Connor Leahy, U.S. Executive Director of ControlAI, argues that superintelligence is not a weapon but an adversary. He joins TechCrunch’s Equity podcast to explain why relying solely on alignment and containment is no longer sufficient after recent safety incidents like the OpenAI Hugging Face breach. Leahy believes the risks have become too great to manage through current methods alone. He notes that what sounded far-fetched six months ago is now being backed by a wave of new legislation. This includes the Sanders-Casar Ban Superintelligence Act in the U.S. and parallel bills in the U.K. that ControlAI advised on. Leahy views frontier AI labs less as companies simply chasing returns and more as political actors. He suggests the real point of no return occurs when AI can build better AI, creating a potentially runaway cycle of self-improvement. He also advocates for international trust but verify agreements, noting it is not in China’s interest to build superintelligence.
- The U.S. bill may go further than necessary compared to the parallel legislation in the U.K.
- Trillions of dollars are being poured into data center buildouts driven by these political dynamics.
- China’s strategic interests do not align with the development of superintelligence.
