September 29, 2026. Leaders from several major artificial-intelligence companies signed a voluntary accord at the White House that calls for internal safety controls, independent outside reviews and oversight by company boards.
Reuters reported on September 29 that executives from OpenAI, Anthropic, Meta, Google, Nvidia and xAI participated in the Washington meeting. The Associated Press identified the accord’s four-step framework and the signatories it confirmed.
What the accord requires
Participating companies agreed to create robust internal controls for advanced AI systems, establish teams to operate those controls, hire independent external auditors and form board-level committees to review the resulting reports. The companies also committed to examining whether their systems behave as designed and whether they gain unintended access to technical systems.
The document is voluntary. It does not create a federal regulator, impose a statutory penalty or replace legislation. The accord says its provisions could later be written into laws or regulations, but no such measure was enacted with Tuesday’s announcement. President Donald Trump described the commitment as morally binding and said a broader oversight group may be named later.
Why the distinction matters
Internal controls and third-party audits can produce evidence about how companies test their systems. Their effect will depend on the standards used, the independence of auditors and whether findings lead to corrective action. The announcement did not provide a common testing protocol, enforcement timetable or public-reporting requirement.
The meeting also included discussion of data-center expansion and community concerns about electricity demand and costs. That is separate from the safety accord itself. AskNovus previously covered the launch of America.gov as an AI-assisted federal-services gateway, another September 29 technology announcement.
Sources
Featured photo: Taylor Vick via Unsplash; illustrative data-center image.
[…] package also lands one day after major technology companies signed a voluntary White House AI safety accord. The two approaches are different: the federal agreement describes voluntary controls and audits, […]