What we learned mapping a year's worth of AI-enabled cyber threats

Anthropic mapped 832 banned malicious accounts onto MITRE ATT&CK, a concrete read on how attackers now build with AI.