Six AI firms sign morally binding White House accord on superintelligence

Six leading AI companies signed the White House Accord on Superintelligence on 29 September 2026, an agreement President Donald Trump called morally binding. Meta, Google, OpenAI, Anthropic, Nvidia, and xAI signed the document, subtitled the Joint Commitment on Frontier Responsibilities, which Trump posted to his Truth Social account; it commits the firms to four layers of controls but carries no enforcement mechanisms or legal implications.rollcall.com+1

The accord arrives as voluntary self-policing at a moment when AI firms have faced scrutiny over incidents in which major companies inadvertently hacked into outside organisations during safety tests. Trump has broadly rejected calls for more regulation of the technology, citing competition with China and threats to economic growth. The document leaves open the possibility that it may one day make sense to codify these steps into laws or regulations.

What the accord commits companies to

The accord outlines four layers of controls and audits that the White House believes signatories should implement in addition to any other precautions. Companies must implement robust internal controls to monitor model capabilities and alignment during training and deployment around cybersecurity, biosecurity, and chemical threats. The controls must also ensure models do not hack or access technical systems in unintended ways. The document offers no definition of what constitutes robust internal controls.

The second layer empowers an internal team to ensure all controls, monitoring, and detection operate as intended and that any issues are remediated. The third requires partnering with an independent external auditor or evaluator to carry out independent assessments of whether the controls are working. The fourth designates an independent committee of the board of directors to oversee and receive reports from the teams operating the controls and the internal and external auditors. The document specifies no frequency for external audits or evaluations.

The document states that participating companies will meet regularly to establish standards and best practices to improve the safety of their systems. Trump said he would name someone to oversee the agreement in coming days after consulting with industry. The agreement appears to allow companies to pick their own evaluators, appoint their own oversight boards, and decide whether to publish their results.

Oversight left to the companies

None of the controls listed in the agreement involve government regulators or include a commitment to publicly detail the findings of independent evaluations. Trump announced the deal following a White House luncheon with top tech executives including Elon Musk, Mark Zuckerberg, Jensen Huang, and Dario Amodei. Trump said the agreement was almost like a constitution and that the biggest people in the world signed it. Zuckerberg said it was a start and an accord that the whole industry could come to.

Amodei, chief executive of Anthropic, said the technology has very real risks and that the mechanism for addressing those risks is still under discussion. Asked why tech leaders should be trusted to self-regulate, Trump replied that they are outstanding people. He also claimed he was seeing tremendous self-policing and that the companies understand they have to self-police. Trump said the accord would be morally binding and would allow tremendous self-policing of products and technologies.

On the same day, Speaker Mike Johnson pushed back on calls to impose regulations on artificial intelligence, pledging that lawmakers would continue to assess and further deliberate in the days ahead without committing to legislation. In the Senate, Chairman Ted Cruz objected when Senator Mark Warner asked for unanimous consent to pass an unnumbered bill establishing an AI Safety Board within the Commerce Department. The bill would have required frontier developers to give the board access to new models at least 45 days before release.

Congress takes a back seat

The bill would also have required frontier developers to report safety or security incidents within 30 days, or within 72 hours if they pose an imminent threat to national security. Warner said his request was really simple and had been as pared down as possible, adding that he had not talked to a single member on either side of the aisle who did not think Congress needs to move. Senator Brian Schatz questioned whether companies rolling toward the biggest IPOs in human history would voluntarily regulate themselves.

Cruz said he is continuing to work on an AI catastrophic risk bill with Senate Majority Leader John Thune and Senator Amy Klobuchar, and said he shares concerns about warnings on AI safety. Senator Thom Tillis said he doubted AI regulation could be solved in one meeting, describing the gathering as getting the process started. Schatz said that if a catastrophic event related to AI occurs before the election, he thinks the Senate should return to Washington to deal with it.

Also on 29 September 2026, Trump signed an executive order requiring federal agencies to use the term Super Intelligence and the acronym SI instead of Artificial Intelligence and AI in official correspondence, public communications, websites, reports, and policy documents. Trump tasked his Assistant for Science and Technology to work with other agencies to develop a definition within 60 days, as no working definition of Super Intelligence yet exists. Trump said he will never stifle the growth of a technology that will be bigger than the industrial revolution.