About This Session
Software security has fundamentally changed. Advances in AI are accelerating vulnerability discovery, shrinking the time between disclosure and exploitation, and making sophisticated attacks accessible to a wider range of adversaries. Organisations need to be taking this threat seriously. Don't panic, but do take action. This talk will walk through the seriousness of the situation before focusing on remediations -- how to mythos-proof your code and systems. We'll look at running LLMs defensively as well as more traditional security methods to provide protection.
Topics
- AI Models
- Anthropic
- Security