Start the day here

Policy — Meta — Regulation

Zuckerberg Told Trump a National AI Regulator Was Flawed

Business Insider reported on September 3 that Mark Zuckerberg told President Trump, on a private call the week of August 17, that a proposed national AI regulator was a flawed idea. The proposal, championed by Google DeepMind’s Demis Hassabis and favored by some White House officials, would create an independent body modeled on FINRA, the securities industry’s front-line self-regulator under SEC supervision. The imagined job: review advanced models and test them for risk before broad deployment.

One person familiar with the call said Zuckerberg did not demand Trump kill the plan outright. He argued that anyone appointed to such a body should reflect Trump’s light-touch approach to AI. Trump placed the call. That detail matters. The Oval Office is now an informal appellate court for Silicon Valley objections after policy gets drafted downstairs.

6 min read
A hand holds a yellow-cased phone showing a calculator app over a folder of tax forms

FINRA for models, MPA for vibes

Hassabis floated a FINRA-style “standards body” in a July essay and briefed White House officials this summer. FINRA writes and enforces rules for thousands of broker-dealers, funded by member fees, with SEC oversight. An AI twin could hire the technical talent actual agencies struggle to keep. Critics fear voluntary pre-release testing becomes soft law, then hard law.

David Sacks, Trump’s former AI and crypto czar, wants a different metaphor: the Motion Picture Association’s voluntary ratings system, which he has called a way to forestall “a DMV for AI.” He already proved the informal channel works. In May he called Trump the morning of a planned AI executive-order signing and helped scrap it. Zuckerberg’s August call sits in that same pipeline: private pressure after public process.

The China month

Zuckerberg’s August essay stated the policy preference without needing a secret phone line. Any rule that slows a model release, even by a month, would “add significant risk to American leadership” over China. That sentence is the whole theory. Speed is sovereignty. Oversight is lag.

The administration still claims it wants to balance innovation and security. Officials told BI the FINRA-style idea remains under consideration alongside a voluntary industry group. Meta declined to comment. The White House offered a balancing platitude. Nobody denied the call.

Judgment

A FINRA for AI would still be industry money supervising industry machines, with a government shadow. Zuckerberg’s objection is not that self-regulation is too cozy. It is that any queue at all is too slow.

Spreading open weights was Meta’s earlier safety story. Calling a national reviewer flawed is the companion move: keep the state out of the release calendar. If American AI policy is going to be written on phone calls after the staff work is done, say so. Do not pretend the FINRA memo was ever the last word.

Live scoreboardFollow the AI race on AI Wars

Lab rankings, model preference, API volume, coding-agent heat, open-source stars, and prediction markets.

Related stories

An empty operating room with a black surgical table under twin ceiling lights and wall monitorsBusiness

How OpenAI Wired ChatGPT Into Epic's Patient Charts

Today

Gold dome and white Corinthian columns of the Massachusetts State House against a clear blue skyBusiness

Massachusetts's 120-Day AI Evaluators

Today

Yellow freight locomotive pulling tanker cars past railway speed-limit signsBusiness

Will OpenAI Restart Astra Before Its Safeguards Catch Up?

Today

Letters

0

No letters yet.

Write a letter