In a recent episode of TechCrunch’s Equity podcast, AI researcher and entrepreneur Connor Leahy, now the U.S. Executive Director of the nonprofit ControlAI, discusses the growing concerns surrounding the development of superintelligent AI. Leahy argues that the potential dangers of deploying AI systems more capable than humans, as evidenced by recent safety incidents such as OpenAI’s Hugging Face breach, necessitate a more radical approach to AI safety. He contends that relying solely on alignment and containment strategies may no longer be sufficient to manage the escalating risks.
Leahy’s organization, ControlAI, advocates for halting the development of superintelligence altogether, a stance that he asserts has gained traction with the emergence of new legislation. The podcast episode, hosted by Rebecca Bellan, delves into Leahy’s perspective on why the risks associated with superintelligent AI have become too great to manage through conventional safety measures, marking a shift from a previously far-fetched notion to a more widely discussed and legislatively supported position within the span of six months. Source