China's Ministry of Foreign Affairs has refuted allegations that the country encourages businesses or individuals to collect or store users' data through illegal means. The statement came during a regular press briefing on Thursday, addressing concerns surrounding DeepSeek, a Chinese artificial intelligence (AI) service that recently topped the U.S. iPhone app store charts.
Spokesperson Guo Jiakun emphasized that China prioritizes data privacy and security, operating within legal frameworks. He criticized the overextension of national security concerns and the politicization of trade, economic, and technological issues. Guo also reiterated China's commitment to protecting the legitimate rights and interests of its companies globally.
DeepSeek, known for its advanced performance and low training costs, has adopted an open approach to AI model development. Its latest models, including V3, R1, and Janus Pro, are freely available for download. The company has also published research papers detailing training methods, enabling other developers to replicate them using their own datasets.
DeepSeek's models operate locally without requiring an internet connection, preventing the transmission of user data to third parties. This feature distinguishes DeepSeek from closed models developed by companies like OpenAI and Google. Venture capitalist Marc Andreessen described DeepSeek-R1 as "a profound gift to the world."
DeepSeek's success has prompted other AI developers to adopt more open strategies. Following DeepSeek's rise, OpenAI announced free access to its ChatGPT search functions, eliminating the need for user sign-up. Alibaba, a leading Chinese internet company, unveiled its Qwen2.5-Max model, trained using methods similar to DeepSeek-V3, and launched a web service offering free access to its models.