Trump signs a voluntary AI safety accord with six tech chiefs

President Donald Trump and the leaders of six AI companies signed a voluntary accord at the White House on September 29, 2026 that commits their firms to internal controls and outside audits of frontier models. The document carries no legal force, and it says nothing about deepfakes, synthetic performers or AI video.

Trump posted the text on Truth Social late on Tuesday, the Associated Press reports via NPR. The accord, titled "White House Accord on Super Intelligence: Joint Commitment on Frontier Responsibilities", is signed by Trump alongside Google chief executive Sundar Pichai, Anthropic chief executive Dario Amodei, Meta chief executive Mark Zuckerberg, OpenAI president Greg Brockman, Nvidia chief executive Jensen Huang and Elon Musk, whose xAI is now part of SpaceX.

The commitments have four layers, technology.org reports: internal controls that track what a model can do during training and deployment, naming cybersecurity, biosecurity and chemical risks, and asking that models "do not hack or access technical systems in unintended ways"; an internal team that confirms the controls work; an independent external auditor or evaluator assessing the same controls from outside; and a committee of each company's board that receives the audit reports and sees that flagged issues are fixed. The signatories also agree to meet regularly on standards.

Asked whether the document binds anyone, Trump told reporters it is "morally binding". "I think I'm seeing tremendous self-policing, and they understand that they have to self-police," he said, per the Associated Press. House Speaker Mike Johnson described the accord as voluntary. The document itself keeps the door open: "Over time, it may make sense to codify these steps into laws or regulations."

Two more pieces came with it. Trump said he is considering a committee of about 10 people to watch over the effort and that he would name an official to lead White House AI policy, per Reuters' account carried by the Cyprus Mail. A separate executive order directs federal agencies to use "Super Intelligence" and "SI" in official communications in place of "artificial intelligence" and "AI", per Economy Middle East.

The reception split along the usual line. Senator Mark Warner said Congress should pass mandatory testing, evaluation and incident reporting for the most advanced models, in his statement of September 29: "The companies building the most powerful AI systems are warning us that the technology is advancing faster than our safeguards," he said. "The president's response? To rename it and tell the companies developing it to regulate themselves." Senator Elizabeth Warren made the self-regulation argument in the same terms, per Crypto Briefing. Voluntary pledges have a history here: the Biden administration gathered similar voluntary commitments from seven AI companies in July 2023, including Amazon, Anthropic, Google, Inflection, Meta, Microsoft and OpenAI, according to the White House fact sheet of July 21, 2023.

For creators, the gap is the story. Nothing in the accord, and nothing in federal law, requires a disclosure when an AI-generated performer appears in an advertisement or when a synthetic voice is used in a video. The binding rules that do touch that work are coming from states: California's SB 1050, signed on September 16, 2026, requires a clear disclosure when a video or audio advertisement uses an AI-generated performer, and SB 1111, which the Assembly passed on August 28, would treat impersonation through a digital replica as false impersonation. The rest of California's 2026 AI package sat under a September 30 sign-or-veto deadline, per Transparency Coalition's September 25 update.

The accord says its four steps may one day be codified into law; nothing in it binds a company today.