The White House AI accord: what tech companies agreed to

On Tuesday, September 29, leaders of several of the largest artificial intelligence companies met at the White House and signed a document titled the White House Accord on Super Intelligence: Joint Commitment on Frontier Responsibilities. If you saw the headlines and wondered what it actually says, here is a plain summary.

According to Forbes, which published the text, the companies agreed to four layers of oversight. Each company commits to internal controls that monitor what its models can do and whether they behave as intended, with attention to cybersecurity and biosecurity threats. Each also commits to an internal team that checks those controls and fixes problems, an independent outside auditor or evaluator, and a board committee that reviews the reports and makes sure issues are addressed.

KPBS reported that signers included the leaders of Anthropic, Google, Meta, OpenAI, Nvidia and xAI. President Trump called the accord "morally binding." The document itself describes the steps as voluntary and says it may make sense over time to write them into laws or regulations. KPBS also reported that some of these practices were already in place at the companies in some form.

Reactions have been mixed. Some researchers quoted by KPBS said voluntary commitments are a start, while others argued that legal liability and regulation would do more. Those debates are likely to continue as lawmakers consider AI rules.

If you use AI tools at work or at home, the accord does not change anything you need to do today. It is still a good habit to read a tool's privacy settings and to double-check important answers against a trusted source.