AI Risks Unpacked at PBC Connect

After a full day of conversations at PBC Connect – RSAC 2026, Dana and Mark regroup to recap the prevalent ideas from onstage, in the audience, and between sessions.

We revisit the launch of the State of AI Risk Management 2026 report, the growing gap between visibility and control, and the reality that many teams still feel overwhelmed by fragmented tooling and slow mobilization. The hosts also return to a warning echoed throughout the day: a surge of AI-generated vulnerabilities is coming, and many organizations are still working to understand what that means in practice.

From agent behavior and non-determinism to prompt injection, model poisoning, and the idea of presuming breach, the conversation connects the threads from the day’s panels into a broader look at how security teams are preparing for AI systems that are already here.

Stay tuned for the session recording of this Purple Book Community panel – coming soon.