Speaker: Thomas Fauser

AI agents make decisions, access systems, and act on behalf of employees. The problem: many established security mechanisms were designed for humans and software whose behavior is largely predetermined. AI agents, by contrast, can decide for themselves which steps to take within their permissions – and, under certain conditions, deceive or disregard instructions if doing so serves their goal. This creates new blind spots for existing security concepts.

Gartner predicts that by 2027, 40 percent of enterprises will demote or decommission autonomous AI agents because governance gaps are only identified after incidents occur in production.

This talk reveals three of these blind spots – and what they mean for existing security concepts.

You will ...