The White House Accord on Super Intelligence calls for companies training and deploying frontier models to work with an “independent external auditor or evaluator.” But its text does not define who qualifies, how the reviewer is chosen, what independence requires, what evidence the reviewer may examine, or whether anyone outside the company will see the results. The Accord establishes a place for external review in a four-layer oversight structure; it leaves the terms that would make audits comparable and publicly legible unspecified.
What does the White House Accord require?
The September 29, 2026 Accord describes four layers of controls and oversight. Its external reviewer is one layer, not a substitute for internal monitoring or board responsibility.
As an Amazon Associate I earn from qualifying purchases.
- Internal controls: Monitor model capabilities and alignment during training and deployment, including cybersecurity, biosecurity, chemical threats, and unintended system access.
- An internal oversight team: Check that monitoring and detection work as intended and that problems are remediated.
- External assessment: Partner with an “independent external auditor or evaluator” to assess whether controls, monitoring, and detection operate as intended.
- Board oversight: An independent board committee receives reports and oversees remediation.
The Accord says participating companies will meet regularly to establish standards and best practices, and that codifying steps into laws or regulations may make sense over time. Those are commitments and possibilities in the text, not a defined audit standard or a present legal mandate. Read the White House Accord on Super Intelligence.
The Tool Desk
Outbyte PC Repair FREEClear out junk files and repair common Windows errorsFree Scan →Outbyte Driver Updater FREEScan for outdated or missing drivers - takes under a minuteDriver Scan →Who gets to audit frontier AI?
The Accord does not name an auditor or specify qualifications, accreditation, or a selection procedure. It also does not set a conflict-of-interest test. So, based on the published text alone, there is no stated answer to who is eligible or who chooses the reviewer. The Accord calls for companies to partner with an external auditor or evaluator, but does not say whether the company, its board, or another party selects that partner.
#1 Best Overall
This is a gap in the Accord’s text, not proof that no other law, contract, or company policy could govern a particular review.
What does “independent external auditor” mean?
The Accord uses the phrase but does not define it. It does not state how financial or business ties, repeat engagements, or other relationships should be assessed, nor does it specify how an auditor’s independence should be protected in practice. Without those criteria, the text does not establish a common test for whether two companies’ reviewers are independent in the same sense.
A 2026 paper by Miles Brundage and coauthors defines frontier AI auditing as “rigorous third-party verification of frontier AI developers’ safety and security claims, and evaluation of their systems and practices against relevant standards, based on deep, secure access to non-public information.” The authors propose AI Assurance Levels (AAL-1 to AAL-4), ranging from time-bounded system audits to continuous, deception-resilient verification. This is a research proposal, not a requirement adopted by the Accord. Read the paper on frontier AI auditing.
What could an audit examine, and against what standard?
The Accord says an external reviewer should assess whether controls, monitoring, and detection are operating as intended. It does not define a required scope, name a standard against which to measure performance, or set minimum access to models, logs, or other evidence. The paper’s emphasis on secure access to non-public information helps clarify a practical issue: a reviewer needs access to evidence sufficient to verify the company’s claims, but the Accord does not specify what access is required.
Rank #3
For readers evaluating a proposed audit regime, the useful questions are:
- Can the evaluator inspect the evidence needed to test the company’s claims?
- Is the review measured against a stated standard, or only against controls defined by management?
- Is the scope a one-time assessment or recurring, and how often does it happen?
- Who receives the findings, and are identified problems required to be corrected?
- Can the public tell what was assessed, even if sensitive evidence must remain confidential?
Will the public see audit results?
The Accord does not require publication of findings or specify a reporting cadence. It says an independent board committee receives reports and oversees remediation, but does not say whether reports go to regulators, the public, or anyone beyond that committee. The text also does not say what happens when an audit identifies a problem beyond the board committee’s oversight of remediation.
The result is an unresolved transparency question: an external assessment may occur, but the Accord does not establish what information about its scope, conclusions, or corrective actions will be publicly available.
Is the federal review framework the same as the Accord audit?
No. Executive Order 14409, dated June 2, 2026, sets out a separate federal model-review framework. It directs agencies to develop a classified benchmarking process for advanced cyber capabilities and for deciding when a model meets the order’s “covered frontier model” designation. It also directs the administration to design a voluntary framework for developer engagement with the government, including providing covered models for up to 30 days before release to other trusted partners, subject to safeguards, and collaborating on partner selection.
The order explicitly says that section does not authorize mandatory federal licensing, preclearance, or permitting for new models. That government-facing framework is distinct from a company’s external audit under the later Accord. Read Executive Order 14409.
What remains unsettled?
The Accord creates a commitment to external assessment but leaves important operating details open. Until its signatories establish additional standards or disclose how they implement reviews, the text does not answer who conducts an audit, what makes the auditor independent, what evidence must be examined, what standard applies, how often reviews occur, or whether findings become public. Its own stated next step is for signatories to meet regularly to establish standards and best practices.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
Recommended Free Tools
