Proofpoint’s CEO Discusses the Urgent Need for AI Integrity Frameworks
In the rapidly advancing realm of artificial intelligence, the distinction between human and machine behavior has blurred, especially concerning security risks. Sumit Dhawan, the CEO of Proofpoint, emphasizes that AI agents now mimic human actions, representing a similar risk profile. This assertion raises critical questions regarding the governance of AI behavior. During a recent interview, Dhawan highlighted that AI operates in a non-deterministic manner and is susceptible to manipulation through techniques such as prompt engineering. As a result, the establishment of a tailored integrity framework to oversee AI agents is essential.
Traditional security protocols were predominantly developed to address predictable, Boolean logic. AI agents, however, defy these conventional patterns, which necessitates a shift towards behavioral drift detection as the primary defense model. Dhawan draws an intriguing parallel to enterprise insider risk programs, which monitor human behavior for deviations from expected norms. In the same manner, AI agents require a robust mechanism to detect and respond to anomalous behaviors.
“AI lacks a universal code of conduct,” Dhawan stated. “It does not inherently possess integrity; thus, we must integrate this into a technological layer—essentially creating an AI behavior safeguard layer.” This perspective not only underscores the need for ongoing vigilance in the AI landscape but also emphasizes a proactive approach in developing frameworks capable of managing AI’s unpredictable nature.
During the interview conducted by Information Security Media Group at the RSAC Conference 2026, Dhawan further discussed the evolving roles of Chief Information Security Officers (CISOs) in response to the mounting pressures from AI-driven threats. He noted a burgeoning divide among CISOs, with some opting for proactive measures to implement AI safeguards, while others adopt a more wait-and-see approach. This bifurcation reflects the urgency of addressing the complexities introduced by AI technologies.
Moreover, Dhawan addressed the importance of evolving cybersecurity measures. The rise of AI-driven threats has pushed cybersecurity vendors to transition from traditional machine learning methods to more sophisticated language model-based detection systems. This shift illustrates the necessity of adapting to a landscape where AI tools not only pose risks but also necessitate innovative countermeasures.
Proofpoint’s AI security platform is designed to extend its existing human insider risk model to encompass AI agents. This evolution signifies a broader trend in the industry, reflecting the integrative strategies that organizations must adopt to contend with the challenges posed by AI systems. By bridging the gap between human-centered security approaches and AI-driven threats, Proofpoint aims to enhance its security framework and better protect organizations.
Dhawan, who spearheads Proofpoint’s human-centric security strategy, brings a wealth of experience from his previous roles at VMware, Instart, and Citrix. His background in scaling enterprise software businesses and driving transformative marketing strategies positions him uniquely to tackle the multifaceted challenges that today’s cybersecurity landscape presents.
As organizations increasingly rely on AI technology, Dhawan’s insights highlight a pressing need for businesses to cultivate a culture of security awareness that encompasses both human and machine elements. The goal is not merely to respond to existing threats but to build a comprehensive understanding of how AI agents can be managed and secured effectively.
In conclusion, Dhawan’s emphasis on establishing integrity frameworks for AI agents aligns with a larger narrative in cybersecurity. As AI capabilities expand, so too do the potential risks—underscoring the urgency for organizations to proactively develop robust security measures. By fostering a collaborative approach between human and machine behaviors, businesses can better navigate the complexities of the digital landscape and safeguard their assets against increasingly sophisticated threats. As this dialogue continues, it becomes clear that the future of security will hinge on the integration of human-centric principles with innovative AI governance practices.