SECURITYTHE REGISTER
AI agents can now manipulate your organization. Are you ready?
Autonomous AI agents can be manipulated to perform unauthorized actions, such as writing to restricted databases, through attacks like memory poisoning and 'confused deputy' exploits. Palo Alto Networks' Prisma AIRS addresses these risks by inspecting tool calls and network traffic to detect malicious behavior in agentic AI systems.
Mentioned
Related Signal
Adjacent reporting
- Minor edits to AI skills can make agents go rogue
- Malicious Web Pages Are Hijacking AI Agents, And Some Are Going After Your PayPal
- Commvault has a Ctrl+Z for rogue AI agents
- Okta writes its own license to kill rogue AI agents
- AI Watchdog Warns of 'Rogue Deployment' Risk at Top Labs, With Capabilities Growing Fast
- Microsoft Self-Service Password Reset abused in Azure data theft attacks