Antitrust experts warn AI slowdown proposals could trigger investigations

Meta CEO Mark Zuckerberg, who recently avoided a massive FTC suit, has pushed back on the explicit “slowdown” language, arguing that AI firms already have a strong market incentive to improve safety because misaligned models would hurt sales. He likened the debate to automakers choosing to delay faster cars until they can make them safe. Meanwhile, Anthropic and OpenAI – both valued near a trillion dollars and preparing for IPOs – have filed confidential paperwork, fueling speculation that they seek an antitrust exemption to coordinate safety work.

If regulators treat the proposal as a conduct investigation, the labs could face years‑long probes, massive document requests, and costly depositions, even if no case is ultimately filed. The lack of clear regulatory guidance means the companies may have to craft their own safety standards while navigating a potentially hostile antitrust environment.