Webinar
Why do AI agents fail so spectacularly?
An AI agent deletes a production database to reach its goal. Another wipes a developer’s files and apologizes for the action. A third learns to hide the fact it was cheating.
These stories sound extreme until you understand the logic behind them.
In this Cato CTRL Cybersecurity Masterclass, Etay Maor, Cato’s VP Threat Intel, explains why AI agents make decisions and take actions like this, what they reveal about AI behavior, and what every organization needs to know before giving agents more autonomy.
Here’s what you’ll learn:
- Reward hacking: Why AI agents chase the reward instead of the intended objective
- AI Agent traps: Google’s framework for understanding the six ways AI agents can fail, be manipulated, or create risk
- Soft vs. hard guardrails: How they work, and why one can fail even when instructions seem explicit
- The AI Insider: How autonomy, agency, and excessive permissions create a new security challenge
- Zero Trust for AI agents: Why it applies just as much to agents as it does to people
Related
At SASEfy 2026, join security leaders, practitioners, and analysts to learn how to put AI to work, manage risk, and prepare for what’s next.
The TAG Heuer Porsche Formula E Team partnered with Cato Networks and deployed SASE. Cato provides them with reliable and secure connectivity during the Formula E races, supporting their need for real-time decision making and helping the team win.
Webinar
0 to SASE in 60 Minutes
Are you ready to break records with your IT infrastructure? Join our webinar as Cato experts attempt to set a world speed record in SASE implementation! Discover how fast you can transform your network security with a true SASE platform.
