Agentic AI Security Roundtable

The Agentic AI Security Roundtable will open a practical debate on how to safely deploy AI agents into the real operations of organizations. What approaches and permissions can we grant them? How can we control their behavior and prevent unwanted actions? What happens when an agent makes a mistake, falls for manipulation, or gains access to information and systems they shouldn't have? We will discuss topics such as identity and access management, working with sensitive data, prompt injection and other attacks, monitoring and auditability, human-in-the-loop mechanisms, and responsibility for autonomously performed actions. Above all, we will seek the balance between the autonomy that makes agents useful and the control that ensures they are safe. The roundtable is built on open expert discussion and sharing experiences among participants, rather than on classical presentations. For CISOs, CIOs, and CTOs, cybersecurity specialists, AI and IT architects, AI and automation leaders, risk and compliance professionals, developers of agent systems, and other experts responsible for the safe deployment of AI in organizations.