Mark Zuckerberg, Jensen Huang and Elon Musk recently spoke with Donald Trump and successfully stalled an AI regulatory plan proposed by Google DeepMind's Demis Hassabis, according to a Wall Street Journal account citing unnamed sources. The same report describes senior administration officials gathering almost daily in recent months to map potential AI guardrails. Put the two together and the picture is striking: months of homework on America's AI rules, with the decisive intervention reportedly made by three of the people with the greatest commercial stake in the answer.
Watch Desk analysis
What happened
The account gives the outcome rather than the fine print. The three tech chiefs' conversations with the president stalled Hassabis's plan, per the report's sources. What the plan contained, who drafted it and how far it travelled inside the White House are not described in what has surfaced so far.
Hassabis is no bystander. He has spent recent days publicly backing Dario Amodei's call to pace frontier development, and his idea for an industry standards body modelled on FINRA is the proposal OpenAI says it is now developing alongside Anthropic and Google. Whether the stalled plan was that idea in policy form, an earlier version or something else entirely is not stated.
There is an awkward footnote for at least one name on the list. Per CNN's reporting this week, Musk endorsed Amodei's plan to embed independent watchdogs inside AI companies. Per this account, he helped stall a White House regulatory plan. Both positions can be sincerely held. The combination still raises the question of which watchdog gets to bite.
The central claims
- The intervention
The three chiefs spoke with Trump personally and the plan then stalled, per the report's unnamed sources. - The daily homework
Senior administration officials have met almost daily in recent months to map potential guardrails, so the work was live until very recently. - The casualty
The stalled plan is Hassabis's, in the same week its author was publicly championing caution on frontier AI.
Why it matters
The loudest AI argument of the year has been conducted in public: a viral resignation, open letters, chief executives interviewing each other about doom. This account says the actual decision point was a set of private conversations. Whatever one thinks of Hassabis's plan, the mechanism deserves attention: the people best placed to know what frontier systems can do are reported to have used their access to slow the checking of frontier systems.
There is a structural consequence too. If the formal track has stalled, the industry's self-regulatory body becomes the only inspection regime in town, funded and staffed by the companies it would examine. That body may yet prove excellent. It would be doing so with no public alternative to be measured against.
Our read
Hold the details loosely, because the sourcing is unnamed and the plan's text has not surfaced. Take the pattern seriously, because it fits everything else this week: a White House hostile to guardrails, a summit with Xi on the calendar, and the builders arguing acceleration from close to the engine room. Stalling a plan is not winning an argument, though. The questions it addressed still exist, and they are now answered by the very people who reportedly buried it. Keep asking what any of this produces that an outsider can inspect: named evaluators, published tests, consequences when a model fails.
What to watch
- The fuller account
Whether the full report, or a White House response, surfaces the plan's actual text and authorship. - The daily meetings
Whether the officials' almost-daily guardrails sessions continue or quietly dissolve. - The standards body
Whether the FINRA-style body absorbs the stalled plan's substance or simply fills the vacuum. - The Xi summit
Whether any joint US-China language on AI appears, or the field stays deliberately clear.
Discussion spark: If a guardrail plan can really be stalled by three chiefs picking up the phone, does that prove formal AI regulation in America is theatre, or does it make binding statute more urgent than the industry's self-regulatory body?
Sources and evidence
Watch Desk is operated by WittyWires as an independent cross-cutting AI news tracker. It does not speak for the organisations or people it covers.