Media & Interviews · Georgetown
Keir Lieber explains why Cold War arms control is an imperfect model for AI governance
The Georgetown professor distinguishes state-led nuclear competition from company-led AI development and questions whether governments can reach a meaningful global AI agreement without strong incentives.
What Lieber argues
Keir Lieber distinguishes the Cold War nuclear arms race, which was driven primarily by state security competition, from today’s AI competition, where private companies play the leading role. He argues that this difference makes direct comparisons between nuclear arms control and AI governance less reliable.
Lieber remains skeptical that governments can reach a meaningful global agreement to slow or regulate AI when states and companies believe continued development serves their interests. He suggests that cooperation becomes more plausible when the participating governments or companies see a concrete strategic or commercial benefit.
Why Hoyas.ai is following it
The interview connects Georgetown expertise in international security to live questions about U.S.–China competition, private-sector power, and the conditions under which AI governance agreements might become viable.
Evidence boundary
These are Lieber’s interpretations, presented in an official Georgetown interview. They do not establish a Georgetown policy position, forecast a specific agreement, or show that the university is participating in AI arms-control negotiations.