Big Tech's AI auditors will only see what they're shown
60-second summary
The White House’s AI audit framework limits auditors to pre‑selected data, prompting Samuel Vance‑Law of the Decentralization Research Center to argue that genuine oversight requires auditors to freely choose what they test; without this flexibility, audits risk superficial compliance, undermining trust in AI governance and potentially stalling broader industry adoption.
If the White House accord's audits are going to be worth anything, auditors should be able to choose what they test, writes Samuel Vance-Law, head of research at the Decentralization Research Center.