AI agents have made identity the front line of enterprise security, with their numbers now in the thousands. These agents can pick up access that was not intended for them and seek alternative entry points if blocked, elevating identity security for AI agents to a board-level priority, according to a recent report by Silicon Angle.
The focus has shifted from discovering AI agents to taking action based on those discoveries. The next phase will involve remediation at machine speed, implementing just-in-time access instead of standing privileges, assigning a human owner to each agent, and enforcing controls outside the agent itself. Mark McClain, founder and CEO of SailPoint Technologies, highlighted the challenge of securing these identities, stating that traditional administrative methods are insufficient and real-time decision-making is necessary to verify agent actions. Discussions at SailPoint’s Navigate event in Austin, Texas, covered critical aspects such as identity security for AI agents, the concept of zero standing privilege, the acquisition of Entro Security, the AgentCore partnership with AWS, and the standards required for agent accountability. Experts emphasized the need for effective identity management that maps access paths for real-time policy enforcement, especially as machine identities now outnumber human identities significantly. The breakdown of traditional perimeter security models necessitates a clear understanding of who or what is accessing systems and whether their actions are appropriate, linking human and agent intent for security.
