The US government’s new artificial intelligence task force has formally begun work. After its first official meeting on October 15, the White House panel adopted a charter and a 120-day window to answer core questions about how Washington should manage the rapid development of advanced AI systems.
Key takeaways
- The task force is chaired by Director of National Intelligence Jay Clayton.
- Its review covers cybersecurity, biosecurity and “model control and accountability.”
- It will also implement a White House accord on advanced AI signed by major AI company leaders earlier this month.
Who sits on the task force
Alongside Clayton, three vice chairs lead specific workstreams. Defense Undersecretary Emil Michael oversees protocols for AI labs to report breaches and hacks and coordinates with frontier developers. Federal Trade Commission Chair Andrew Ferguson leads a law enforcement and consumer-protection committee that includes the Justice Department. Office of Personnel Management Director Scott Kupor works on critical infrastructure and, with the Labor Department, on AI’s impact on the workforce. Members include Vice President JD Vance, Treasury Secretary Scott Bessent and science adviser Michael Kratsios, among others.
The three risk areas
- Cybersecurity and critical infrastructure: how powerful AI models could be used to find and exploit software vulnerabilities, and how to defend power grids, water systems and networks.
- Biosecurity and dual-use applications: preventing AI from lowering barriers to creating biological threats while preserving benefits for medicine.
- Model control and accountability: ensuring developers can monitor, correct and report problematic model behaviour.
The President’s Council of Advisors on Science and Technology will evaluate risk in each area.
The open question is enforcement: a review can identify risks, but it is less clear what tools the government will use to make companies comply.
Industry and civil society
The task force says it will consult frontier model developers, consumer-protection groups, civil liberties advocates and law enforcement at federal and state levels. The relationship between government and AI labs has been tense at times, with disputes over government access to models and the use of export controls as leverage.
What to expect by February
A 120-day timeline points to recommendations in mid-February 2027. Observers will watch for proposals on incident reporting, testing requirements for the most capable models and how the federal approach interacts with state AI laws. More in our Technology & AI section.