Two-Thirds of Accounts Banned by Anthropic Were Linked to Cyberattack Preparation, Report Finds
AI cybersecurity
AI threats
Anthropic
cyberattacks
DeFi security
malware
smart contracts
zero-day exploit
Anthropic reports that 560 of 832 policy-violating accounts examined over the past year used AI to prepare for cyberattacks, with threat levels nearly doubling in the second half of the study period....
Jun 04, 2026