On September 29, President Trump hosted some twenty tech executives for a White House lunch and signed the one-page “White House Accord on Super Intelligence: Joint Commitment on Frontier Responsibilities.”

The signatories: Google’s Sundar Pichai, Anthropic’s Dario Amodei, Meta’s Mark Zuckerberg, OpenAI’s Greg Brockman, Nvidia’s Jensen Huang, and SpaceX’s Elon Musk. (Sam Altman was notably absent from the room, per some coverage.)

The accord commits the companies to four layers of controls: robust internal safety processes, internal safety-team review, independent external auditors, and board-level oversight committees. Trump called it “morally binding” — “almost like a constitution.”

A companion executive order directs federal agencies to rebrand AI as “Super Intelligence” / “SI” in public communications, a move announced at the UN General Assembly the week before. (The signed document also went viral for misspelling “Unites States” under Trump’s signature.)

Critics call it self-policing with no enforcer: no audit frequency, no thresholds, no penalties. Bill Gates told NBC, “No one thinks self-regulation is enough.” And on September 30, the FTC opened an investigation into AI agent risks at OpenAI, Anthropic and others — details still unconfirmed, but the accord’s signed commitments now serve as a benchmark that investigation can measure against.

Why it matters

This is the clearest articulation yet of the US federal government’s AI stance: keep industry self-governance while the China race dominates, with real regulation parked as future possibility. Even as a non-binding pledge, the accord creates a documented record of each company’s safety practices — immediately usable by the FTC, state attorneys general, and plaintiffs. It also pulled Anthropic’s Amodei, previously the industry’s loudest slowdown advocate, under a “superintelligence” banner: the safety debate has been absorbed into the race framing.

Good reads