How AI agents could undermine computing infrastructure security – Help Net Security

In this Help Net Security video, Ev Kontsevoy, CEO at Teleport, explores the risks AI agents pose to computing infrastructure, particularly when exposed to social engineering attacks.

Unlike traditional software, AI agents aren’t fully deterministic, making them more vulnerable to manipulation. Like all software, we’ll examine how these agents can be exploited through implementation flaws and why this raises new security concerns.