Premier Cho Jung-Tai (卓榮泰) announced at a cabinet meeting on Thursday that the use of new Chinese AI model DeepSeek is prohibited throughout all of Taiwan’s government agencies. All agencies are instructed to restrict the use of any Chinese products or non-official software.
Cho cited concerns about the information security risk posed by DeepSeek, in addition to ideological censorship and the possibility of privacy infringement. This Wednesday, the Education Ministry indicated that public schools are also prohibited from purchasing or using Chinese information and communication products. Cho said that if public universities or research institutes need to use any Chinese software, hardware, or services, they can apply for approval per the Cabinet’s cybersecurity regulations.
The use of generative AI is also prohibited in the production of confidential documents by the Cabinet and affiliated agencies, and business contractors cannot provide generative AI with information that must be kept confidential or only disclosed upon consent of the agency. As for the private sector, Cho has asked the Digital Affairs Ministry and Science and Technology Ministry to evaluate risks and issue warnings accordingly.
President Lai Ching-te (賴清德) has pointed out that several other countries, among them the United States, France, Japan, South Korea, Italy, and Australia, have already banned the use of DeepSeek AI within their digital infrastructures or adopted risk control measures. Given the global demand for AI and the way China has positioned itself to spread their values through biased language models, Lai stresses the importance of speeding up development of domestic AI technology and linking with other democratic countries to resist totalitarianism.