AI-powered Cyber-Attacks Up Significantly in the Last Year, Warns CrowdStrike

CrowdStrike's 2026 Global Threat Report warns of an 89% increase in AI-enabled attacks in 2025, with adversaries using LLMs and ML to scale and improve phishing, reconnaissance and malware development; examples include Russian APT activity (Fancy Bear's LameHug embedding LLM prompting) and criminal groups (Renaissance Spider) leveraging AI for more convincing phishing. The report notes AI is currently used to optimize existing methods rather than create wholly new vectors, and recommends stronger identity verification, AI-focused security awareness training, threat intelligence monitoring, and clear incident response planning.