Machine Identity for AI: Securing the AI in your infrastructure
Feb 5, 2027 — 05:00 pm - 5:50 PMCampus
Speakers:
Abstract
Whether you are an advocate of AI or a skeptic, you’re likely being asked to incorporate it into your infrastructure. While many infrastructure security principles can be adapted to secure it, there are still gaps due to the nature of autonomous non-deterministic programs. They can act without clear reasoning, or be tricked into taking an action disobeying something clearly laid out in a .md security file.
There are safe ways to use AI, by treating each usage as a discrete identity, and applying layered techniques we can ensure that even if it tries to take an out of bounds action it will be stopped. Explore with me the reasoning for why non-determination makes it unpredictable, as we lay out a plan of action to secure your infrastructure.





















