Six major AI labs sign voluntary White House 'Superintelligence Accord' with no enforcement teeth

nordicinst · x · 2026-10-01

Executives from Google, Anthropic, Meta, OpenAI, xAI, and Nvidia signed the "White House Accord on Super Intelligence" after a luncheon hosted by President Trump. The voluntary pact asks companies to: maintain robust internal controls to monitor model capabilities and prevent rogue behavior; empower an internal team to do that work and remediate failures; partner with an empowered external monitor for independent assessments; and set up a board committee to receive reports.

WIRED calls it a "fancy pinky-swear": breaking the public commitments would at most count as illegally lying to consumers and wouldn't trigger major reforms, and the sweeping regulation some inside the labs wanted didn't materialize. Meanwhile the FTC is preparing to investigate several AI companies over potential issues.

Related event: Six Major AI Firms Sign Voluntary White House Superintelligence Accord(14 posts)→

Original post →

More from Safety

Safety channel →