AI giants push self-regulation while Trump dismisses safety fears as hoaxes

The emerging split is striking: the biggest labs are publicly asking for a standards body that can test and rein in their own frontier models, while the White House is led by a president who calls AI risk a hoax and cheers Nvidia’s push to “run as fast as you can.” On one level, this is classic regulatory‑arbitrage theatre: a self‑regulator modeled on FINRA could give industry a hand on the steering wheel of frontier oversight. On another, it is a tacit admission from OpenAI, Anthropic and Google that they expect catastrophic‑risk narratives, whistleblowers and agent incidents to keep ramping, and want a credible forum to absorb that pressure.([amp.dw.com](https://amp.dw.com/en/ai-leaders-call-for-regulations-trump-dismisses-hoaxes/a-79284638))

For the AGI race, a firm, well‑funded standards body could be a double‑edged sword. If designed with real external power and transparency, it could slow reckless rollouts and force more robust evals before each new Astra‑class model ships. If captured, it could legitimize cartel‑like coordination on pace and access, disadvantaging smaller labs and open‑source efforts that cannot pay to play. Trump’s dismissal of AI dangers raises another risk: that US statutory regulation will lag far behind whatever voluntary regime the labs can stitch together, shifting the center of meaningful rule‑making to Brussels, London or even Seoul. The next 12–18 months will determine whether “pacing the frontier” becomes a concrete governance mechanism or just branding for continued maximal scaling.