Key Takeaways:
The Trump administration is considering creating an independent AI regulator modeled after the Financial Industry Regulatory Authority to review frontier model safety, following complaints from Silicon Valley leaders that current ad-hoc measures are too arbitrary.
Treasury Secretary Scott Bessent helped draft the proposal, which would establish a self-regulatory organization reporting to the Securities and Exchange Commission, according to people familiar with the matter. The plan is under review by White House Chief of Staff Susie Wiles.
"The strength of this approach is it would be technically focused, while at the same time supporting innovation and incentivising responsible behavior," Demis Hassabis, chief executive officer of Google DeepMind, said in a post on X Tuesday, laying out a similar vision for a standards body that would review models up to 30 days before release. Hassabis explicitly cited FINRA as the model for his proposed framework.
The push for a formal regulatory structure comes after the US government conducted ad-hoc reviews of Anthropic's Mythos and OpenAI's Sol — evaluations that drew criticism for lacking technical expertise and opaque decision-making around release timing. White House AI advisor Sriram Krishnan has previously ruled out an "FDA for AI" within the executive branch, making the self-regulatory model a potential compromise that could address industry concerns about heavy-handed government oversight.
A Self-Regulatory Middle Ground
Establishing the body as a self-regulatory organization akin to FINRA — a private, nonprofit watchdog funded by the industry it oversees — could bypass political resistance to a traditional federal regulator. Under Hassabis' framework, frontier labs would voluntarily share models for review before release, with formalization following once the assessment protocol proves effective.
The proposal has drawn a mixed response. Anthropic, which has endorsed some of the nation's toughest state-level AI laws in California, New York, Illinois and Massachusetts, has argued that transparency and self-reporting are "no longer sufficient safety measures for the most powerful AI systems," according to Cesar Fernandez, the company's head of US state and local government relations. But critics including former White House AI czar David Sacks have accused Anthropic of pursuing "regulatory capture" through fear-mongering, arguing the company is trying to trap smaller startups in red tape.
What's at Stake for the Industry
The creation of a formal AI regulator would introduce new compliance costs and approval timelines for major model releases, potentially slowing product launches for companies including OpenAI, Google and Meta. Tech giants with strong compliance teams may benefit relative to smaller players, while the framework could provide the regulatory clarity that institutional investors favor for long-term industry stability.
The last time the US government imposed structured pre-release review on an emerging technology was in the financial sector after the 2008 crisis, when the Dodd-Frank Act expanded SEC and FINRA oversight of derivatives and high-frequency trading. That regulatory expansion added an estimated $70 billion in annual compliance costs across the industry, according to a 2015 study by the American Action Forum, though proponents argue it reduced systemic risk.
The White House has not set a timeline for a decision on the Bessent proposal. Hassabis said his proposed standards body could be "ratcheted up if the seriousness of the situation demands," suggesting the framework is designed to scale with the technology's advancement.
This article is for informational purposes only and does not constitute investment advice.
