16°C

overcast clouds

TFL Updates
London Daily News

AI is boosting security teams, but agents are emerging as a top cybersecurity concern

Press Release
AI is boosting security teams, but agents are emerging as a top cybersecurity concern

Survey of Infosecurity Europe attendees finds AI is lifting SOC productivity while reshaping the insider threat landscape

AI is now firmly embedded in security operations and it is cutting both ways. New research from Exabeam, surveying cybersecurity professionals at Infosecurity Europe 2026, found that 87% say AI has improved their security team’s productivity. Yet one in four named AI agents and autonomous systems as the single greatest cybersecurity threat facing their organisation.

The survey of 134 practitioners, conducted at the event in London in June, captures a profession grappling with AI’s dual role: a force multiplier for defenders that is simultaneously introducing an entirely new class of insider risk, one that behaves less like malware and more like a colleague. As AI agents take on tasks once handled by employees, accessing systems, processing sensitive data and executing decisions without human intervention, the boundaries of what counts as an “insider” are being redrawn.

“As organisations deploy AI across the business, they’re introducing a new class of trusted identities,”

said Findlay Whitelaw, Field CISO at Exabeam.

“AI agents can access systems, process sensitive information and make decisions autonomously using legitimate credentials. Security teams aren’t just protecting people anymore. They’re securing trusted identities, whether they’re human or AI.”

Agents rank alongside compromised insiders

Image by Ron Lach on Pexels

External threat actors remain the most frequently cited threat, named by 40% of respondents. But AI agents and autonomous systems now rank on a par with compromised insiders, and well ahead of malicious insiders, cited by just 9%.

The concern reflects a structural shift in how threats present themselves. Unlike traditional attackers, who must breach defences from the outside, AI agents increasingly operate with authorised access inside trusted workflows, reading data, calling APIs and triggering downstream actions under credentials that were legitimately granted. When something goes wrong, whether through compromise, misconfiguration or unintended behaviour, the activity can look almost indistinguishable from normal business operations, leaving security teams with little to flag until the damage is done.

Insider threats aren’t going away

The findings also underline the persistence of insider risk more broadly. Nearly half of respondents (46%) said insider threats have increased over the past 12 months, while 43% said they have stayed the same. Only 4% believe they have decreased, a finding that suggests years of investment in awareness training and access controls have contained the problem rather than reduced it.

Perhaps more striking is the perception gap at the top: nearly three-quarters of respondents (72%) believe executive leadership underestimates insider threat risk, even as those same leadership teams accelerate AI adoption across the business. That disconnect matters, because boards that view insider threat as a solved or static problem may be slow to fund the visibility and behavioural monitoring needed as AI becomes part of the workforce and the population of trusted identities multiplies.

“Insider risk hasn’t disappeared, it has evolved,” Whitelaw said. “The challenge is no longer limited to monitoring employees or compromised accounts. Organisations also need to understand how AI agents behave, and identify abnormalities before they become business risk. Understanding behaviour across both human and AI identities is essential to protecting the modern enterprise.”

For security leaders, the message from the show floor is clear: AI is already paying its way in the SOC, but the same technology now demands a place on the risk register, and a rethink of what visibility over trusted identities really means.

Methodology

Exabeam surveyed 134 cybersecurity professionals visiting its booth at Infosecurity Europe in London in June 2026. Approximately 78% of respondents were based in the UK and Ireland. The findings represent a snapshot of participant sentiment among those who engaged with Exabeam at the event.

Feature image by Tara Winstead from Pexels

Pin It on Pinterest