Why One AI Safety Leader Wants to Ban Superintelligence

Connor Leahy, U.S. Executive Director of nonprofit ControlAI, argues that AI safety risks have grown too severe to manage through alignment and containment alone. He is pushing to stop companies from developing superintelligence altogether, calling it not a weapon but an adversary. On TechCrunch's Equity podcast, Leahy discusses the Sanders-Casar Ban Superintelligence Act, why frontier AI labs act as political actors, and the danger of AI that can build better AI. He also makes the case that international agreements could work, and that China may have little incentive to pursue superintelligence.