The Case for an Independent AI Oversight Commission
A joint campaign between GIE Foundation and the Align AI Foundation for an independent federal commission on artificial intelligence.
The Case for an Independent AI Oversight Commission
Every general-purpose technology that reshaped American life eventually acquired an institutional referee: Railroads got the ICC, securities got the SEC, aviation got the FAA, pharmaceuticals got the FDA. Artificial intelligence has reached that moment - and the institutions built for a slower century have not caught up.
Watch a three-minute explanation of the problem, the precedent, and the proposed AI Oversight Commission:
Why We’re Behind This
GIE Foundation exists to study how democratic institutions absorb the shocks of a faster and more unstable century, and to find where governing capacity has fallen behind the forces it's meant to manage.
Artificial intelligence is now the clearest case we study: A technology advancing faster than the institutions responsible for it, governed in the U.S. by fifty state legislatures working at cross purposes, a Justice Department litigation campaign, and a patchwork of voluntary pledges companies can drop the moment they're inconvenient.
We reached this proposal from the institutional side of the problem. The Align AI Foundation, a philanthropic funder of independent AI safety and alignment research, reached the same conclusion from the technical side: Its research shows that alignment failures are no longer a future hypothesis, and that the findings have no institution positioned to receive them or act on them.
Together, we've published a single proposed framework: The American Artificial Intelligence Leadership and Accountability Act, establishing the Artificial Intelligence Oversight Commission (AIOC).
Read more about it below.
The United States currently governs its most consequential technology through a patchwork of state statutes, executive orders, consent decrees, and voluntary commitments. We propose that Congress do for artificial intelligence what it did for securities markets in 1934: Create an expert, bipartisan, independent commission, the Artificial Intelligence Oversight Commission (AIOC), created through the American Artificial Intelligence Leadership and Accountability Act.
The framework is disclosure-based, not licensing-based. It touches fewer than a dozen frontier developers and statutorily exempts everyone else. Read the position paper, structure and understand why this matters, below, courtesy of the Align AI Foundation.
What We Are Proposing
Full Position Paper + Legislative Framework
A position paper on the American Artificial Intelligence Leadership and Accountability Act, which would establish the Artificial Intelligence Oversight Commission. This includes a full section-by-section legislatve architecture, risk register, bipartisan landscape analysis and an implementation timeline.
For policymakers, advisors, legislators and the media: The overview of the proposed Commission and the Act.
Two-Page Policy Brief
A state-of-play overview on the current risks and harms of large language models (LLMs), the physical and social costs of the data center buildout, and the potential future risk of unaligned AI, with sources throughout. Written for donors, journalists, and anyone new to the field. Courtesy of the Align AI Foundation.
Current Harms Case Study
Public Campaign Presentation
A public-facing overview presentation making the case for the AI Oversight Commission, built for general audiences, including a focus on what LLMs are already doing, why voluntary self-regulation isn't oversight, what the Commission would do instead. Courtesy of the Align AI Foundation.