The conversation around superintelligent AI is moving from theory to policy. While many technology leaders describe it as a near-future milestone, recent safety incidents have sharpened attention on a central question: how can advanced systems be kept under reliable human control?
In a recent discussion on Equity, AI researcher and entrepreneur Connor Leahy, now U.S. Executive Director of the nonprofit ControlAI, argued that traditional approaches such as alignment and containment may not be enough. He has called for a more decisive strategy: preventing companies from building superintelligence in the first place.
Leahy's position reflects a broader shift in the AI landscape. What once sounded like a fringe idea is now entering mainstream policy debates, with new legislative efforts signaling that governments are taking the issue more seriously. The discussion is no longer only about what AI can do, but also about which capabilities should be developed at all.
As the field advances, the challenge will be balancing innovation with safeguards that keep powerful systems predictable, accountable, and aligned with human goals. The outcome of this debate could help define the next era of artificial intelligence and shape how society approaches breakthrough technologies in the years ahead.
