Former Treasury Secretaries Call for AI Safety Oversight Body

Former Treasury Officials Propose AI Safety Framework

Two former U.S. Treasury Secretaries are calling for new government structures to manage the systemic risks posed by artificial intelligence, drawing on their experience managing financial crises to argue that AI development demands similar oversight mechanisms.

Henry M. Paulson Jr., who served as Treasury Secretary from 2006 to 2009, and Robert E. Rubin, who held the position from 1995 to 1999, made their case in a Washington Post opinion piece published September 3, 2026. The officials argue that while warnings about AI risks have become "louder, more urgent and increasingly credible," concrete solutions remain scarce.

Why it matters

The proposal carries weight because it comes from officials who managed the 2008 financial crisis and earlier economic turbulence. Their argument positions AI safety not as a theoretical concern but as a systemic risk requiring the kind of institutional safeguards that now protect the financial system. This framing could influence how policymakers approach AI governance, particularly as Congress and regulatory agencies debate oversight frameworks.

Drawing Parallels to Financial Regulation

The core of Paulson and Rubin's argument rests on a comparison between financial system safeguards and the current state of AI development. Just as the financial sector built mechanisms to prevent and manage systemic risks after repeated crises, they contend that AI development now requires similar institutional protections.

The officials' perspective is informed by their direct experience implementing crisis management during periods of severe financial instability. Paulson oversaw Treasury during the 2008 financial crisis, while Rubin managed economic policy during the Asian financial crisis and the collapse of Long-Term Capital Management in the late 1990s.

The Current Gap in AI Governance

According to the former Treasury Secretaries, the AI field faces an imbalance: growing recognition of potential dangers without corresponding development of risk management structures. Their call for solutions suggests they see AI safety as requiring proactive institutional design rather than reactive crisis response.

The specifics of what oversight body or regulatory framework Paulson and Rubin envision were not detailed in the extracted portions of their opinion piece, though their emphasis on systemic risk management points toward centralized coordination rather than fragmented approaches.

The proposal adds to ongoing debates about AI governance as the technology's capabilities expand and its deployment accelerates across critical sectors of the economy and society. The involvement of figures with crisis management credentials may lend urgency to discussions about establishing formal oversight mechanisms before problems emerge.

The opinion piece was first reported by The Washington Post.