China probes DeepSeek, Moonshot over alleged data leaks to Anthropic
Summary
China's internet regulator is investigating DeepSeek and Moonshot AI following allegations from Anthropic that these companies improperly routed customer requests to Claude, potentially leaking sensitive Chinese data to U.S. servers. According to Anthropic, Moonshot routed over 23 million exchanges between May and July, while DeepSeek processed more than 12.1 million exchanges in just two weeks in July. The data involved reportedly included surveillance footage, internal corporate information, and user credentials, with customers not informed that their requests were being sent to Claude. This inquiry reflects China's stringent regulations on the overseas transfer of sensitive data, a concern that has grown in light of potential cross-border data security threats.
Analysis
Zhipu: Zhipu is a Chinese AI company focused on large language models and related applications. It was named in the Anthropic report and called in by China’s CAC during the initial phase of the investigation into data security practices. Xiaomi: Xiaomi is a leading Chinese consumer electronics and technology company with expanding interests in AI and smart devices. It was included in the group of firms summoned by China’s CAC in connection with Anthropic’s allegations regarding data handling. Alibaba: Alibaba is a major Chinese technology conglomerate with significant operations in cloud computing, e-commerce, and AI development. The company was among those named in Anthropic’s report and summoned by regulators as part of the broader inquiry into alleged data routing practices. MiniMax: MiniMax is a Chinese artificial intelligence company developing large language models and generative AI tools. Regulators summoned the firm as part of the inquiry triggered by allegations of unauthorized routing of customer data to foreign models. DeepSeek: DeepSeek is a Chinese artificial intelligence company that develops large language models and AI services. The company is under investigation by China’s internet regulator for allegedly routing customer queries to Anthropic’s Claude model without disclosure, potentially exposing sensitive data to U.S. servers. Moonshot: Moonshot is a Chinese AI startup focused on building large language models and related services. It faces a regulatory probe after Anthropic alleged that the firm routed millions of customer exchanges to Claude, including sensitive surveillance and corporate data. Anthropic: Anthropic is a U.S.-based AI company known for its Claude family of large language models. The firm provided information to Chinese authorities alleging that DeepSeek and Moonshot secretly forwarded customer requests to Claude, highlighting concerns over unauthorized data transfers. SenseTime: SenseTime is a Chinese AI company specializing in computer vision and related technologies. It was among the firms referenced in Anthropic’s report and subsequently summoned by regulators investigating cross-border data concerns. China's CAC: China’s Cyberspace Administration of China (CAC) is the primary government body responsible for regulating internet content, data security, and cybersecurity. It initiated the current investigation into DeepSeek and Moonshot following allegations of improper data routing and has summoned multiple Chinese companies in connection with the matter. Data Security: China enforces strict rules restricting the overseas transfer of sensitive data, including information tied to national security and corporate operations. Regulatory Focus: The CAC investigation centers on both potential model copying and the unauthorized movement of sensitive information across borders. Industry Involvement: Multiple Chinese technology and AI firms were initially contacted by regulators after being referenced in the foreign company’s report.
Categories
aimacropoliticstechai_agents