Hard Problems in AI Security at Meta and how We’re Addressing Them | RSAC Virtual 2025

There are few technology areas more important than the intersection of AI and cybersecurity in 2025. For AI to deliver on its promised value of autonomously solving manual tasks, we must ensure it behaves securely. And for AI to improve our cybersecurity, we need to identify where AI-driven forms of automation add value—even as the rapid growth of AI capabilities makes answering this a moving target.

How do we succeed? In this talk, I’ll lay out what I see as the core open questions in the field and provide opinionated answers.