The trillion dollar fight over who started AI

16 Sept 2026 · 1 min

Watch on YouTube

0:53

In this video segment, Dan Novaes explains that modern AI systems are trained on massive amounts of human‑created content—books, articles, photos, music, and code. Creators such as writers, artists, publishers, and other rights holders are suing AI companies, claiming that copyrighted works were used without permission. AI firms counter that model training may be legally protected in certain contexts. Novaes argues that the legal outcome could reshape the entire AI industry: if courts require broader licensing of training data, the cost of building powerful models could rise dramatically, while creators could gain a new revenue stream. He notes that licensing agreements between AI companies and major content owners are already emerging, underscoring that the dispute is not merely about ownership of a single image or article but about who gets paid when human creativity becomes part of AI training.

Chapters

AI learned from an enormous amount of human-created content— books, articles, photos, music, code. But now, creators are asking a simple question: should AI companies have been allowed to use all of it? Writers, artists, publishers, and other rights holders have brought lawsuits arguing that copyrighted work was used to train AI without permission. AI companies have pushed back, arguing that training models can be legally protected in certain circumstances. And the outcome could reshape the entire industry. Because if courts require companies to license much more of their training data, building powerful AI could become significantly more expensive. Publishers and creators could also gain a completely new source of revenue. We're already seeing AI companies sign licensing agreements with major content owners. So this fight isn't just about who owns an image or an article. It's about something much bigger. Who gets paid when human creativity becomes part of what teaches AI?