South Korea bans new downloads of China's DeepSeek AI

South Korea's personal information protection regulator has announced a ban on new downloads of the Chinese artificial intelligence (AI) chatbot DeepSeek in South Korea. This measure aims to ensure that the AI model complies with South Korea's personal data protection laws. The regulator is actively working to safeguard user data privacy.

The agency stated that DeepSeek will only be reopened to South Korean users after "improvements and remedies" are made regarding personal data protection. Previously, DeepSeek garnered widespread global attention and quickly rose to the top of the South Korean app store charts in a short period, exceeding one million weekly active users. This rapid growth underscored its potential but also raised concerns.

DeepSeek's rapid popularity has also raised concerns in several countries around the world regarding privacy and national security, leading to restrictions being imposed on it. In addition to South Korea, Taiwan and Australia have also banned the use of the application on all government devices. The Australian government emphasized that the ban was not due to concerns about the application's Chinese background, but because it poses an "unacceptable risk" to national security. This highlights the broad range of security considerations.

Italian regulators have also taken similar measures, requiring DeepSeek to address concerns about its privacy policy before it can be relisted in app stores. Data protection authorities in France and Ireland have also inquired with DeepSeek about how it handles citizens' personal information, including whether it stores data on servers within China, as its privacy policy suggests. Furthermore, DeepSeek has stated that, like other generative AI tools, it may collect information such as email addresses and dates of birth, and use input prompts to improve its products. Such data handling practices are under increasing scrutiny.

In the United States, lawmakers have also proposed a bill to ban the use of DeepSeek on federal devices, citing monitoring risks. At the state government level, Texas, Virginia, and New York have already introduced similar regulations for employees. DeepSeek's "large language model" (LLM) has reasoning capabilities comparable to U.S. models such as OpenAI's o1, but is reportedly trained and operated at a fraction of the cost, raising questions about the billions of dollars being invested in AI infrastructure in the U.S. and elsewhere. This cost-effectiveness adds another layer to the debate surrounding AI development and deployment.