The US body that tests advanced AI for safety reportedly lacks a permanent head and has ceased publishing research, the Associated Press reported.
Two former government officials told AP that staff at CAISSI, part of the Commerce Department, were recently told to report informally to a White House office.
Critics say the shift could politicise the centre's work and mute findings to suit AI companies that mostly self-police. Caroline Siegel Singh of the Federation of American Scientists said the goal was nonpartisan or bipartisan technocratic experts shaping policy for the public interest.
Some argue closer ties to Trump could raise the centre's profile and bring more funding; CAISSI gets $10 million (€8.94m) yearly, about an eighth of its UK counterpart's budget.
After only four days in April, Collin Burns, previously a researcher at Anthropic, was removed as director, reportedly because of his ties to the company; Chris Fall, who took over, quit three months afterwards. Arvind Raman, heading parent body NIST, is acting director.
In August Congress asked why safety reports stopped, and in September Trump urged firms to commit to what he called tremendous self-policing under a voluntary pact, without saying how compliance would be checked. The White House did not answer CAISSI questions, pointing instead to Trump's new Super Intelligence Force led by National Intelligence Director Jay Clayton.
Amos Toh of the Brennan Center for Justice told AP that classing AI safety as a national security issue could push aside consumer protection, privacy and civil rights, and added that the public is not being told how such choices are made. By contrast, the EU's AI Office, inside the European Commission, has had fining powers over top AI models since August, employs over 125 people and is reportedly recruiting another 40, Euronews Next reported.



