Back to today’s briefing

Story perspectives

Fraudulent Labs Access Claude: Anthropic Alerts US Authorities

2/23/2026

39 5 Full Breakdown

Summary

Story summary

Anthropic says three China-based labs—DeepSeek, Moonshot AI, and MiniMax—used fraudulent accounts to access Claude, generating over 16 million exchanges in a distillation campaign.
Jacob Klein, Anthropic's head of threat intelligence, says the distillation gains were substantial.
Anthropic has shared findings with United States authorities and urges protecting AI progress amid United States–China competition due to safety concerns.

1 of 1

Story summary

Anthropic says three China-based labs—DeepSeek, Moonshot AI, and MiniMax—used fraudulent accounts to access Claude, generating over 16 million exchanges in a distillation campaign.
Jacob Klein, Anthropic's head of threat intelligence, says the distillation gains were substantial.
Anthropic has shared findings with United States authorities and urges protecting AI progress amid United States–China competition due to safety concerns.