Key Takeaways
- AI is widely seen as a productivity booster: 88% of surveyed cybersecurity professionals report that AI has improved their security team’s efficiency.
- Yet AI‑driven agents are also viewed as a major risk: one‑in‑four (≈25%) consider AI agents and autonomous systems the single greatest cybersecurity threat to their organization.
- External threat actors remain the most frequently cited danger (40%), but AI agents rank on par with compromised insiders and ahead of malicious insiders (9%).
- Insider‑threat perceptions are worsening: 46% say insider threats have increased over the past year, while only 4% believe they have decreased.
- A large majority (72%) feel that executive leadership underestimates insider‑threat risk, indicating a gap between security teams and senior management.
- The survey, conducted at Infosecurity Europe 2026 with 134 professionals (≈78% UK/Ireland based), provides a snapshot of how practitioners are navigating the dual role of AI as both enhancer and emerging threat vector.
Overview of the Exabeam Survey Findings
Exabeam released results from a questionnaire administered to cybersecurity professionals who visited its booth at Infosecurity Europe 2026 in London. The poll captured the attitudes of 134 practitioners regarding the influence of artificial intelligence on security operations. By focusing on a real‑world event audience, the survey offers a timely pulse on how frontline defenders are experiencing AI’s integration into everyday workflows and threat landscapes. The data reveal a clear dichotomy: AI is simultaneously empowering teams and introducing new classes of risk that demand revised defensive strategies.
AI’s Positive Impact on Security Productivity
A striking 88% of respondents affirmed that AI has bolstered their security team’s productivity. This endorsement underscores the practical benefits defenders are observing—such as accelerated threat detection, automated triage of alerts, and enhanced correlation of disparate data sources. Professionals highlighted that AI‑driven analytics reduce the manual burden of sifting through massive volumes of log information, allowing analysts to concentrate on higher‑value investigations and proactive threat hunting. The high approval rate suggests that, for many organizations, AI tools have moved beyond experimental pilots to become integral components of the security stack.
AI Agents as a Rising Threat Vector
Despite the productivity gains, one in four participants identified AI agents and autonomous systems as the single greatest cybersecurity threat facing their organization. This perspective reflects growing concern that intelligent agents, once deployed, can operate with legitimate credentials and trusted access, thereby blurring the line between benign automation and malicious activity. Respondents warned that the autonomy of these systems enables them to make decisions, access sensitive data, and lateral‑move across networks without direct human oversight, creating a new class of insider‑like risk that traditional perimeter defenses struggle to contain.
Comparison with Traditional Threat Categories
When asked to rank threats, 40% of respondents cited external threat actors as the most frequent danger, maintaining the historical primacy of outside attackers. Notably, AI agents and autonomous systems were rated similarly to compromised insiders, placing them ahead of malicious insiders, which only 9% flagged as a top concern. This positioning indicates that security professionals view the risk posed by AI‑driven entities as comparable to the danger of legitimate accounts that have been hijacked, rather than to the threat of deliberate insider malfeasance. The nuance lies in the trusted nature of AI agents: they often possess authorized access, making anomalous behavior harder to distinguish from normal operations.
Evolution of Insider‑Threat Perceptions
The survey also shed light on how insider‑threat landscapes are shifting. 46% of respondents reported that insider threats have increased over the preceding 12 months, while 43% said the level has remained unchanged, and only 4% perceived a decrease. These figures suggest that, contrary to narratives of declining insider risk, many organizations are experiencing a rise—or at least a stagnation—in internal security challenges. The persistence of insider concerns highlights the need for continual vigilance, even as external threats continue to dominate headlines.
Leadership’s Underestimation of Insider Risk
A significant disconnect emerged regarding executive awareness: 72% of participants believe that leadership underestimates the risk posed by insider threats. This perception points to a potential gap in communication or prioritization between security teams and senior management. When leaders downplay insider dangers, organizations may under‑invest in the monitoring, analytics, and governance capabilities necessary to detect subtle abusive behaviors—whether originating from compromised credentials, negligent employees, or autonomous AI agents. Bridging this gap is essential for aligning security investments with the actual threat landscape.
The Dual Identity Challenge: Human and AI
Findlay Whitelaw, Field CISO at Exabeam, emphasized that AI’s proliferation introduces a “new class of trusted identities.” Unlike human users, AI agents can authenticate, process sensitive information, and execute decisions autonomously while still operating under legitimate credentials. Consequently, security teams must broaden their protective focus beyond people to encompass these machine‑based identities. Whitelaw argued that effective defense now requires continuous behavioral baselining for both human and AI actors, enabling the detection of deviations that may signal misuse before they evolve into tangible business risk.
Methodological Context
The data stem from a straightforward intercept survey conducted at Exabeam’s booth during Infosecurity Europe 2026. A total of 134 cybersecurity professionals participated, with roughly 78% hailing from the United Kingdom and Ireland. The authors caution that the findings represent a snapshot of the sentiments of those who engaged with Exabeam at the event, rather than a statistically representative sample of the global security community. Nevertheless, the concentrated expertise of the respondents offers valuable insight into prevailing attitudes among practitioners who are actively evaluating AI‑enabled security solutions at a major industry gathering.
Implications for Security Strategy
Taken together, the survey results paint a nuanced picture: AI is undeniably enhancing operational efficiency, yet it simultaneously expands the attack surface by creating trusted, autonomous entities that can masquerade as legitimate users. Organizations must therefore adopt a balanced approach—leveraging AI for detection and response while instituting robust identity‑centric controls, continuous behavior monitoring, and clear governance over AI agent deployment. Additionally, raising executive awareness of insider‑threat evolution—particularly the emerging risks from AI agents—will be critical to securing appropriate resources and fostering a culture of shared responsibility for security across the enterprise.