OpenAI and Anthropic Join Forces — To Make Sure You Can't Use Free Open-Source AI
Rivals in a rare alliance, critics call it straight: 'regulatory capture'
OpenAI and Anthropic compete fiercely for customers. But in Washington, the two leading AI labs have found something they agree on: warning policymakers about Chinese open-weight AI models. The catch? Critics, including Trump adviser David Sacks, say this amounts to 'regulatory capture' — rules meant to improve AI safety could instead entrench the largest companies by making it harder for competitors to release models.
Anthropic CEO Dario Amodei argues that open-weight models are harder to keep safe because, once released, developers lose the ability to revoke access or update safety guardrails. Open-model supporters say that's exactly the point — no single company should control who uses AI or how. Meanwhile, over 1,000 startup founders and researchers urged the White House not to shut off Chinese open-weight AI. Former Biden adviser Suresh Venkatasubramanian said losing access to Chinese models would be a 'big problem for doing research.'
On the other side, U.S. Trade Representative Jamieson Greer has called Chinese distillation a form of IP theft. Treasury Secretary Scott Bessent said Chinese models must meet the same standards as American ones. Sam Altman met with lawmakers and the White House the same week to discuss upcoming model releases. And on a state level, OpenAI endorsed a weaker Massachusetts AI safety bill than the version Anthropic backed — the stronger version required independent testing every six months and public reports; OpenAI's version didn't include those provisions.
It's a neat arrangement: two competitors tell the government that open alternatives are too dangerous, and the rules that follow make it harder for anyone without a billion-dollar lobbying budget to compete. BenPig thinks this feels like the two biggest amusement parks asking the city to shut down the free swings — for safety, of course.