Tech billionaire and venture capitalist Chamath Palihapitiya has spoken on AI intellectual property debate following allegations from top US labs OpenAI and Anthroractice of copying the capabilities of a rival’s expensive AI model at a fraction of the cost. Pushing back against the complaints of the US frontier labs, Palihapitiya argues that tech giants cannot complain about competitors copying their systems when their own foundational AI models were built by scraping copyrighted human data from the open internet without permission.
What US frontier AI labs claim
Palihapitiya’s comment came soon after the economic and moral defense of distillation was highlighted in an analysis of Anthropic’s own Fable model. The company blamed Chinese AI companies to have used distillation to train their models, thereby, calling it ‘illegal’.Major AI labs have spent billions of dollars on computing power, data curation and human feedback to build their models. Yet, to gather their initial training datasets, they scraped millions of copyrighted books, news articles, and proprietary code from the public internet without compensating the original human creators. Their legal defense has rested on “fair use”.Palihapitiya points out in a long post on X (formerly Twitter) that distillation applies this exact same logic, only turned back onto the tech giants themselves. In distillation, a competitor’s AI is simply “learning” from the output data of another AI. He also called the sudden anxiety gripping Silicon Valley driven by economics rather than ethics. Distillation ‘allowed’ Chinese companies to effectively free-ride on the astronomical research and development costs borne by US pioneers.
What Chamath Palihapitiya said
From Anthropic’s Fable model on the economic, moral, ethical and legal opinion of distillation of Anthropic’s Fable model:Whether it’s a moral problem is genuinely contested. The labs trained their models on the open internet — copyrighted books, articles, code — largely without permission, and their fair-use defense is essentially “learning from data is transformative.” Distillation is the same argument turned against them: a model learning from another model’s outputs.It’s hard to construct a moral principle that permits the first and forbids the second, which is why critics call the labs’ objections hypocritical rather than principled.Distillation is contentious because it lets a smaller model absorb much of a frontier model’s capability by training on its outputs — effectively free-riding on billions of dollars of compute, data curation, and RLHF work.The DeepSeek episode made this concrete: if you can extract 80% of the value of a $1B training run for $5M by querying the API, the economics of frontier labs get shaky.That’s the core anxiety — it’s a moat problem before it’s anything else.Legally, it’s mostly a contract issue, not a copyright one. Model outputs likely aren’t copyrightable (no human author), so the labs’ real weapon is terms of service — every major API prohibits using outputs to train competing models. But ToS violations are breach of contract, hard to detect, hard to prove, and nearly unenforceable against a foreign entity. There’s no statute against distillation itself. So the practical answer: it’s an economic problem dressed in legal clothing, with a moral argument that cuts both ways depending on whose training data you start counting from.
