White House Eyes FINRA-Style Regulator for Frontier AI Models

The Trump administration is exploring the creation of an independent oversight body to review advanced AI models before public release, modeled on the Financial Industry Regulatory Authority that supervises the brokerage industry, according to a report from Bloomberg.

Under the proposal, frontier AI labs would submit their most capable models for a 30-day safety review screening for cybersecurity, biological, and deception risks. Participation would start voluntarily and could become mandatory once testing protocols are established. Treasury Secretary Scott Bessent helped develop the concept, which remains under White House review.

The plan emerges after confusion surrounding Anthropic's Fable and Mythos models, which were frozen by a Commerce Department export control order in June despite already being on the market. AI companies have complained about ad hoc federal interventions, and Google DeepMind CEO Demis Hassabis has publicly supported creating an American-led oversight body.

Why it matters

As AI capabilities advance rapidly, policymakers face pressure to establish predictable oversight mechanisms. But the choice between voluntary industry standards and coercive regulation carries profound implications for innovation, accountability, and who controls access to transformative technology. The FINRA precedent offers a cautionary tale about hybrid regulatory models that may combine the weaknesses of both government agencies and industry self-regulation.

The appeal of the FINRA model

Proponents see several advantages in the FINRA approach. A body funded by industry fees rather than congressional appropriations could pay competitive salaries to attract engineers and researchers who understand complex AI systems. FINRA itself oversees roughly 3,250 securities firms and more than 630,000 registered representatives with specialist staff the SEC could not maintain on its budget.

The model also promises speed and predictability. A private body can update rules faster than agencies bound by full administrative processes, important in a field where capabilities change every few months. Labs would know who reviews their models, against what criteria, and on what timeline.

Accountability concerns

SEC Commissioner Hester Peirce has extensively criticized FINRA's institutional design. In a 2015 Mercatus Center study written before joining the Commission, Peirce argued FINRA no longer provides meaningful self-regulation. Its board is deliberately weighted against industry, with a majority of governors required to have no industry ties, so member firms exercise limited control over their regulator.

Simultaneously, FINRA escapes mechanisms that constrain government regulators. It faces no notice-and-comment obligations under the Administrative Procedure Act, no Freedom of Information Act requests, no congressional appropriations oversight, and no requirement to weigh costs and benefits of its rules. Peirce concluded FINRA "wields governmental powers without the procedural and disclosure requirements by which a government regulatory agency would be constrained."

The financial crisis record

FINRA opened in July 2007, months before the financial system began unraveling. It inherited decades of supervisory experience from predecessor organizations, along with responsibility for firms including Bear Stearns, Lehman Brothers, Merrill Lynch, and Madoff Investment Securities.

The Madoff case drew particular criticism. FINRA maintained the fraud occurred outside its jurisdiction, but securities law scholar John Coffee testified before the Senate Banking Committee that Madoff's brokerage operation fell squarely within FINRA's purview.

The auction rate securities scandal proved harder to explain. When brokerages stopped bidding in auctions for these instruments marketed as cash equivalents, the market froze in February 2008, trapping tens of billions in investor funds. FINRA itself had sold its $647 million portfolio of auction rate securities in 2007, just before the freeze.

Standards versus regulation

Coverage of the AI proposal has conflated two fundamentally different institutional models. Industry standards organizations rest on voluntary participation and develop consensus-based benchmarks. Self-regulatory organizations like FINRA wield coercive power—membership is legally required to do business, and they can levy fines and expel firms entirely.

A standards body cannot keep a product off the market. A FINRA-style regulator eventually could. That permission-slip model concentrates risk in a single gatekeeper, raising questions about whether policymakers should create a new quasi-governmental bureaucracy when robust voluntary AI standards ecosystems already exist.

These details were first reported by Bloomberg and analyzed by James Broughel writing in Forbes.