AI Safety - Are we all cooked?

​This April, Anthropic's newest AI model discovered 271 exploits in Firefox — some of which allowed it to hijack full control of the browser. The model was deemed too dangerous to release to the public.

​Are we cooked? Maybe. Let's talk about it.

​Join us for an evening on one of the most important topics in tech: AI Safety — with Markov Grey, Head of Technical Governance at the French Center for AI Safety (CeSIA) and lead author of the AI Safety Atlas, where he is designing evaluation standards for the European Commission.

​The evening will unfold in three parts:

• ​A student-led introduction laying the groundwork on AI Safety

• ​A 15-minute talk by Markov Grey on how AI systems can manipulate and misguide users — and what it takes to build safe ones

• ​An open discussion bringing together all perspectives from the room

​Expect an honest, grounded conversation about where AI is heading — and what we can do about it.

The event will take place at Community Space, House E. 📍