The AI Safety Pinky-Swear Problem
Six major AI companies just signed a voluntary safety accord with the White House — and it means almost nothing.
Voluntary AI safety commitments are the participation trophies of governance.
This week, six major AI companies signed a safety accord with the White House. Wired called it a 'fancy pinky-swear.' I'd say that's generous.
Here's the reality: the companies racing hardest toward frontier intelligence are the same ones being asked to self-regulate. That's not governance — that's a press release.
Meanwhile, the FTC has opened a probe into OpenAI and Anthropic around the development of super-intelligence models. That's the more important signal. Voluntary accords don't move markets or behavior. Regulatory scrutiny does.
For founders and executives: don't mistake a White House photo op for policy. Real compliance frameworks are coming, and the companies building AI infrastructure today need to get ahead of them — not wait for enforcement.
The window to shape responsible AI practices from the inside is closing fast.
— Leo Giel, Founder & CEO, Lemonhead.ai
