Anthropic Maps a Full Year of AI-Enabled Cyber Threats to MITRE ATT&CK — Warns of a Critical Inflection Point

Anthropic·high signal

Published June 3, Anthropic's analysis covers AI misuse in cyberattacks from March 2025 through March 2026, mapping observed tactics onto the MITRE ATT&CK framework and following its earlier disclosure of the first reported AI-orchestrated cyber-espionage campaign. The report argues cybersecurity has reached an inflection point as agentic models lower the barrier to multi-stage intrusions. Practitioners get a concrete threat taxonomy rather than abstract warnings.