White House Accord on Super Intelligence Signed by Trump and Six AI Giants

Voluntary self-policing pacts are not new: in February 1975 the Asilomar Conference produced informal guidelines that steered recombinant-DNA research for years before Congress codified bio-safety rules; in 2009 the tech industry’s Global Network Initiative tried a similar ‘code of conduct’ approach on surveillance, with mixed results. The new accord fits that lineage of frontier technologies briefly governed by elite consensus before formal law catches up. It also echoes earlier moments—such as the 1921 Washington Naval Conference or the 1997 Kyoto Protocol—where the U.S. balanced international rivalry (now China) against domestic industrial advantage, choosing minimal constraints to keep its lead. Whether this matters a century from now hinges on two structural trends: (1) the growing transfer of quasi-regulatory power from elected bodies to a handful of private platforms who own the compute and data, and (2) the recurrent American preference for market-led innovation until a crisis forces statutory intervention. If future accidents or AI-driven inequities prove rare, this 2026 pact may be remembered like the early internet’s RFC 1087 “Ethics and the Internet”—a footnote. If, however, a major SI failure occurs, historians may cite this moment as the missed chance when voluntary guardrails replaced enforceable law in the name of beating China.