Anthropic says it caught China using Claude to develop an anti-torpedo weapon, tell me about this news in (https://www.hindustantimes.com/world-news/anthropic-says-it-caught-china-using-claude-to-develop-an-anti-torpedo-weapon-ai-101789099095684.html)
11-09-2026
According to the report published by Hindustan Times based on Anthropic’s Threat Intelligence findings, Anthropic detected and disrupted multiple instances of its AI model, Claude, being misused for military and state-sponsored operations:
- Anti-Torpedo Weapons Development: Anthropic revealed that actors in China utilized Claude to assist in developing software and guidance systems for an anti-torpedo weapon intended for the People’s Liberation Army (PLA) Navy. This was highlighted as part of a emerging category of AI misuse involving conventional weapons development.
- Broader Military Software Use: The company detailed other similar misuse cases across multiple countriesincluding Russia and Yemenwhere actors attempted to use AI models to write control and guidance code for drones, rockets, and missile systems.
- State Surveillance Operations: Beyond weapons development, Anthropic disclosed that it shut down several state-aligned surveillance campaigns. These operations leveraged Claude to analyze intelligence, monitor dissidents and ethnic minorities, and automate social media tracking.
- Anthropic’s Response: Anthropic terminated the associated accounts, reinforced safeguards across its models to block weapons-related and surveillance queries, and shared intelligence with security partners and relevant authorities.