Superintelligence is coming. Should we let it?

AI companies have presented superintelligent systems as an eventual outcome of current research, but recent safety incidents have raised questions about the wisdom of that trajectory. The Equity podcast examined those concerns and whether society should allow development of systems that could exceed human capabilities.

Connor Leahy, who now serves as the U.S. executive director of ControlAI, told host Rebecca Bellan that his organization advocates for stopping companies from building superintelligence. He said recent failures show alignment and containment strategies may no longer be sufficient to limit the risks posed by highly capable AI.

Leahy also noted that proposals to prevent or strictly limit development of superintelligent systems — ideas that many considered unlikely just months ago — are starting to gain traction in policymaking circles. He pointed to a wave of new legislation backing more radical approaches to AI safety.

The episode credits Theresa Loconsolo as an audio producer for Equity; her TechCrunch bio says she joined the outlet in 2022, previously worked as a producer at a four-station conglomerate, and holds a bachelor's degree in Communication from Monmouth University. Theresa can be reached at theresa.loconsolo@techcrunch.com.