AI companies have been talking about superintelligent AI as though it is inevitable. However, recent safety incidents like the OpenAI-related Hugging Face breach are highlighting the potential dangers of deploying AI systems that exceed human capability.

The core question being examined is what happens when we cannot reliably control what these advanced systems do. TechCrunch's Equity podcast explored this issue with Connor Leahy, an AI researcher and entrepreneur, discussing the implications of developing AI systems we may not be able to manage.