AI Cyberattacks: Autonomous Intrusions & Evolving Threats

AI is now actively carrying out cyberattacks with minimal human input. Researchers documented intrusions where AI ran exploitation workflows autonomously, generating thousands of commands across dozens of sessions. This comes from Check Point’s AI Security Report 2026. Attackers are orchestrating AI across multiple stages of an attack without needing human intervention. They achieve this by getting capable AI models and removing their safety controls. Attackers gain AI capabilities through abusing commercial models, using stolen AI credentials, or purchasing access to AI tools built for cybercrime. AI is also widely used to develop malware, allowing less experienced attackers to create more capable tools. A growing category of malware even communicates with AI models during execution to generate commands or adapt its behavior in real-time. The expertise barrier separating capable attackers is disappearing, posing new challenges for cybersecurity.