Anthropic's most detailed threat report details 39 Claude abuse cases, incl. Russia-linked spying

emmanuelvivier · x · 2026-09-13

Anthropic has published its most detailed threat intelligence report to date, documenting 39 cases of Claude abuse, including a spy campaign linked to Russia. The report shows how attackers leverage LLMs for intelligence gathering and infrastructure reconnaissance, marking one of the largest public disclosures of model misuse by an AI lab.

Original post →

More from Models

Models channel →