TLDR:
- South Korea has suspended new downloads of Chinese AI chatbot DeepSeek due to privacy violations, while existing users can still access the service via web or previously downloaded apps
- The Personal Information Protection Commission (PIPC) identified issues with DeepSeek’s data processing policies and communication functions with third-party providers
- DeepSeek has gained over a million weekly users in South Korea before the ban, reaching top positions in app stores
- Multiple countries including Taiwan, Australia, and Italy have banned DeepSeek from government devices over privacy and security concerns
- DeepSeek’s AI model reportedly matches capabilities of US competitors like OpenAI’s o1 but at lower operational costs
South Korea has suspended new downloads of the Chinese artificial intelligence chatbot DeepSeek, citing privacy concerns and data protection violations. The suspension, which began on February 15, 2025, at 6:00 p.m. local time, affects both iOS and Android platforms but does not impact the web-based version of the service or existing app users.
The Personal Information Protection Commission (PIPC) of South Korea implemented the suspension after identifying problems with DeepSeek’s data processing policies and its communication functions with third-party service providers. The commission’s analysis, which began immediately after DeepSeek’s launch in the country, revealed that the platform did not fully comply with local privacy laws.
DeepSeek had experienced rapid growth in South Korea, attracting over a million weekly users and reaching top positions in app stores. The platform’s success, however, drew increased scrutiny from regulatory bodies, including the National Intelligence Service (NIS), which expressed concerns about excessive personal data collection and its use in AI system training.
Security Concerns
Technical investigations have uncovered security vulnerabilities in DeepSeek’s mobile applications. Both Android and iOS versions were found to transmit certain data to company servers without encryption, raising additional privacy concerns among security experts.
In response to the regulatory action, DeepSeek has appointed a local representative in South Korea and acknowledged its failure to adequately consider domestic privacy laws during the service’s launch. The company must now implement improvements to comply with the Personal Information Protection Act before downloads can resume.
The PIPC has advised existing users to exercise caution when using the platform, specifically recommending against entering personal information in the DeepSeek input window until the company announces final compliance measures.
The South Korean government’s response extends beyond the PIPC’s actions. Several government agencies have prohibited their employees from downloading the chatbot to work devices. Acting President Choi Sang-mok has described DeepSeek as a “shock” with potential impacts extending beyond the AI sector.
South Korea’s actions align with a broader international response to DeepSeek’s rapid expansion. Taiwan and Australia have implemented bans on the app’s use on government devices, with Australian officials stating their decision was based on “unacceptable risk” to national security rather than the company’s Chinese origins.
European regulators have also taken action. Italy’s data protection authority has temporarily banned DeepSeek, similar to its previous action against ChatGPT in 2023. French and Irish data protection authorities have requested clarification about DeepSeek’s handling of citizens’ personal information, particularly regarding data storage on Chinese servers.
The United States has seen both federal and state-level responses to DeepSeek’s expansion. Lawmakers have proposed legislation to ban the app from federal devices, citing surveillance concerns. States including Texas, Virginia, and New York have already implemented restrictions for government employees.
DeepSeek’s privacy policy indicates that, like other AI tools, it collects various types of personal information, including email addresses and birth dates. The company also uses input prompts to improve its service, a practice common among AI providers but one that has raised concerns about data handling and storage.
Despite these regulatory challenges, DeepSeek’s technical capabilities have garnered attention in the AI industry. The platform’s large language model (LLM) reportedly matches the reasoning capabilities of leading US models while operating at lower costs.
Beijing has responded to international concerns by stating that internet companies are welcome to operate in China as long as they follow local laws and regulations. Chinese officials have emphasized that they do not request companies or individuals to collect or store data in violation of laws.
The suspension in South Korea remains in effect as of February 17, 2025, with the PIPC working to prevent similar privacy compliance issues in future AI service launches. The commission plans to improve guidance for companies entering the Korean market.
DeepSeek users who downloaded the app before the suspension can continue using the service, and the web platform remains accessible to all users in South Korea.
