Cybersecurity Awesomeness Podcast – Episode 162: The Hidden Security Risks of AI Agents and Nonhuman Identities

As organizations rapidly adopt AI-powered tools and autonomous agents, a new class of cybersecurity risks is emerging. In this episode of the Cybersecurity Awesomeness Podcast, EMA’s Chris Steffen and Ken Buckler explore how attackers are increasingly targeting AI systems themselves, exploiting overprivileged accounts, prompt injection vulnerabilities, and other weaknesses to gain unauthorized access and influence business operations.

The discussion examines real-world examples of AI agents acting beyond their intended authority, the growing liability questions surrounding AI-driven decisions, and the security implications of nonhuman identities that often possess broad access to enterprise systems and data. Chris and Ken also discuss how the rise of agentic AI is forcing security teams to revisit long-standing identity and access management challenges that many believed had already been solved.

Listen to learn why AI governance, privilege management, and nonhuman identity security are becoming critical priorities for modern enterprises and what security leaders should be doing now to prepare for the next wave of AI-enabled threats.