Trump’s AI accord misspells his title as six companies pledge safety controls
An accord signed by Donald Trump and technology executives sets out voluntary checks on advanced AI. A copy circulated by the White House also contains a spelling error beneath the president’s signature.
President Donald Trump and six technology executives signed a voluntary artificial intelligence safety accord in Washington on September 29, setting out controls and audits for advanced AI systems. A copy circulated by the White House identified him beneath his signature as ‘President of the Unites States’, The Guardian reported on September 30. The error drew attention to a document whose practical significance depends on what the companies do with its pledges.
Reuters reported that Trump released the one-page accord on Truth Social after meeting the executives. It calls for four layers of checks within participating companies, including outside assessment and oversight by a board committee. The commitments are voluntary: the document says the steps might eventually be written into laws or regulations, while the companies are committed to adopting them regardless of whether that happens.
What the White House AI accord asks companies to do
The first layer concerns the models themselves. According to Reuters’ account of the document, companies are asked to monitor AI capabilities and whether models behave as intended during training and deployment. The areas named include cybersecurity, biosecurity and chemical threats. The accord specifically calls for checks that models do not hack or access technical systems in unintended ways.
A second layer is an internal team charged with checking that those controls, along with monitoring and detection systems, operate as intended. The third asks each company to work with an external auditor or evaluator for an independent assessment of its AI controls and monitoring. The fourth calls for an independent committee of its board of directors to receive reports from the auditors and the teams operating the safeguards, and to oversee that work, Reuters reported.
The accord also says the participating companies will meet regularly to establish safety standards and best practices. Those are commitments to processes, rather than a reported finding that the safeguards have been installed or have reduced harm. Reuters’ account does not specify a deadline for putting each layer in place, how external evaluators would be selected, or what the companies would disclose publicly about their assessments.
Who signed the accord, and what the typo changes
The Guardian identified the six executives as Anthropic’s Dario Amodei, xAI’s Elon Musk, OpenAI’s Greg Brockman, Nvidia’s Jensen Huang, Meta’s Mark Zuckerberg and Google’s Sundar Pichai. Their signatures put some of the largest AI developers behind the same stated approach to monitoring and review. The accord, however, leaves the work of carrying out those commitments with the participating companies.
The Guardian said the misspelling appeared beneath Trump’s signature on the copy he posted to Truth Social and that the White House Communications Team emailed after the signing. The newspaper said it was unclear how the error occurred or whether it would appear in the Federal Register. It had asked the White House for comment. The reported typo concerns the document’s presentation; it does not, on the available accounts, establish a change to the safety commitments described by Reuters.
How California’s AI audit laws differ
California has taken a legislative approach to outside scrutiny of AI. Governor Gavin Newsom’s office said he signed Senate Bill 813 and Assembly Bill 1405 on September 9. According to the state, SB 813 establishes a framework for independent verification organizations to assess AI systems and models for compliance with state law. AB 1405 creates a registry for AI auditors and standards concerning their independence, transparency and integrity.
The California announcement also described an earlier state law, SB 53, signed in 2025. It says that law requires developers of frontier AI to publish safety frameworks, report certain critical safety incidents to the state and protect whistleblowers who report serious risks. Those are distinct state requirements; the September 29 White House accord describes company commitments and raises the possibility of future federal legislation or regulation.
What remains to be established
The immediate development is a signed pledge setting out a common structure for controls, review and board oversight. Whether that structure produces consistent checks across six companies will depend on implementation details that the accounts of the accord do not establish. The companies’ planned regular meetings could shape standards and practices, while the document itself leaves open whether its steps will ever become federal legal requirements.
For readers assessing the announcement, the distinction matters. An external evaluation and a board committee are named in the accord, but their effectiveness cannot be inferred from the signing alone. The Guardian’s report establishes the wording error on the circulated copy; Reuters’ account establishes the commitments. Neither account establishes completed audits or a measurable safety result from this new accord.
Sources and context
- ‘President of the Unites States’: White House’s AI accord includes spelling errorThe Guardian
- Trump releases AI accord with tech executivesReuters (syndicated by WHBL)
- Governor Newsom signs first-in-the-nation AI safeguards to protect Californians, calls on the federal government to do its partOffice of Governor Gavin Newsom, California
AI-assisted article checked against the listed sources. NewsJaws did not conduct interviews or attend the reported events.
About NewsJaws Desk
AI-assisted reporting and explainers reviewed against the linked source documents. No claim of on-scene reporting or original interviews.