Excerpt from TechRadar Report, Published on November 9, 2025
As enterprises race to integrate Gen AI and AI Agents into daily operations, they are facing growing cybersecurity challenges. A new report by Proofpoint warns that these technologies, while revolutionary, are creating an entirely new class of insider risks — dubbed “agentic workspaces.”
According to the study, two in five organizations cited data loss via public or enterprise Gen AI tools as a top concern. Over a third of respondents also fear that sensitive corporate data may be used in AI model training, potentially exposing trade secrets or personal information.
AI Agents, often operating with high – level access privileges, amplify these risks. More than a third (38%) of organizations identified unsupervised data access by AI systems as a critical threat, while 54% admitted to lacking visibility and control over Gen AI deployments. This lack of oversight leaves digital environments vulnerable to accidental leaks and malicious exploitation.
Ryan Kalember, Chief Strategy Officer at Proofpoint, explained, “We’ve entered a new era of data security where insider threats, relentless data growth, and AI – driven change are testing the limits of traditional defenses.” He emphasized that future protection depends on unified, AI – powered solutions that adapt in real time and secure both human and agent activity.
Despite these emerging risks, humans remain the weakest link. Sixty – six percent of organizations attributed major data loss incidents to careless employees or contractors. Another third blamed malicious insiders or compromised accounts.
To manage this dual threat from humans and AI, Proofpoint recommends adopting behavioral cybersecurity and analytics. Research shows that just 1% of users are responsible for 76% of data loss events — underscoring the need for behavior – aware, adaptive defense strategies that evolve alongside Gen AI technologies.
To delve deeper into this topic, Visit TechRadar.
