US AI Safety Center Staff Asked to Report to White House
The Associated Press reported on Nov 8 that staff at CAISSI, the Commerce Department's AI safety evaluation unit, were recently asked to report informally to a White House office. The center has no permanent leader and has stopped publishing model safety assessments.

What happened
The Associated Press account cites two former government officials who spoke anonymously to protect current staff. The two said staff at CAISSI were recently asked to report informally to a White House office; the AP does not name the office, and the White House did not answer the AP's specific questions about the center.
CAISSI, the Center for Advancing Innovation and Standards for Super Intelligence, is an independent division of the Commerce Department, housed in the National Institute of Standards and Technology, and tests frontier models for safety before deployment. It began as the AI Safety Institute set up by the Biden administration in November 2023, was later renamed the US Center for AI Standards and Innovation, and was renamed again recently.
In a statement, the White House said Trump had formally established his Super Intelligence Force and commissioned a risk report from the President's Council of Advisors on Science and Technology, and that it will ensure continued US leadership on super intelligence. The statement did not answer the AP's specific questions about CAISSI. Critics said assessments in areas such as cybersecurity and bioweapons could be politicized or softened.
Key facts
- Report and sources: published by the Associated Press on Nov 8, based on two former government officials who spoke anonymously
- Core of the report: CAISSI staff were recently asked to report informally to a White House office; the AP does not name the office
- Nature of the change: the arrangement is an informal reporting line, not an announced formal restructuring
- Where it sits: CAISSI is an independent division of the Commerce Department, housed in the National Institute of Standards and Technology
- Name history: it began as the Biden-era AI Safety Institute set up in November 2023, was later renamed the US Center for AI Standards and Innovation, and was renamed CAISSI recently
- Leadership gap: a director hired in April was fired within days; his successor Chris Fall resigned in July 2026; NIST Director Arvind Raman has served as acting director since
- Budget: $10 million a year, about one-eighth of its British counterpart's
- Reports halted: since August, Congress has asked why the center stopped publishing model safety assessments
- Earlier move: on Oct 12, The Information reported three senior officials would move to the White House to report to Office of Science and Technology Policy Director Michael Kratsios and National Cyber Director Sean Cairncross; Commerce Secretary Howard Lutnick had invited the two to co-manage the center about six months earlier
- Responses and context: the White House did not answer the AP's specific questions; critics at the Brennan Center for Justice and the Federation of American Scientists warned findings could be politicized; on Oct 16 the center jointly published incident-response guidance with the NSA and CISA
Why it matters
the center is the US government's main civilian body that tests frontier models for safety before deployment. After the reporting line moved toward the White House, whether its findings stay independent and public is the next point Congress and outside observers are watching.
CompaniesUS government
Follow HIMIAO Horizon
No sign-up · Unsubscribe any timeNotifications
A ping on your phone or computer when the brief is out. No email needed.
Comments
Related
- Nov 5Google to Schools: No AI, No Search
- Oct 29US Government Claude Purchases Cut 40% as Anthropic Expands OneGov Deal With GSA
- Oct 18Commerce Ministry Notice No. 48: 23 AI Technologies Proposed for Export Control List, Registration Required Before Export
- Oct 11White House Tells AI Companies Incident Reporting Is Mandatory, After Anthropic's Disclosure
- Nov 4India Releases Three Homegrown Foundation Models: EKA-1 and Dhir 2B, 5B
Today
Nov 9 Monday- Brief
- GitHubGitHub Weekly Vol.5: 6 Projects Worth a Look This Week
Tested or read — weeklyAll →
- Quiz
- Call
Latest news
All →- Nov 9Researchers Find Macs Build an Undocumented Personal Knowledge Graph, Scoring Contacts Daily
- Nov 9Apple Wants Siri to Act Like an Agent, Bloomberg Reports, as Early as 2027
- Nov 9Amazon's Alexa Starts Shopping by Text: Message It to Search and Buy
- Nov 9Shield AI to Buy Radar Maker Echodyne for $5.3B, Raising $2.5B at a Doubled Valuation
- Nov 9NVIDIA AI Day Seoul: Physical AI Enters Its Golden Age, With Samsung, Hyundai and LG on Board
- Nov 9NVIDIA Board Opens Independent Probe as CEO Huang Reportedly Targeted by Foreign Intelligence Effort
™