CAISI Director Resigns After Three Months Amid U.S. AI Standards Uncertainty

Chris Fall, director of the Centre for AI Standards and Innovation (CAISI), has resigned after serving for just three months, according to multiple media reports citing confirmation from the agency. No explanation was provided for his departure, marking another leadership change for the organisation responsible for developing AI standards and evaluation methods in the United States.

Fall took over the role after his predecessor, Collin Burns, left less than a week into the job. Reports at the time said Burns was pushed out because of his previous employment at Anthropic during a period of tension between the AI company and the Trump administration. Before joining CAISI, Fall held senior leadership positions at the Department of Energy, including director of the Office of Science during President Donald Trump’s first administration and acting director of the Advanced Research Projects Agency-Energy.

Leadership changes follow broader shifts in AI oversight

Before Burns and Fall, the organisation was led by venture capitalist David Sacks in his capacity as the White House AI and crypto czar. Sacks stepped down in March, continuing a period of frequent leadership turnover at the agency.

CAISI operates under the National Institute of Standards and Technology (NIST) and is responsible for developing technical standards, testing methodologies and cybersecurity assessments for advanced AI systems. Despite that mission, the agency was not directly involved in one of the year’s most significant AI policy disputes.

In June, the U.S. Commerce Department used an export control directive that temporarily forced Anthropic to withdraw its Mythos and Fable AI models from the market. The restriction was lifted later that month after Commerce Secretary Howard Lutnick said he was satisfied with the company’s AI safety commitments.

Questions grow over the future of AI standards.

Earlier this month, the White House introduced a new AI safety oversight initiative called Gold Eagle to coordinate cybersecurity vulnerability reporting across several federal agencies. Although the Commerce Department and the Department of Homeland Security were included, CAISI was notably absent from the programme.

The agency’s future role has also come under renewed scrutiny after Google DeepMind CEO Demis Hassabis publicly called for an independent, industry-led organisation to establish frontier AI standards, drawing comparisons with the Financial Industry Regulatory Authority (FINRA). The proposal closely mirrors the objectives for which CAISI was originally established.

Fall’s resignation also comes as policymakers debate how to respond to increasingly capable open AI models developed in China. The release of Moonshot AI’s latest Kimi model sparked renewed discussion over potential restrictions on Chinese open-weight AI systems, prompting criticism from technology leaders including David Sacks, who argued that regulation should focus on safety rather than protecting domestic AI companies from competition.

Although CAISI has published evaluations of Chinese open-weight models such as Z.ai’sChris Fall, CAISI, NIST, AI standards, artificial intelligence, U.S. Commerce Department, Anthropic, David Sacks, AI regulation, Gold Eagle GLM-5.2 and DeepSeek V4 Pro, the agency has disclosed little about how its model-testing process operates. Questions surrounding those evaluation methods remain unanswered as scrutiny of U.S. AI governance continues to grow.