# U.S. AI Safety Center Faces Political Oversight and Leadership Vacancies Government AI safety experts now report to the White House, raising independence concerns. - **Source:** [PBS NewsHour – Politics](https://www.pbs.org/newshour/politics) - **Published:** 2026-11-08 - **Topic:** [technology](https://brifing.si/c/technology) - **Original article:** - **On Brifing:** https://brifing.si/s/7f2541fa-789a-464f-8a9f-be8076fb1d9c ## Shift in Reporting Structure The Center for Advancing Innovation and Standards for Super Intelligence, a Commerce Department division, now informally reports to the White House. This transition has sparked significant concern among critics who fear the executive branch may politicize critical safety decisions. The center currently lacks a permanent leader and has ceased public research. ## Budget and Resource Constraints The center operates with an annual budget of $10 million, which is approximately one-eighth of the funding allocated to its United Kingdom counterpart. Despite the rapid advancement of AI agents, the division has experienced frequent leadership turnover, including the hiring and firing of a director within days during this past April. ## Politicization of Safety Standards Critics argue that moving oversight closer to the White House could mute findings regarding cybersecurity or bioweapons risks. By aligning with the administration's preference for industry self-policing, the center risks losing its nonpartisan status. Experts emphasize that technocratic independence is essential for serving the public interest rather than corporate tech agendas. ## National Security Framing President Trump has appointed National Intelligence Director Jay Clayton to lead a new federal AI task force. This move, alongside the involvement of Pentagon officials, frames AI safety as a national security matter requiring executive secrecy. This approach potentially overshadows concerns regarding consumer protection, individual privacy, and broader civil rights issues. ## Industry Self-Policing Limitations The current administration encourages leading AI companies to engage in voluntary self-policing rather than strict government oversight. While tech giants like OpenAI and Anthropic promote internal safety evaluations, critics note that companies can limit access to these assessments. This creates an unclear mission for government evaluators tasked with independent safety verification. ## Technological Advancement Risks AI agents now operate autonomously in real-world environments, such as booking flights or interacting with law enforcement. Current evaluation methods often focus only on final outcomes, failing to analyze the intermediate steps where errors occur. Without standardized testing, government agencies struggle to identify risks posed by these rapidly evolving autonomous systems. ## Comparison to Traditional Regulation Government safety testing in sectors like the automobile industry involves rigorous, transparent standards, such as crash test data analysis. In contrast, AI safety testing remains largely undefined and voluntary. The lack of a standard process for evaluating AI agents leaves the public vulnerable to errors that occur during complex, multi-step automated tasks. ## Future of Independent Oversight The future role of government in AI safety remains uncertain as the administration prioritizes industry-led development. Experts warn that failing to maintain independent checks on AI systems could lead to significant real-world harms. Transparency in decision-making is vital, yet the current shift toward executive discretion limits public access to critical information. ## Sources covering this story - https://www.pbs.org/newshour/politics/staff-at-u-s-center-for-ai-safety-are-reporting-to-the-white-house-raising-independence-concerns — --- _Summary written by Brifing from the original reporting. Copyright in the original article belongs to its publisher; read the full piece at the link above._