ControlAI 联合创始人 Connor Leahy：超级智能不是武器，而是对手

My Analysis of the Text

Okay, so here's what's running through my mind. I see the core issue being that AI companies are hyping up superintelligent AI as if it's a foregone conclusion, but recent incidents, like that OpenAI Hugging Face breach, are exposing the real dangers of deploying systems that are smarter than we are. Essentially, if we can't reliably control what these systems do, what's the outcome? The text then references a TechCrunch podcast episode where Rebecca Bellan is talking to Connor Leahy, who has a background in AI research and entrepreneurship, and is now the US Executive Director of […]. The main point, as I read it, is a concern about safety and control as AI capabilities advance. The mention of the podcast provides a context, likely where this concern will be further explored.

各大人工智能公司一直在大谈超级人工智能，仿佛这已是必然趋势，但近期发生的安全事件（例如 OpenAI 的 Hugging Face 数据泄露事件）正表明，部署能力超越人类的 AI 系统存在潜在危险。那么，当我们无法可靠地控制这些系统的行为时，会发生什么？在本期 TechCrunch 的《Equity》播客中，Rebecca Bellan 邀请到了人工智能研究员、创业者兼现任 […] 美国执行总监的 Connor Leahy。