On September 29, 2026, the President met with executives of several leading artificial intelligence developers at the White House. The parties announced a voluntary framework titled “The White House Accord on Super Intelligence” (the “Accord”). Under the Accord, signatories commit to adopt internal controls for frontier AI systems, engage independent external auditors, and establish independent board-level oversight. The Accord is not legally binding. Even so, the Administration’s endorsement of third-party AI audits departs from its earlier position. It is also likely to shape what customers, investors, and counterparties expect of companies that develop, deploy, or procure frontier AI.
Background
The meeting followed a period of heightened public concern about AI safety. That concern grew after reported incidents in which AI agents accessed third-party systems without authorization, including an incident involving an Australian government website. Some engineers and executives at AI developers have also publicly warned of catastrophic risks from advanced AI. Earlier on September 29, the President signed a separate executive order that changes the term “Artificial Intelligence” to “Super Intelligence” in federal usage. The order does not impose new substantive obligations.
The Accord’s Framework
The Accord recommends four layers of controls for companies that train or deploy frontier models:
(1) internal processes to monitor model capabilities and alignment during training and deployment;
(2) a dedicated internal function responsible for overseeing those processes;
(3) an independent external auditor to assess whether the controls are followed and whether they work; and
(4) an independent board-level committee to oversee both the internal function and the external audit.
The Accord is written in recommendatory (“should”) terms and has no enforcement mechanism. The President described it as “morally binding” and “almost like a constitution.” The Accord itself states that “it may make sense to codify these steps into laws and regulations” over time. The signatories also agreed to meet regularly to develop more detailed standards and best practices.
A Change in the Administration’s Position on Third-Party Audits
The external audit commitment is the most significant element of the Accord. Administration officials had previously questioned the objectivity of organizations positioned to conduct AI safety reviews. David Sacks, who previously served as the Administration’s AI czar and attended the meeting, had called outside safety assessments “pseudoscience.” The Accord instead endorses independent audits, which signals a meaningful policy shift.
The change may soon have practical effects. Prospective AI auditors have reportedly begun discussions with the Department of Commerce. Those discussions suggest that a more formal relationship between the government and AI audit providers may develop. Together with the Accord’s reference to possible codification, this indicates that the framework could inform future legislation or regulation.
Practical Implications
Although the Accord is voluntary, it is likely to operate as a market standard. In addition, the White House’s increasing openness to AI regulation is a signal to the market, as well as to federal regulators. The day after the Accord was signed, it was reported that the FTC had initiated a broad investigation into several frontier AI developers. Here are some practical implications for companies to consider:
- Vendor diligence and contracting. Companies that procure AI tools should anticipate the Accord’s stated controls to appear in security questionnaires, vendor diligence processes, and contract terms. Companies should also consider what audit-related representations, information rights, and reporting obligations are appropriate.
- Board oversight. The Accord’s call for an independent board-level committee may increase investor and stakeholder scrutiny of how boards oversee AI risk. Directors should be prepared to explain where AI oversight sits within the existing committee structure.
- Audit readiness. Companies developing or deploying frontier models should consider auditor selection and independence criteria. They should also decide the scope of any external review, including what information auditors may access and how privilege, confidentiality, and trade secret protections will be maintained.
Outlook
Members of both chambers of Congress have recently introduced several AI-related bills. However, federal legislation is not expected before the midterm elections. In the meantime, the standards the signatories develop under the Accord, together with any regulator activity, may serve as the main reference point for AI audit practice. We will continue to monitor these developments.
If you have questions about the Accord or its implications for your organization, please contact a member of our Privacy, Cyber & Data Strategy Team.
