White House Reporting Raises Questions About Federal AI Safety Center

Staff at a Commerce Department unit that promotes innovation and sets standards for advanced intelligence, a division tasked with assessing AI safety, have recently been asked to report informally to a White House office. As of November 8, 2026, the center has no permanent director and has stopped publishing its research.

White House Ties Threaten Independent AI Safety Oversight

A Commerce Department center meant to assess AI safety is being drawn closer to the White House just as it lacks a permanent director and has stopped publishing research. The informal reporting arrangement raises a basic question: can government experts independently scrutinize powerful AI companies if their work is routed through the executive branch?

Critics fear findings could be softened to suit political priorities or the technology industry, which has largely been left to police itself. The center receives $10 million annually, and its director’s position has been vacant since two leaders departed in succession. Congress began asking in August why the division’s public reports had ceased.

The stakes are growing as AI agents act with less human prompting, while no standard process exists to evaluate their real-world conduct. President Donald Trump has asked leading companies to undertake voluntary self-policing, but his administration has not explained how it will monitor compliance. The case for independent, transparent government testing is therefore becoming more urgent, not less.

White House Could Give Underfunded AI Safety Center More Clout

A small Commerce Department division charged with assessing AI safety is moving into closer contact with the White House. Critics warn that this could compromise independence, but bringing officials nearer to President Donald Trump’s administration could also raise the center’s profile and help it secure resources it has long lacked.

The division has an annual allocation of $10 million—about one-eighth the budget of its U.K. counterpart—and has gone without a permanent leader after two directors left. It has also stopped publishing research, prompting questions from Congress. OpenAI has called for more money and support for the center.

The concern that government oversight could be politicized deserves attention, but so does the gap between rapidly advancing AI and the government’s ability to evaluate it. Trump has asked leading companies to take part in voluntary safety measures and commissioned an AI risk report from his science and technology advisers. The administration says it will pursue continued U.S. leadership in the field; a more prominent safety division could be part of that effort.