Chris Fall has resigned from his position as Director of the Center for AI Standards and Innovation (CAISI), the central US regulatory authority for the safety of AI models. The US Department of Commerce confirmed the resignation on Monday, July 20, 2026, just three months after Fall took office in April. The agency did not provide a reason.
CAISI leadership temporarily transitions to NIST chief Raman
Fall was previously the Director of the Office of Science at the US Department of Energy under the first Trump administration and held positions at the Advanced Research Projects Agency-Energy as well as the Office of Naval Research. Department spokesperson Kristen Eichamer stated according to Axios that the position was intended as a transitional solution from the beginning. A permanent appointment is expected to be established within a few weeks. Until then, Arvind Raman, confirmed as the 18th Director of the National Institute of Standards and Technology (NIST) since June 30, 2026, and previously Dean of Engineering at Purdue University, will temporarily take over the CAISI leadership in addition to his own position. CAISI is subordinate to NIST and emerged in 2026 from the former US AI Safety Institute, which the Trump administration renamed and restructured as part of a realignment of AI policy. Since then, the agency has been developing standards and testing procedures for advanced AI systems and evaluating unpublished models from companies like Anthropic, Google DeepMind, and OpenAI for safety risks before their market launch. For these evaluations, CAISI requires close, ongoing contact with the labs – precisely this continuity is now facing another disruption due to the change in leadership.
Predecessor Burns held the position for only four days
Fall is already the third AI official in the government to leave within a year. His direct predecessor, Collin Burns, had to vacate the CAISI position in April 2026 after only about four days. According to TechCrunch, the decisive factor was his previous work at Anthropic, which was viewed within the government as a conflict of interest – a sign of ongoing tensions between the White House and the AI company, which has repeatedly advocated for stricter AI regulation in the past. There has also been no permanent appointment at the next higher level for months: David Sacks, who was responsible for overarching AI policy in the White House as the “AI and Crypto Czar,” left office in March 2026 without a successor being named. The repeatedly short tenures complicate, according to several US media, a continuous testing practice for new AI models, especially since CAISI must work closely with the development labs to gain access to unpublished systems. For the model providers, this means additional uncertainty about which testing standards will apply in the long term.
Vacant leadership meets dispute over Chinese models
The vacancy comes at a time when the government is internally debating restrictions on inexpensive Chinese AI models. The trigger is, among other things, the competitor Kimi K3 from the Chinese provider Moonshot AI, which, according to a report by SAN, can compete with US top models in specific tasks like frontend programming – at significantly lower computing costs per request. According to data from the provider OpenRouter, the weekly share of Chinese models in usage has been over 30 percent since February 2026. Within the government, measures such as adding Chinese AI companies to the Department of Commerce’s Entity List or a joint security warning from the NSA and the White House have been considered but not yet implemented. One reason against this: many Chinese models are available as open weights that companies have already downloaded and can therefore continue to use independently of future export regulations. Ironically, in this debate, CAISI, as the responsible testing authority, currently lacks a firmly confirmed leadership that could represent a position on the opportunities and risks of Chinese models and provide advisory input.
It remains open who the Department of Commerce will appoint as the permanent CAISI chief and whether this personnel decision will also resolve the position of the overarching AI Czar in the White House, which has been vacant since March. Until a permanent successor is in place, the technical evaluations for the next generation of models will continue under a temporary dual leadership – while the decision on how to handle Chinese AI models remains pending.
