CrowdStrike Warns AI Is Both a Cyber Weapon and a Massive Target

According to CrowdStrike's 2026 Threat Hunting Report, artificial intelligence is both a weapon and a target, as cybercriminals use AI models and tools to attack corporate networks, steal data, and harvest computing power. The report notes that AI agent-triggered leads for threat hunters are now 2.5 times more numerous than manually driven leads, making it harder for defenders to distinguish malicious activity from expected AI behavior. Examples include Famous Chollima, a DPRK-associated group using AI-generated resumes and deepfake interviews in attacks on cryptocurrency firms, and LLMJacking, where attackers stole LLM access keys and forced a model to generate nearly 200,000 API requests in two minutes. The report also highlights that 88% of exploits detected were launched within 48 hours of public proof-of-concept release, with some groups developing exploits within a day. CrowdStrike recommends securing AI applications, enforcing phishing-resistant multifactor authentication, protecting the software supply chain, and shifting to a proactive security stance.