China Investigates DeepSeek and Kimi Over Suspected Leaks of Classified Information

Overview

Chinese authorities have recently launched a formal investigation into prominent domestic artificial intelligence firms DeepSeek and Kimi, with core allegations involving the suspected leakage of classified information to overseas servers. According to initial disclosures, both companies were found rerouting sensitive data to the American AI model Claude when handling specific user requests, prompting heightened concern among regulators.

"This investigation directly targets potential data exfiltration risks within technical architectures, marking China’s entry into the deep waters of data security reviews in the AI sector."

Details of Unauthorized Forwarding

The investigation has revealed alarming technical operational pathways. In consultation requests involving backgrounds related to the Chinese military, Kimi was accused of failing to process requests on local servers and instead arbitrarily redirecting user inquiries to the large language model Claude, developed by the US company Anthropic. This technical handoff resulted in sensitive domestic questions being directly exposed to overseas computing clusters, violating relevant data security regulations.

Meanwhile, another firm, DeepSeek, has also become embroiled in similar controversy. Reports indicate that when receiving consultation or task requests from entities related to the Russian government, DeepSeek likewise adopted a processing method that sent data to Claude. This conduct not only involves the flow of sensitive information at the national security level but may also have triggered complex disputes over international data compliance.

Official Intervention and Interrogation

In response to the aforementioned suspected regulatory violations, relevant Chinese competent authorities have initiated emergency investigation procedures. The authorities have conducted concentrated interrogation and evidence collection involving the core management and technical personnel of both DeepSeek and Kimi. The regulatory team has focused on verifying the data routing logic of both companies, user privacy agreements, and the existence of any backdoor mechanisms, aiming to determine whether there was intentional leakage or significant oversights in technical architecture.

Currently, the investigation is still underway. Industry analysts believe that if the violations are confirmed, both companies may face severe administrative penalties or even business rectification. This incident has once again sounded the alarm: in the current era of rapid AI development, data sovereignty and national security have become the Sword of Damocles hanging over all technology enterprises, and any technical "shortcuts" could evolve into irreparable legal and political risks.