Trump’s AI Standards Chief Resigns After Just Three Months

Chris Fall, the director of the Center for AI Standards and Innovation (CAISI), resigned in July 2026 after roughly three months on the job, the agency confirmed to multiple news outlets. He is now the third person to exit the U.S. government's top AI standards role in a matter of months. His predecessor, Collin Burns, lasted less than a week. Before Burns, venture capitalist David Sacks held the broader White House AI and crypto czar title until he stepped down in March. No reason was given for Fall's departure.

What happened

Detail	Fact
Departing director	Chris Fall
Role	Director, Center for AI Standards and Innovation (CAISI)
Time in post	Approximately three months
Predecessor	Collin Burns, pushed out in April 2026 after less than one week
Before Burns	David Sacks (White House AI and crypto czar), stepped down in March 2026
Reason for Fall’s exit	Not given

CAISI operates under the National Institute of Standards and Technology (NIST) and is the primary federal body responsible for developing technical standards and testing methods for AI models, as well as assessing cybersecurity risks from those systems.

Before joining CAISI, Fall led the Department of Energy’s Office of Science during the first Trump administration and served as acting director of the DOE’s Advanced Research Projects Agency-Energy. He had also worked in the DOE’s Office of Naval Research.

Burns, the man Fall replaced, was reportedly pushed out in April because he had previously worked for Anthropic, a company the Trump administration had been in conflict with at the time.

Why it matters

CAISI is being sidelined at the worst possible moment

The leadership churn is happening as AI model risk sits at the top of the policy agenda. In June, the U.S. Commerce Department used an obscure export control directive to force Anthropic to pull its Mythos and Fable models from the market. The ban was lifted by the end of the month after Commerce Secretary Howard Lutnick said he was satisfied with Anthropic’s safety plans.

Earlier in July, the White House signed an executive order for a new AI safety program called “Gold Eagle,” which sets up a clearinghouse for cybersecurity vulnerability coordination. The Commerce Department and the Department of Homeland Security were both named in the order. CAISI was not, as CNBC pointed out.

That omission is notable. CAISI was created specifically to do the kind of model risk assessment that “Gold Eagle” is now routing through other agencies. If the body tasked with AI standards is being left out of major AI safety orders and cannot hold onto leadership, it raises real questions about whether it can do its job.

Calls for a different approach are growing

Google DeepMind CEO Demis Hassabis recently called for an independent, industry-run standards body for frontier AI modeled on FINRA (the Financial Industry Regulatory Authority, a self-regulatory body for U.S. broker-dealers). That is almost exactly the role CAISI was created to fill, yet Hassabis is apparently looking past it.

Meanwhile, the administration is reportedly weighing how to restrict Chinese open-weight AI models following the release of a competitive new version of Moonshot’s Kimi model. Open-weight means a model’s weights can be downloaded and run locally, though its training code and data are not public. CAISI has released reports on Chinese open-weight models including Z.ai’s GLM-5.2 and DeepSeek V4 Pro, but has said little about how it actually tests those models.

As of publication, TechCrunch reported it had sent multiple inquiries to the Department of Commerce and NIST since July 9 asking about the LLM evaluation process and had received no response.

For context on how Chinese lab valuations are shifting the competitive picture, see our earlier coverage of DeepSeek’s Liang Wenfeng and his $71 billion valuation.

Our take

Three directors in roughly four months is not a leadership transition problem. It is a signal that the role has no stable political footing. CAISI is supposed to be the serious, technical counterweight to AI hype and risk. Right now it looks like a waiting room.

The “Gold Eagle” order skipping over CAISI entirely is the more worrying detail. If the White House is routing AI safety coordination through DHS and Commerce while leaving its own standards body out, that either means CAISI is being deliberately marginalized or the administration’s left hand genuinely does not know what its right hand is doing. Neither option is reassuring for anyone who needs predictable AI regulation to plan a product roadmap.

If you are building AI-integrated products or advising clients on compliance, do not bet on CAISI output as a stable policy signal right now. Watch what Commerce and DHS actually publish under Gold Eagle instead. Our AI integration work with clients always accounts for regulatory uncertainty, and right now that uncertainty is unusually high.

What to do about it

1. Stop treating CAISI standards documents as settled policy. Treat them as working drafts until the agency has stable leadership.
2. Watch the Commerce Department and DHS for Gold Eagle implementation rules. Those agencies are where AI safety coordination is actually happening.
3. If your product touches frontier AI models, document your own safety and testing processes now. Voluntary documentation is fast becoming the price of doing business, as Anthropic’s June situation showed.
4. Monitor Chinese open-weight model restrictions. Any export control that bans downloadable weights would directly affect self-hosted AI deployments. Check your stack.

The practical takeaway: U.S. AI standards policy is being improvised in real time, so build your compliance posture around what agencies are actually doing, not which agency has the word “standards” in its name.

Source: TechCrunch · AI