Director of the Center for AI Standards and Innovation (CAISI) Chris Fall left his post just three months after being appointed by the administration of President Donald Trump. The US Department of Commerce reported this to CNBC.
Axios previously reported the resignation, citing sources. TechCrunch calls Fall the “AI czar” of the Trump administration.
The department did not officially disclose the reasons for the official’s departure. Commerce Department spokeswoman Kristen Ihamer said National Institute of Standards and Technology (NIST) Director Arvind Raman will temporarily lead the center and continue to oversee CAISI.
The Commerce Department plans to announce a new permanent director in the coming weeks, according to Axios. At the same time, one of the representatives of the department claims that Fall’s appointment was initially considered as temporary, and Raman had already been selecting candidates for this position in recent weeks.
CAISI is part of the US Department of Commerce and is responsible for developing methods for testing and evaluating artificial intelligence systems, as well as preparing safety standards for advanced AI models. The center was created after the Trump administration reorganized the former US AI Safety Institute in April.
As CNBC notes, Fall's departure comes at a time when the Trump administration is forming a new system of government control over the development of artificial intelligence. In addition, after the departure of venture capitalist David Sachs in March from his post as the White House curator for AI and cryptocurrencies, the administration has yet to appoint a new person responsible for this area.
The personnel changes coincide with increased competition between American and Chinese artificial intelligence developers. Last week, the Chinese company Moonshot AI introduced the Kimi K3 language model, saying that it is close in capabilities to leading American developments, and in a number of tests is superior to OpenAI and Anthropic models.
In June, Trump signed an executive order increasing government control over the world's most powerful artificial intelligence systems. The document provides for the voluntary transfer of new models to federal authorities to assess their capabilities before widespread launch and obliges relevant departments to develop a unified testing system within 60 days.
Against this backdrop, OpenAI said it had initially limited the distribution of GPT-5.6 series models at the government's request, and Anthropic had temporarily removed access to Fable 5 and Mythos 5 models after receiving an export directive from the Department of Commerce. Later, both companies opened access to their developments to a wider range of users.
In recent weeks, the administration also announced the creation of the Gold Eagle program, which would coordinate the search for cyber vulnerabilities and regulate companies' access to the most advanced AI models. However, CAISI was not mentioned among the departments participating in the implementation of this initiative.
