Trump, AI Executives Sign Voluntary Pact on AI Oversight
President Donald Trump met with top artificial intelligence executives at the White House on Sept. 29, 2026, and said they signed a voluntary accord to self-police AI development. NPR
House Speaker Mike Johnson joined Trump for the meeting with executives from top tech firms investing in AI infrastructure. Johnson, a Republican from Louisiana, was present as Trump described the outcome to reporters. Reuters reported Sept. 29 that the AI sector had signed an accord on technology standards, according to the House speaker. Reuters
The signers were Trump, Anthropic CEO Dario Amodei, Google CEO Sundar Pichai, Meta CEO Mark Zuckerberg, OpenAI President Greg Brockman, Nvidia CEO Jensen Huang, and Elon Musk. Trump posted the text of the accord on Truth Social late Tuesday, Sept. 29. PBS NewsHour reported that the voluntary accord, signed during the White House meeting, included internal and external reviews. PBS
Under the accord, companies would set up "robust internal controls." That means safety rules and checks inside each company, like a factory safety checklist. They would also hire an "independent external auditor," an outside reviewer, to assess if those controls work. Each company would form a committee within its board of directors to review reports from both sets of auditors. The agreement is voluntary. It is not a law. Axios reported Sept. 29 that Trump said top AI executives agreed to voluntary standards aimed at increasing industry oversight. Axios
The accord states, "Over time, it may make sense to codify these steps into laws and regulations." Codify means to write them into law. The White House Briefing Room lists a fact sheet titled 'President Donald J. Trump Inaugurates The Era of Super Intelligence' dated Sept. 29, 2026, and a presidential action titled 'Inaugurating The Era Of Super Intelligence' dated Sept. 29, 2026. White House Briefing Room
Trump told reporters he was seeing "tremendous self-policing" by AI companies. He said they understand they have to self-police. He said about 10 people would be named to a committee to "watch over the whole enterprise." He said the AI accord would be "morally binding." He said he would name someone to oversee the agreement in coming days after consulting with industry.
The broader context here is that the structure is self-regulatory, not statutory. Board committees, internal controls and outside review create a written record for oversight. They do not create enforceable legal duties by themselves. The language on future laws leaves Congress and agencies as the next decision point. Key questions are what counts as a working control, who qualifies as independent, what standards auditors use, and how boards handle dissenting findings.
In my view, the near-term stakes are less about technical compliance and more about leverage. A signed, public text gives the White House a reference point in future talks with the companies. It also gives lawmakers a draft framework to adopt, amend or reject. The test will be whether auditor reports and board reviews are disclosed so outsiders can verify them, and whether the planned oversight lead sets timelines and public expectations.
