AI-powered cybercrime is getting easier

Axios says researchers are seeing cybercriminals use AI more practically as access costs fall and model capabilities improve. Recent examples include the JadePuffer ransomware case, where an AI agent rewrote exploit code in seconds, a cloud attack compressed into 72 hours, and AI-generated malware used in bank fraud targeting Mexican financial organizations. Researchers say early attacks still make mistakes, but the window for defenders to prepare is shrinking.

Key details: Researchers told Axios AI-assisted attacks are moving beyond experiments, JadePuffer used an AI agent to rewrite exploit code in 31 seconds, Other reports cited AI-assisted cloud attacks and bank-fraud malware.

Why it matters: The AI security story is no longer theoretical: attackers are starting to use agents and model-generated code inside real intrusions.