ControlAI leader urges a ban on superintelligent AI development

In short: ControlAI’s U.S. executive director Connor Leahy is calling for laws that stop companies from developing “superintelligent” AI.

What happened

Connor Leahy, the U.S. Executive Director of the nonprofit ControlAI, spoke on TechCrunch’s Equity podcast about why he thinks the push toward superintelligent AI should be paused or banned.

“Superintelligence” is a term for an AI system that is smarter than humans across many tasks. Leahy argues that treating it like a weapon misses the point. In his view, it is closer to an adversary, meaning something you might not be able to control once it is capable enough.

The podcast discussion comes after recent AI safety incidents, including reports related to OpenAI and a breach connected to Hugging Face. These incidents are used as examples of what can go wrong when powerful systems are deployed without reliable controls.

Leahy said he does not think current approaches like “alignment and containment” are enough. Alignment is the idea of training AI to follow human goals. Containment is like putting the system in a locked room with strict rules, so it cannot cause harm (like keeping a tiger in a cage). He also pointed to new political momentum, including the Sanders-Casar “Ban Superintelligence Act” in the U.S. and parallel efforts in the U.K. that ControlAI advised on.

Why it matters

This debate could shape what kinds of AI products are allowed to be built and sold in the coming years, and how governments respond to safety problems. If lawmakers move toward bans or long pauses, that could slow down new AI capabilities, and it could also change which companies and countries lead AI development.

Source: TechCrunch AI