Cybersecurity leader Bitdefender has recently released an annual report in which cybersecurity professionals detailed their most urgent concerns, key challenges, and threat perceptions shaping security.
Agentic AI, LLMs, and infrastructure breaches top the list of security concerns
The 2026 Cybersecurity Assessment Report is an independent survey of over 1,200 IT and security professionals. Those surveyed come from organizations with 500 or more employees across France, Germany, Italy, Singapore, the U.K., and the U.S.
“The expanding attack surface, the rapid proliferation of AI-powered threats, and persistent operational pressures are forcing organizations to rethink how they approach security from the ground up,” said Andrei Florescu, president and general manager of Bitdefender Business Solutions Group.
“The findings in this report make clear that modern security strategies must go beyond reactive defenses to continuously reduce risk, govern AI adoption, and ensure compliance across an environment where adversaries are faster, more adaptive, and increasingly automated.”
The report shows how seriously security leaders are thinking through AI-native risks and other key trends:
- 45 percent of respondents said internal AI systems and LLMs are their primary environment concerns, followed next by cloud infrastructure and application environments at 44 percent. Third was identity and access management (IAM) systems at 33.3 percent. Despite AI systems as respondents top security concern, 20.4 percent said employees leaking sensitive data into public LLMs as a low or extremely low risk.
- 41.8 percent of respondents reported cloud infrastructure or application breaches in the past 12 months, followed by BEC resulting in financial or data loss at 35.9 percent and ransomware at 25.6 percent. Among U.S. organizations, 54.7 percent reported BEC incidents, which is nearly 19 percentage points above the overall average. Further, 59.2 percent of all respondents confirmed experiencing AI-driven social engineering attacks in the past 12 months.
- 55.9 percent of those surveyed said attackers using AI to generate self-mutating malware is the top threat on the spectrum of AI-driven scenarios. Employees leaking sensitive data into public LLMs (53.5 percent), AI-driven evasion techniques bypassing traditional endpoint detection and response (EDR) signatures (52.5 percent), and deepfakes or voice cloning used in fraud or BEC (51.9 percent) were next up on the list of AI-related threats. Bitdefender notes that current threat intelligence suggests adversaries are using AI to accelerate and refine attacks as opposed to creating new malware. Additionally, agentic AI expands the attack surface emerging as a regional flashpoint – particularly in Singapore (64 percent) and the U.S. (61.6 percent).
What organizations need to address to stay secure
Overall, the report found that the strongest security postures may not belong to organizations with the most tools or the loudest AI strategy, but rather to those that reduce complexity, regain visibility, and focus on attack prevention.
Additional data in the report shows the challenges still facing security leaders as they address new threats:
- At 38 percent, high overhead in maintaining hardening rules and exceptions is the top barrier to reducing the attack surface. Followed next by fear of operational disruption (35.4 percent) and resource constraints (34.6 percent), an indicator that organizations understand the need to reduce exposure, but struggle to act without impacting operations. Further, 34.5 percent reported difficulty securing legacy systems and 33.8 percent reported visibility gaps in determining which legitimate tools are essential for each user. Additionally, 48.8 percent of U.S. organizations reported marked gaps in visibility compared to the overall average of 33.8 percent.
- 76.1 percent of respondents said that they would likely switch cybersecurity vendors due to concerns about data sovereignty, jurisdiction, or foreign government access to their data. The U.S. led at 87 percent, with the U.K. and Germany at 85 percent and 77 percent, respectively.
- Roughly 47 percent of those surveyed said that only partial or no visibility into individual Shadow AI tools or personal accounts used for work. While 51.8 percent said full visibility into sanctioned and unsanctioned AI usage, it’s different at the leadership level with 57.8 percent of managers believing they have full visibility compared to 45.9 percent of practitioners. Just 0.5 percent of managers report zero visibility versus 4.5 percent of practitioners. This indicates that leaders may be underestimating their organization’s true exposure.
“In many ways, the findings in this report reflect an industry actively adapting to fast-changing threats,” the report notes. “But the data also reveals a deeper tension. Complexity continues to outpace simplicity. Compliance often overshadows prevention. AI is reshaping both offense and defense. And throughout the report, one theme appears repeatedly: the gap between perception and operational reality.”
Bitdefender says that the organizations best positioned for the future will be those that reduce complexity, shrink their attack surface, align security with operational reality, and focus on outcomes over optics.