SPEAKER_00 0:00

This is AI Frontiers. Today, the creativity debate was always the wrong fight. For three years, the creative industries argued about whether AI can truly create. In June 2025, two federal courtrooms in California answered a different question. And it turned out to be the one that actually matters. Within two days of each other, two federal district judges ruled that training large language models on copyrighted books was fair use. Judge William Alsop called anthropics training quintessentially transformative. In the parallel case against Meta, Judge Chabria reached the same conclusion on training, though narrowly warning the ruling did not make Meta's practices lawful in general. Both judges declined to treat a machine reading a book differently from a person reading one. That is the analogy AI companies had been selling for years. The courts picked it up and used it against the artists. But here is where the story turns. Anthropic still agreed to pay around $1.5 billion, roughly $3,100 per book, across 482,000 titles. Not for learning, for piracy, for downloading and hoarding a library of stolen copies. The court blessed the training and condemned the pirated library. Judge Al Soup's own words. Pirating copies to build a research library without paying and retaining them for future use was its own use and not a transformative one. That is the whole game. The line these rulings drew was between material. You came by honestly and material you took, and between output that adds something new and output that displaces the original. Provenance and substitution. Everything else was theater. America and Europe have landed in the same place by different roads. America litigates case by case. Europe regulates in advance. Since August 2025, the European Union's Artificial Intelligence Act requires model providers to publish training data summaries, maintain a copyright policy, and respect opt-out rights under EU text and data mining law. Enforcement powers, fines included, switch on in August 2026. The creativity question is a decoy. It is unanswerable, and it is irrelevant to every practical decision a working person has to make. The questions that survive contact with reality are narrower. Where did the training data come from? Did the people who made it get to say no? Does the output displace what it learned from or stand beside it? Stop asking whether the machine is creative. It is the wrong question, and no answer to it changes what you owe. That's AI Frontiers.