It was great to speak at the Commonwealth Club for Zenity's Agentic Security Summit last week

It was great to speak at the Commonwealth Club for Zenity's Agentic Security Summit last week. Loads of informative talks from leaders in the space about what people should _really_ be worrying about when it comes to AI and agents.

Loved Johann Rehberger's continual pushing of the boundaries to show how these systems can be exploited (and by proxy, how you can protect them).

Jack Cable gave a great talk on vulns he's discovered and highlighted why you shouldn't threaten legal action against researchers trying to help (strange we still need to say that in 2025...)

Ben Nassi, PhD had a timeline of AI attacks that just blows you away when you realize what's been compressed into the last 3 years.

Ken Huang as always bringing new work forward, this time introducing the OWASP AIVSS Project for scoring AI based vulnerabilities.

The OG Steve Wilson had a thoughtful presentation on how we need to consider agents as insider threats. Really fast, sneaky ones that lack conscience and accountability.

Allie Howe also had a great presentation at the end showing how to truly secure (and exploit) agentic systems, you *gasp* actually need to understand what's going on under the hood a bit.

Overall, a really fabulous event, thanks Michael Bargury and the rest of the team at Zenity for hosting it!