions, not just billions. That’s why Anthropic settled for $1.5 billion; it was a bargain compared to the potential liability. But now we’re seeing two paths emerge. One is the courtroom battle, which is heading for more crucial rulings in 2026 involving music publishers, visual artists, and others. The other path is the deal-making table. Disney’s billion-dollar bet on OpenAI and the Warner Music settlements with Suno and Udio show that some giants would rather partner than fight. It’s a classic case of “if you can’t beat ’em, join ’em,” or maybe more accurately, “if you can’t sue ’em, invest in ’em.”
So What Happens Now?
2026 is set to be the year we get some real answers, or maybe just more confusing questions. More hearings and rulings are coming, and they’ll either clarify this fair use mess or deepen the divide. I think the pressure will force more settlements and licensing deals, even among the current litigants. The cost and risk of a total loss in court are just too high for both sides. But fundamentally, this isn’t just a legal question. It’s about what we want the future of creativity to look like. Do we treat the entire internet as a free training dataset for a few powerful companies? Or do we build a system where creators get a slice of the AI pie? The courts are trying to apply a 20th-century doctrine to a 21st-century technology, and it’s getting messy. One thing’s for sure: the outcome will determine who profits from the AI revolution for decades to come.
