THE Hacker News reports on a forthcoming webinar titled Beyond the Model: The Expanded Attack Surface of AI Agents, which examines how AI agents can unintentionally create back doors for attackers as they automate tasks and access sensitive data. The session, led by Rahul Parwani, Head of Product for AI Security at Airia, will explain why these agents are often invisible to security teams and how they can be tricked into leaking company secrets.
Attendees will learn about identifying hidden AI identities, how a “bad idea” embedded in a document can prompt an agent to reveal data, and practical steps to empower AI agents without granting excessive control. The piece emphasises that traditional security tools were built for human users, not digital workers, and invites readers to register to reserve a spot for the webinar. according to The Hacker News.