Nov 3, 2026 · 36 min
AI agents expose a new identity-security gap
Democracy without a safety net.
As organizations deploy more automated and agentic systems, permissions once designed for software now create unfamiliar paths to misuse and attack.
- 1Non-human identities now span applications, scripts, and systems that handle sensitive information across organizations.
- 2Agentic tools can unexpectedly use resources and permissions, making visibility and reasonable mitigation central to identity security.
- 3A political fundraising-text investigation shows how personal data and campaign messaging systems can misroute persistent appeals.
Don't miss
Alex Weinert explains how agentic systems can unexpectedly use organizational resources and permissions, while attackers use similar tools for reconnaissance.
The brief
The episode opens by questioning why incident-response plans fail under pressure, then broadens the lens to election security, botnets, spyware, vulnerabilities, and major breaches.
Alex Weinert of Sempris explains that non-human identities include applications, provisioning scripts, and systems scanning email for sensitive information.
As low-code, no-code, and AI-assisted development spread, agentic systems may use resources or permissions unexpectedly, leaving organizations to manage behavior they do not fully see.
Attackers can also use agentic tools to inspect environments, find vulnerabilities, and accelerate reconnaissance, turning identity governance into an active security concern.
The closing investigation traces persistent political fundraising texts sent to the wrong recipient, revealing how campaign messaging systems can attach donations to another person’s number.
Weekly email
State of the Pod, every Monday
The week's most-discussed names and stories from podcasts, with the moments behind them. Free, five minutes, unsubscribe in one click.