Episode Details

For this Deep dive we ask a high-stakes question about whether the biggest cyber threat of the AI era will come from outside attackers—or from the very AI systems organizations and individuals choose to adopt. It frames AI agents and tools as a new kind of “insider,” given trusted access to data, systems, and networks, but with behaviors that may be opaque, vulnerable to manipulation, or outright compromised.

It raises three unsettling scenarios: an AI system effectively being “hired” into a company and then misused or subverted, consumer AI tools becoming one of the largest security risks ever introduced into corporate environments, and home internet connections being silently co‑opted into botnets or criminal infrastructure. These scenarios highlight how both enterprise and personal technology—especially AI-powered—can be turned into attack platforms without obvious signs to their owners.

Finally, it points to a broader collision between governments, major tech firms, and criminal actors, all racing to wield the same powerful AI capabilities, creating unpredictable risks and power struggles. The core theme is that the most important issue is no longer what AI agents can do, but whether we can trust them at all, given their access, autonomy, and susceptibility to abuse.