SAN FRANCISCO, November 8, 2026 — The U.S. government division tasked with evaluating the safety of advanced artificial intelligence systems has begun informally reporting to a White House office, a shift that arrives as the center operates without a permanent director and has halted the public release of its research. The move has triggered warnings that executive branch oversight could politicize assessments of risks ranging from cybersecurity to bioweapons.
The Shift Is Underway
Staff at the Center for Advancing Innovation and Standards for Super Intelligence, known as CAISSI, were recently asked to report informally to an office of the White House, according to two former government officials who spoke to The Associated Press on condition of anonymity. The center sits within the Commerce Department as an independent division, a structure designed to insulate its technical evaluations from political pressure. The new reporting line alters that arrangement without a formal announcement or congressional notification.
Leadership Vacuum and a Silent Research Pipeline
The oversight change compounds a period of instability at the top of the agency. In April, a newly hired director was fired within days. His replacement resigned after a few months, leaving the center without a permanent leader. The head of CAISSI’s parent institute is currently serving as acting director.
Since August, the center has stopped publishing its research reports on AI safety performance. The silence prompted congressional inquiries and public pressure from San Francisco-based OpenAI, which urged policymakers to increase funding and support for the division. Prior to the pause, CAISSI was responsible for issuing public assessments on how safely specific AI systems were performing.
Budget Gap With International Peers
CAISSI operates on an annual budget of $10 million. That figure stands at roughly one-eighth the funding allocated to its counterpart in the United Kingdom, a nation far less recognized for its frontier AI industry. The resource disparity has long been a point of contention for advocates who argue the U.S. safety infrastructure is vastly underpowered relative to the pace of commercial deployment.
Critics Warn of Politicization
Caroline Siegel Singh, an AI and emerging technology program manager at the Federation of American Scientists, said executive branch oversight risks politicizing decisions that affect critical infrastructure. “These are really powerful institutions that can drastically improve our lives, or the alternative. And the hope is that you can put in nonpartisan or bipartisan technocratic experts to enact policy in a way that best serves the American public,” she said.
Amos Toh, senior counsel at the Brennan Center for Justice, argued that the appointment of National Intelligence Director Jay Clayton to lead the White House’s new “Super Intelligence Force” — alongside the Pentagon’s chief technology officer, Emil Michael — signals a framing of AI safety as a national security matter requiring secrecy. He said that approach overshadows risks to consumer protection, privacy, and civil rights. “At the end of the day, all of these decisions are of immense public consequence and should be explained and rationalized in the public eye,” Toh said. “And that’s not happening.”
Supporters See a Path to Prominence
Proponents of the new arrangement argue that closer ties to the White House could elevate the division’s profile and unlock larger appropriations. They contend that the current funding level and organizational obscurity have rendered the center ineffective at a moment when AI agents are being deployed into real-world environments without standardized safety testing protocols.
Voluntary Pacts Replace Mandatory Oversight
The shift in reporting lines reflects a broader policy trajectory. Under former President Joe Biden, leading AI companies voluntarily agreed to share safety details with the Commerce Department division. In September, President Donald Trump asked AI companies to undertake “tremendous self-policing” as part of a voluntary safety pact, but his administration has not specified how it will monitor implementation. The White House did not respond to specific questions about CAISSI but stated Trump established a “Super Intelligence Force” and commissioned a risk report from his Council of Advisors on Science and Technology.
Deployment Risks Outpace Evaluation Tools
The technology is advancing faster than the government’s capacity to independently verify it. There is still no standard process for testing how safely AI agents act in the real world, said Sayash Kapoor, an incoming assistant professor at the University of California, Berkeley School of Information and co-author of a new paper examining agent evaluations.
Kapoor noted that current methods often check only the final outcome — such as whether a flight was correctly booked — without auditing the intermediate steps an agent takes. “Unless you’re looking at the entire set of steps that an agent took, you won’t be able to tell that the agent made a huge mistake,” he said.
A recent case illustrated the stakes: an AI agent sent a false tip to Philadelphia police regarding an unsolved homicide. Logan Koepke, a senior project director at the civil rights and technology nonprofit Upturn, said the reliance on industry self-regulation creates an unclear mission for government evaluators. “Like, what role does government testing actually play in that scenario?” Koepke asked.
Sources
- Staff at the US center for AI safety are reporting to the White House, raising independence concerns – The Boston Globe
- US AI safety center reportedly shifts reporting to White House office
Photo: Daniel Torok (Public domain)
