AI Incident Database: Anthropic accuses Chinese labs of trying to illicitly take Claude’s capabilities

Incident Summary

Anthropic accused three Chinese AI labs of attempting to illicitly extract Claude’s capabilities for use in their own models. Anthropic indicated this could potentially support offensive cyber operations. Limited public details are available about the specific methods used, the extent of any extraction, or whether the attempts succeeded. The incident highlights model misuse and security risks affecting the AI provider and downstream users who rely on Claude’s safeguards.

Incident Details

Type

Safety Incident

Severity

LOW

Status

ACTIVE

Date Occurred

April 6, 2026

External Sources

• https://cyberscoop.com/anthropic-accuses-chinese-labs-ai-distillation-cyber-risk/