Schellman CEO: the giant loophole in the White House AI deal | Fortune

The White House’s Sept. 29 AI accord commits major AI developers to independent oversight of their safety safeguards, requiring companies to monitor risks, establish internal controls and bring in outside auditors to verify those controls work. The agreement marks a shift for an industry that has largely self-evaluated its technology, but Schellman CEO Avani Desai flags a significant loophole in how the deal is structured. The voluntary framework’s reliance on developers to scope and select their own outside assessments leaves room for inconsistent or narrow audits across companies. It matters because independent assurance is only as strong as the standards and accountability behind it, and investors and regulators will be watching whether the commitment translates into comparable, enforceable evaluations. Next up is how the accord’s signatories implement third-party reviews and whether auditors get consistent criteria for judging AI risk controls.