Cybersecurity Today
Cybersecurity Today

Oct 31, 2026 · 58 min

AI agents force CISOs to rethink identity and control

Defend At AI Speed: With Special Guest Adam Ely, Check Point Software

As AI accelerates both attacks and enterprise adoption, security leaders must govern temporary agents without abandoning human judgment or basic defenses.

3 key takeaways
  1. 1AI agents create a new identity problem because their temporary access, rapid creation, and destruction complicate tracking and governance.
  2. 2Organizations should manage AI as untrusted technology, using practical experimentation and controls rather than trying to block adoption.
  3. 3Faster attacks intensify CISO pressure, making peer support, incident planning, segmentation, monitoring, and human judgment essential.

Don't miss

Adam Ely explains why rapidly created and destroyed AI agents require a different approach to identity, temporary access, tracking, and accountability.

The brief

Adam Ely, Check Point Software’s AI security leader, opens with a problem human identity systems were not built to handle: agents that appear, act, and disappear quickly.

A PHP-related crisis frames the CISO’s reality: a widely used technology can expand the attack surface overnight, while AI increases the speed and volume of response demands.

Ely argues that organizations should not try to stop AI adoption. They should treat it as untrusted, learn through hands-on use, and redesign work around practical safeguards.

The episode’s central tension is autonomy versus accountability: AI can analyze logs, handle alerts, and challenge incident decisions, but people still need to define limits and judge consequences.

The standout issue is ephemeral agent identity. Temporary access and rapidly changing agent populations make it harder to know what exists, what it can reach, and who is responsible.

The answer returns to basics: understand exposure and attack paths, segment systems, monitor activity, plan for incidents, and use AI to find gaps without surrendering human judgment.

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.

Listen to the full episode and explore every guest, topic, and moment on PodLume.

AI agents force CISOs to rethink identity and control · PodLume