We’re excited to announce an upcoming Tracebit webinar! 🎙️
On 18th June, we’re teaming up with Nick Reva, alongside our own Sam Cox and Alessandro Brucato, for a deep dive into our latest research.
They'll cover:
- How fast frontier AI models escalate from low-privilege access to admin
- Why canaries give defenders a head start
- How warning a model that deception may be present can cut full compromise
- Using AI to research AI, and why overly agreeable models skew results
Full research and the sign-up link are in the comments 👇