Prioritizing security while adopting AI agents


In this episode of The Security Podcasts featuring Michael Bargury, Co-Founder & CTO at Zenity, we discuss best practices for securely implementing AI agents.

“People from outside of the organization, they can send you an email — they can send you a teams message — that ends up being internal data. And so, when you have these agents that are operating within the enterprise, and they are operating on behalf of users, if an attacker can take control over these agents, and they do it through prompt injection and jail breaking, then that agent now becomes an agent of chaos.”

Prioritizing security while adopting AI agents

Michael Bargury, Co-Founder & CTO at Zenity, shares best practices for securely implementing AI agents.

Listen to more from the Security Podcasts.

Or check it out on our Apple Podcasts or Spotify channels. Don’t forget to like, follow, and rate and review our podcasts!



Source link

Leave a Comment