China investigates DeepSeek and Moonshot AI over alleged data leaks to Anthropic
The Cyberspace Administration of China is investigating artificial intelligence firms DeepSeek and Moonshot AI after Anthropic accused the companies of secretly routing millions of user queries to Claude.
The Cyberspace Administration of China has opened a data security investigation into artificial intelligence startups DeepSeek and Moonshot AI, according to reports by The Information and other outlets. The regulator has reportedly questioned employees at both firms after Anthropic accused the companies of routing sensitive data and millions of user queries through its Claude models.[4][5][6][7]
Anthropic claimed that Moonshot routed more than 23 million exchanges to Claude between May and July, while DeepSeek routed more than 12.1 million exchanges in 14 days in July. The company alleged that user queries were forwarded without customer knowledge and returned as if processed by local models. The forwarded information reportedly included sensitive items such as surveillance footage near People's Liberation Army-linked facilities, police-system data, user credentials, and internal corporate information.[1][8]
Chinese authorities reportedly summoned seven companies named in an Anthropic threat report—including Alibaba, Xiaomi, SenseTime, MiniMax, and Zhipu—before focusing the probe on DeepSeek and Moonshot. The inquiry remains ongoing, with no penalties announced, and shares of Chinese AI developers dropped following reports of the probe.[1][2][3]
Key facts
- The Cyberspace Administration of China opened a data security investigation into DeepSeek and Moonshot AI following allegations from Anthropic.
- Anthropic accused Moonshot of routing over 23 million exchanges between May and July, and DeepSeek of routing more than 12.1 million exchanges over 14 days in July, through its Claude models.
- The data allegedly forwarded without user knowledge included internal corporate details, user credentials, police-system data, and surveillance footage near PLA-linked facilities.
- Chinese regulators questioned employees and summoned seven firms named in Anthropic's report before narrowing the investigation to DeepSeek and Moonshot.
- Shares of Chinese AI model developers declined following reports of the regulatory investigation.
- The probe remains ongoing, and Chinese regulators have not announced any penalties.
Sources · 7 sources
- WS
Wall St Engine@wallstenginePost on X ·
CHINA 🇨🇳 PROBES DEEPSEEK, MOONSHOT OVER ALLEGED SENSITIVE DATA LEAKS TO ANTHROPIC China’s internet regulator is investigating DeepSeek and Moonshot AI after Anthropic alleged the companies secretly routed customer requests to Claude, potentially sending sensitive Chinese data to U.S. servers, per The Information. Anthropic says Moonshot routed more than 23M exchanges to Claude between May and July, while DeepSeek routed more than 12.1M exchanges in just 14 days in July. The data allegedly included surveillance footage near PLA-linked facilities, internal Chinese corporate information, police-system data and user credentials, with customers not told their requests were being processed by Claude. China’s CAC reportedly summoned seven companies named in Anthropic’s report, including Alibaba, Xiaomi, SenseTime, MiniMax and Zhipu, before focusing its investigation on DeepSeek and Moonshot. For Beijing, the issue is now as much about cross-border data security as model copying, since China tightly restricts the transfer of sensitive data overseas. The probe remains ongoing and no penalties have been announced. Source: The Information,
Open source - BL
Bloomberg@businessPost on X ·
Shares of Chinese AI model developers fell after a report that the country’s regulators have opened a probe into startups DeepSeek and Moonshot AI over data security concerns https://t.co/hVaWGQ799l
Open source - QU
Quartz@qzPost on X ·
China is investigating DeepSeek and Moonshot over data that flowed to Anthropic: The Cyberspace Administration of China summoned all 7 companies named in Anthropic's threat report, then narrowed its focus to the 2 firms https://t.co/tWmLLhWc04 https://t.co/AHdG2KZLev
Open source - BL
Bloomberg@businessPost on X ·
China has opened a probe into data security at DeepSeek and Moonshot AI following Anthropic’s accusations that the companies had been routing sensitive data through its Claude models, The Information reported. https://t.co/hjNi94qQba
Open source - DE
DecryptArticle ·
China Probes DeepSeek and Moonshot Over Alleged Data Leaks to Anthropic's Claude China's internet regulator is investigating DeepSeek and Moonshot AI after Anthropic accused both of secretly routing millions of user exchanges through Claude to train their own models.
Open source - TE
Techmeme@TechmemePost on X ·
Sources: the Cyberspace Administration of China is investigating DeepSeek and Moonshot over alleged data leaks to Anthropic and has questioned employees (The Information) (Visit Techmeme dot com for the link and full context!)
Open source - *B
*Walter Bloomberg@DeItaonePost on X ·
CHINA PROBES DEEPSEEK, MOONSHOT OVER POTENTIAL DATA LEAKS TO ANTHROPIC - THE INFORMATION ANTHROPIC ALLEGED DEEPSEEK, MOONSHOT ROUTED SENSITIVE DATA TO CLAUDE- THE INFORMATION
Open source - TI
The Information@theinformationPost on X ·
China is investigating DeepSeek and Moonshot AI for potential data leaks to Anthropic. @jingyanghk explains the allegations: "Anthropic said that Moonshot and DeepSeek forwarded users' requests, including sensitive data such as police security camera footage, without their own customers', their own users' knowledge, and passed over the answers to their customers as if these were processed by their own models when, in fact, they were processed by Claude."
Open source