The ‘first’ AI-run ransomware attack still needed a human

New details show an AI agent handled the technical execution of a real-world ransomware attack, reportedly the first known case of its kind. However, a human still selected the victim, prepared the infrastructure, and provided stolen credentials, making the attack only partially autonomous.

The case suggests AI is beginning to automate operational cybercrime tasks, but full end-to-end autonomy remains limited. The near-term risk is not independent AI hackers, but human attackers using AI agents to scale and accelerate intrusions.