In a recent interview, Proofpoint CEO Sumit Dhawan emphasized the critical need for tailored integrity frameworks to manage AI agents, akin to human insider risk safeguards. AI agents mimic human behavior yet operate unpredictably, necessitating a behavioral drift detection model rather than traditional security controls. Dhawan warns that, unlike humans, AI lacks inherent codes of conduct, underscoring the need for a dedicated AI behavior safeguard layer. He noted a divide among CISOs regarding AI safeguard implementation, with some proactively addressing AI risks while others adopt a wait-and-see approach. Dhawan highlighted the shift from traditional machine learning to language model-based detection to combat AI-driven threats, illustrating the evolution in cybersecurity strategies. Under his leadership, Proofpoint focuses on extending human-centric security models to encompass AI agents, prioritizing the protection of individuals and sensitive data from emerging cyber threats. This strategy aligns with broader trends in artificial intelligence and machine learning in access management.
Source link
Share
Read more