US CAISSI Safety Agency Faces Leadership Vacuum, Halts Reports and Is Forced to Report to White House

The CAISSI agency under the US Commerce Department has experienced the dismissal of its director and the resignation of a successor since April 2026. It still has no permanent leadership, has stopped updating its public safety reports, and has been required to report informally to an office in the White House.

Factual Reconstruction

According to reports from the Associated Press and Euronews, CAISSI stands for Center for Advancing Innovation and Standards for Super Intelligence and is part of the Commerce Department. In April 2026, former Anthropic researcher Collin Burns was dismissed only four days after becoming director. Three months later, his successor Chris Fall resigned. Currently, Arvind Raman, the head of its parent agency NIST, serves as acting director. In August 2026, Congress asked why the center had stopped publishing safety reports. In September, Trump asked AI companies to commit to a voluntary agreement involving "tremendous self-policing," but did not explain how compliance would be checked.

Recently, staff members have been required to report informally to a White House office. CAISSI has an annual budget of $10 million, about one-eighth that of a comparable UK institution. The EU AI Office, by contrast, has more than 125 employees and is recruiting 40 more.

Mechanism Analysis

Frequent leadership changes directly caused the interruption in report publication. Burns was quickly removed because of his ties to Anthropic, and Fall later resigned, leaving the agency without a stable decision-maker. The acting director is concurrently the head of NIST, meaning that daily work may be distracted by other matters at the parent agency. The arrangement requiring reporting to a White House office changes the original path of operating through the Commerce Department hierarchy.

The limited budget further constrains the agency's capacity. Annual funding of $10 million is insufficient to support large-scale independent testing and forms a clear gap with the staffing of the EU AI Office. This resource constraint, combined with the leadership vacuum, makes it difficult for the agency to continuously produce public assessment results.

Industry Impact

CAISSI's function is to test the safety of frontier AI in areas such as cyberattacks and bioweapons. The suspension of reports means that outsiders cannot promptly obtain its assessment conclusions. After being required to report to the White House, the assessment process may move closer to the administrative decision-making chain, reducing the transparency originally oriented toward the public.

The voluntary self-regulation agreement promoted by Trump in September requires support from external review mechanisms. CAISSI's current situation makes it difficult for the agency to play that role. AI companies had previously relied mainly on a narrative of self-regulation, but after this institutional change they face the reality of weakened external oversight.

Strategic Judgment (Analysis, Not Fact)

Based on the available facts, the leadership vacuum and the adjusted reporting path may weaken CAISSI's independence. The gap in budget and staffing limits its comparability with the EU's counterpart agency in terms of enforcement capacity. If assessment conclusions need to align more closely with White House decision-making, issues related to the public interest, such as privacy and civil rights, may be relegated to a secondary position. The long-term effect of this change depends on whether permanent leadership and report publication are restored later, but in the short term a contraction of oversight functions has already taken shape.

Historical experience shows that if safety assessment agencies lack stable leadership and independent reporting channels, their conclusions are easily influenced by short-term policy direction. CAISSI's current state provides the latest case of this pattern.