Six AI Companies Sign White House Accord on Safety Controls and External Audits
2026-09-29 ยท Governance
Google, Anthropic, Meta, OpenAI, xAI, and NVIDIA signed a White House accord committing to internal safety controls, an empowered internal checking team, independent external auditing, and board-level oversight.
Six companies signed the White House Accord on Super Intelligence on September 29: Google, Anthropic, Meta, OpenAI, xAI, and NVIDIA. Greg Brockman signed for OpenAI. The accord commits signatories to four layers of oversight: internal safety controls, an empowered internal checking team, independent external auditing or evaluation, and an independent board committee that oversees reports and remediation. Signatories also commit to regular meetings on shared standards and practices.
Attendance at the earlier White House meeting is not the same as signing the accord; SIRDA's list of signatories reflects the signed document itself. The accord is a voluntary agreement among the companies, not a new statutory regulator or an enforceable treaty, though its text discusses the possibility of future codification into law or regulation.
For SIRDA's comparison of governance proposals, the accord is a concrete corporate oversight arrangement that sits alongside proposals for public governance authority. Who runs the internal checking team, what the external auditors can actually see, and what the board committee can compel are the open questions. Trump's September 27 dinner with Anthropic CEO Dario Amodei remains a separate, earlier activity entry, not part of this accord.