AI agents pose insider risks due to unsupervised access and lack of visibility controls, the report claims. 66% of major data loss events are caused by negligent employees or third-party contractors. Proofpoint recommends behaviorally aware and adaptive security to protect both human and AI activities
As companies rush to deploy generative artificial intelligence (Gen AI) and AI agents, they face the same challenges they face when deploying new technologies quickly: the risk of data leaks, leaks, and breaches.
A new Proofpoint report argues that “agency workspaces” represent a new type of insider risk, rivaling human error in severity. Two in five responding organizations said data loss through public or corporate GenAI tools is their top concern, and more than a third said they were concerned about sensitive data being used in AI training.
The report further explains that AI agents often operate as privileged superusers, which only makes the situation worse. More than a third (38%) reported unmonitored data access by AI agents as a “major threat” and 54% said they lacked sufficient visibility and control over Gen AI tools. This means that AI agents are left to their own devices, which causes problems.
you may like
test the limits
“We are entering a new era of data security, where insider threats, relentless data growth, and AI-driven change are testing the limits of traditional defenses,” said Ryan Kalember, chief strategy officer at Proofpoint.
“Fragmented tools and limited visibility are putting organizations at risk. The future of data protection depends on integrated, AI-powered solutions that understand content and context, adapt in real-time, and protect information across both human and agent activities.”
Despite AI being a problem, humans remain the weakest link in this cybersecurity chain. Two-thirds (66%) of organizations blame their most severe data loss events on “negligent employees” and third-party contractors, while 31% cite compromised users. A third (33%) pointed to malicious insiders.
To mitigate threats from both humans and AI agents, Proofpoint advises pivoting to behavioral cybersecurity and analytics. “Just 1% of users are involved in 76% of data loss events, highlighting the importance of a behaviorally aware and adaptive security strategy.”
Apparently, two-thirds (65%) have already implemented AI-enhanced data security capabilities.
The best antivirus tool for every budget
Our top picks based on real-world tests and comparisons
Follow TechRadar on Google News and add it as your preferred source to get expert news, reviews, and opinions in your feed. Be sure to click the follow button!
Of course, you can also follow TechRadar on TikTok for news, reviews, and unboxings in video form, and get regular updates on WhatsApp.
