As AI CEOs clash over regulation, a new culture war is brewing

The fight over AI regulation has stopped being a technical argument and started looking like a political identity test. In just a few days, the debate over how to control frontier AI has widened from safety circles into a broader clash over power, who gets to set the rules, and which companies get to benefit from them.

On one side are Anthropic’s Dario Amodei and OpenAI’s Sam Altman, who are unusually close on this one. Amodei wants frontier labs to slow down when needed, allow independent evaluators, and line up around shared safety standards. Altman has backed that broad direction and said OpenAI will let independent evaluators operate inside the company.

On the other side are Meta’s Mark Zuckerberg and Nvidia’s Jensen Huang, both arguing that the industry does not need fresh regulation. Huang said safety and speed can coexist. Zuckerberg went further, saying market pressure, legal liability, and customer distrust will already push companies toward safer systems. He even pointed to Meta delaying its Muse agent for months so it could focus on security and safety.

That pushback is not limited to the biggest names. Cohere warned that a safety regime built around a small cluster of dominant Silicon Valley firms could look “a cartel by any other name,” especially if those firms got an antitrust exemption to coordinate on rules everyone else would have to follow. Cohere’s Aidan Gomez wants evidence-based risk assessments, mandatory transparency, independent testing tied to actual capabilities, and oversight that is not riddled with conflicts of interest. Cohere chief AI officer Joelle Pineau agreed that regulation is necessary, but said the companies building frontier systems should not be the only ones writing the rules.

The politics around all this are getting uglier by the week. The Wall Street Journal reported that Zuckerberg, Huang and Elon Musk separately lobbied Donald Trump against a proposed industry-funded AI oversight body modeled partly on FINRA. Trump, meanwhile, has brushed off calls for a slowdown, treated regulation as a threat to U.S. competition with China, and said the country’s main safeguard is a “high IQ president.” Around the same time, the New York Post and the Department of War were both taking shots at effective altruism, which has become closely tied to AI safety. That’s the problem now: the argument is no longer just about testing models. It’s about which faction gets to define reality first.