The Zuckerberg Trump AI regulator call that took place during the week of 17 August has thrown fresh uncertainty over White House efforts to create an independent body to vet advanced AI models before deployment, with Meta’s chief executive telling President Donald Trump directly that he opposed the proposal, according to a senior White House official directly familiar with the conversation.
The plan under debate would establish an organisation modelled on the Financial Industry Regulatory Authority (FINRA), the private, not-for-profit membership body that writes and enforces rules for over 3,000 brokerage firms operating in U.S. securities markets under the supervision of the Securities and Exchange Commission (SEC). Applied to AI, the concept would see frontier models reviewed and tested for potential risks before broader commercial release, a form of pre-deployment gating that has no direct equivalent in FINRA’s own remit, which has little involvement in signing off on new investment products.
Origins of the FINRA-Style Proposal
The idea originated with Demis Hassabis, the Nobel Prize-winning Google DeepMind scientist, who set out the case in a July essay for a new U.S.-led Frontier AI Standards Body to handle risks including emerging cybersecurity threats. According to the Council on Foreign Relations, Hassabis published that framework on 14 July. He subsequently briefed White House officials about the concept this summer, according to an administration official familiar with the matter.
Hassabis has since pressed for speed. He told Axios he wants the body operational within “months,” ideally before the end of 2026, according to AI Weekly. That timeline now looks considerably more contested given the resistance emerging from major industry players. Google did not respond to a request for comment.
The institutional architecture being discussed has attracted serious backing at a senior level. According to Lawfare, Treasury Secretary Scott Bessent helped develop a proposal for an independent frontier AI regulator modelled on FINRA and reporting to the SEC, a version of which is now under review by White House Chief of Staff Susie Wiles. That level of Treasury and chief-of-staff involvement signals how far the idea had progressed before industry pushback began to register.
What the Zuckerberg Trump AI Regulator Call Reveals About White House Policymaking
Senior White House officials had previewed the FINRA-style proposal separately with major tech companies, including Meta, OpenAI and Anthropic, in mid-August, before Trump spoke directly with Zuckerberg. According to a second person familiar with the matter, Zuckerberg did not ask Trump to alter his position outright, but suggested that any appointees to the regulatory body should reflect Trump’s own approach to AI, widely characterised as light-touch. That person also confirmed that Trump placed the call first.
The exchange fits a pattern. In May, David Sacks, Trump’s former AI and crypto czar, called Trump the morning of a planned executive order signing and persuaded him to cancel the ceremony, as POLITICO reported at the time. On the 21 August episode of his ‘All-In’ podcast, Sacks outlined an alternative he said has support from Elon Musk: a voluntary industry body modelled on the Motion Picture Association (MPA), which administers the film rating system giving movies ‘R’ or ‘PG-13’ labels. Sacks has described the FINRA-style option as ‘a DMV for AI’ where models ‘get lined up in a queue, waiting to get their test done.’
The White House is therefore weighing two distinct paths: a FINRA-style regulator with mandatory pre-deployment testing, or a voluntary industry group on the MPA model. These are not equivalent in regulatory terms. A mandatory pre-deployment gate changes the incentive structure for developers fundamentally; a voluntary ratings system does not, though proponents of the latter argue, as Sacks put it on his podcast, that the MPA approach historically ‘forestalled more intrusive, heavy-handed government action.’
The biggest AI companies have not publicly endorsed or opposed either option. Anthropic co-founder Jack Clark posted favourably about the FINRA idea on X in July; Anthropic declined to comment when approached. OpenAI and Google did not respond to requests for comment. A spokesperson for Meta declined to provide a statement.
The Zuckerberg Trump AI regulator call did not kill the proposal. It remains under active consideration inside the White House, alongside the voluntary MPA-model alternative, with the administration’s stated position being that it is ‘committed to balancing innovation and security in AI policymaking,’ according to a White House spokesperson. Whether Hassabis’s target of a body running before the end of 2026 survives that internal competition is now the live question.