Anthropic claims Claude AI used for missile projects, global espionage
AI was used to develop missile guidance software in Yemen and to power cyber operations, a report alleges.
Anthropic has accused Moonshot AI of secretly routing customer requests intended for its Kimi chatbot through fraudulent accounts to Anthropic's Claude models. The US artificial intelligence developer discovered that Moonshot concealed this practice, presenting Claude-generated answers as its own. Moonshot allegedly used 5,380 fraudulent accounts to send around 300,000 requests through a network to Claude Opus, part of a broader effort to access technology unavailable in China.
Anthropic attributed over 23 million Claude exchanges to Moonshot between May and July, alleging the company used intermediaries and numerous accounts to evade restrictions and detection systems. This alleged activity is detailed in Anthropic's latest threat intelligence report, covering interventions from December 2025 to August 2026.
Moonshot did not release an immediate public response to these allegations. Other China-based AI labs, including Moonshot, DeepSeek, and MiniMax, have also been accused of attempting unauthorized distillation or routing live traffic to Anthropic's models. Distillation is an AI training technique where a smaller model learns from outputs of a larger system, but Anthropic considers it improper when competitors misappropriate outputs in violation of access restrictions or service terms.
Written by urgent.news from Arabian Post's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
This story
This is one outlet's version. Read the fullest account.
- Anthropic accuses Moonshot over concealed Claude routing thearabianpost.com
- Anthropic disrupts Russian, Chinese AI campaigns targeting its Claude models japantimes.co.jp
