ControlAI's Leahy: Stop Building Superintelligence — SkimNews
Get the Tech newsletter
Daily tech — startups, AI labs, chips, the launches that shape the next decade. Free.
- ControlAI's new U.S. Executive Director Connor Leahy appeared on TechCrunch's Equity podcast to argue for stopping companies from developing superintelligence rather than trying to safely contain it
- Rebecca Bellan hosts the Equity episode that examines whether humanity can reliably control AI systems more capable than humans
- OpenAI's Hugging Face breach is cited as a recent safety incident showing the real-world dangers of deploying advanced AI systems
- Leahy contends the risks have grown too great to manage through alignment and containment approaches alone
- A wave of new legislation is now backing what Leahy says sounded far-fetched just six months ago, lending political weight to the moratorium push
- AI companies have continued framing superintelligent AI as inevitable even as safety incidents mount, according to the episode's framing
Why it matters: ControlAI's hardline position — halting superintelligence development rather than pursuing alignment — breaks from the dominant AI safety consensus of containment and control. The fact that legislation Leahy previously dismissed as far-fetched is now materializing gives the moratorium camp political momentum that didn't exist six months ago, putting AI labs on notice that a regulatory ceiling on capability is no longer hypothetical.
Ask SkimNews
