Factbox-Governments and regulatory bodies intensify oversight of DeepSeek

Global Responses to DeepSeek's AI Technology

In January, Chinese artificial intelligence company DeepSeek announced it had created an AI model that could compete with ChatGPT, but at a significantly reduced cost. However, the company's security measures and privacy protocols have attracted attention and concern in several countries.

DeepSeek's privacy policy reveals that it retains a variety of personal information, including user queries and uploaded documents, on servers located in China.

International Actions Regarding DeepSeek

• Australia: In February, the Australian government prohibited the use of DeepSeek on all official devices, citing potential security threats.
• Czech Republic: The Czech authorities imposed a ban in July, preventing public sector employees from utilizing DeepSeek's services due to worries about data protection.
• France: France's data privacy regulator announced plans for early 2025 to question DeepSeek about its AI system and to assess any privacy implications for users.
• Germany: In June, Germany requested that Apple and Google remove DeepSeek from their app stores over data security concerns, according to a data protection official.
• India: At the start of February, India's finance ministry advised staff to refrain from using AI platforms such as ChatGPT and DeepSeek for government work, warning of risks to the confidentiality of official information.
• Italy: Italy's competition authority concluded an investigation into DeepSeek, which had been accused of not adequately informing users about the possibility of generating inaccurate content. The case was closed after DeepSeek agreed to binding commitments. In January 2025, the app was also blocked due to insufficient transparency regarding personal data usage.
• Netherlands: At the end of January, the Dutch privacy authority initiated an inquiry into DeepSeek's data handling practices and advised Dutch citizens to be cautious when using its software. Later, the government prohibited civil servants from using the app, referencing national policy on countries with offensive cyber capabilities.
• Russia: In early February, President Vladimir Putin directed Sberbank to partner with Chinese experts on collaborative AI projects, as reported by a senior executive at the bank.
• South Korea: In mid-February, South Korea's data protection agency halted new downloads of the DeepSeek app after the company admitted it had not fully complied with certain privacy regulations. Earlier that month, the industry ministry had temporarily restricted employee access to DeepSeek due to security issues, but the service was reinstated by late April.
• Taiwan: In February, Taiwan barred government agencies from using DeepSeek, citing security threats. Authorities also expressed concerns about potential censorship and the risk of data being transferred to China.
• United States: According to an April report by the New York Times, the U.S. administration is considering sanctions that would prevent DeepSeek from acquiring American technology and is discussing the possibility of restricting access to its services for U.S. users. In December, nine members of Congress urged the Secretary of Defense to add DeepSeek and other Chinese tech firms to a list of companies allegedly supporting the Chinese military. Additionally, in August, seven Republican senators requested the Commerce Department to investigate data security risks associated with Chinese open-source AI models like DeepSeek.

Compiled by Mateusz Rabiega, Paolo Laudani, and Tristan Veyet in Gdansk; Edited by Matt Scuffham