Cybersecurity professionals cite agentic AI, LLMs, and cloud infrastructure breaches as top concerns

As reported by Channel Insider, a recent annual report by Bitdefender highlights the most pressing concerns, challenges, and threat perceptions among IT and security professionals.

The 2026 Cybersecurity Assessment Report surveyed over 1,200 professionals across six countries. Findings indicate that internal AI systems and large language models (LLMs) are the primary environmental concerns for 45% of respondents, followed closely by cloud infrastructure and applications at 44%. In the past 12 months, 41.8% of respondents experienced cloud infrastructure or application breaches, with business email compromise (BEC) incidents affecting 35.9% and ransomware impacting 25.6%. Notably, U.S. organizations reported a higher incidence of BEC at 54.7%. AI-driven social engineering attacks were confirmed by 59.2% of all respondents.

The top AI-driven threat identified is attackers using AI for self-mutating malware (55.9%), followed by sensitive data leaks into public LLMs (53.5%) and AI-driven evasion techniques bypassing traditional endpoint detection and response (EDR) (52.5%). Agentic AI is also expanding the attack surface, particularly in Singapore and the U.S. Organizations face challenges in reducing their attack surface due to high overhead in maintaining hardening rules (38%), fear of operational disruption (35.4%), and resource constraints (34.6%). Furthermore, 76.1% of respondents would consider switching vendors due to data sovereignty concerns, and nearly half lack full visibility into shadow AI tools and personal work accounts.

Source: Channel Insider