When Your AI Has a Login and a Mind of Its Own

About this episode

The episode discusses the implications of autonomous AI agents in digital systems with insights from a Microsoft strategist.

Autonomous AI agents are no longer science fiction, they’re already reshaping how we work, build, and protect digital systems. In this episode, Jeremy Huval and Ryan Patrick are joined again by Richard Diver, Security and Identity Strategist at Microsoft, to break down what "agentic AI" really means and why it matters now. Richard unpacks the core building blocks of agent behavior (like entitlements, autonomy, and memory) and shares where the biggest risks lie as organizations rush to adopt agent-based systems. From identity sprawl to memory poisoning and the need for lifecycle management, this episode gives insights for security and GRC leaders who want to get ahead of the next wave of AI-driven innovation. Meet Ryan: https://www.linkedin.com/in/ryan-patrick-3699117a/ Meet Jeremy: https://www.linkedin.com/in/jeremyhuval Meet Richard Diver: https://www.linkedin.com/in/rdiver/

People in this episode

Hosts: Jeremy Huval, Ryan Patrick

Guest: Richard Diver

Topics covered

• AI
• autonomous agents
• digital security
• identity management
• risk management

Keywords

• agentic AI
• entitlements
• autonomy
• memory poisoning
• identity sprawl

Mentioned in this episode

Organizations: Microsoft

More episodes of Trust vs.

• Cybercrime Has Gone Cartel · December 18, 2025 · 29 min
• Hacking Back, Legal Limits, and the Future of Cybersecurity Regulation · November 20, 2025 · 40 min
• Why Compliance Fatigue Is a National Security Risk · October 23, 2025 · 42 min
• Why SOC 2 May Not Prove Security Anymore · October 9, 2025 · 41 min
• Chaos, Culture, and Cyber Resilience · September 25, 2025 · 37 min
• Uncle Ronnie and the Problem with AI Security · September 11, 2025 · 32 min