DeepSeek: Innovative, Yet Posing Privacy Risks in South Korea

The newly-made popular AI tool, DeepSeek AI, may pose national security risks in South Korea. However, should its usage be limited, or is this morally incorrect? (Photograph via Unsplash)

DeepSeek is a Chinese artificial intelligence startup company that has made shockwaves in the tech industry. The company’s latest artificial intelligence model, the R1, is similar in strength to OpenAI’s o1 model and was released at the end of 2024. However, Deepseek claimed to use only around 6 million dollars in training their AI model. In comparison, OpenAI has spent over 100 million dollars on its GPT-4 model. Many experts believe that Deepseek combined Nvidia A100 chips with other cheaper and less efficient chips, which helped them ultimately lower the cost of performing tasks.

Because of DeepSeek’s brilliant thinking, stock prices dropped significantly for tech firms that create high-end processing chips such as Nvidia and Nasdaq. With this innovative thinking, Deepseek has shown the world that high budgets and high-quality chips are not the only methods of advancing the artificial intelligence industry.[BBC]

However, even with all of these technological advancements, Deepseek has faced scrutiny over privacy concerns. On February 6th, 2025, Korea’s finance ministry revealed plans to ban DeepSeek from government computers that are connected to external networks. Other government agencies would soon follow as the foreign, military, trade, and environmental ministries also blocked access to the Chinese AI. [YONHAP]

The sudden removal of deepseek in government agencies is a result of reports from the NIS (National Intelligence Service) that show breaches of privacy and national security. The NIS has found that DeepSeek, unlike other AI chatbots like OpenAI’s Chat-GPT, could potentially be storing information including users’ chatting records or keystrokes. Deepseek also tends to avoid answering sensitive topics. For example, when asked about what happened in 1989 in Tiananmen Square, it will try to avoid the topic by asking if it could talk about something else. The NIS also claims that advertisers could have full access to user information. Users' information will be stored on Chinese servers which could be acquired by the Chinese government. South Korea will now be the third country to restrict and warn others about deepseek following Australia and Taiwan. [CPI]

The banning of Chinese corporations over security concerns is not a new situation. TikTok, another Chinese technology company, was also being restricted by the U.S. government for security reasons. However, the censoring of information could be a bigger problem than just privacy concerns. TikTok and DeepSeek are incredibly popular around the world and the censorship of crucial information like the Tiananmen Square Incident could be fatal. Because companies like DeepSeek are Chinese-owned they are subject to following Chinese regulations. Therefore the Chinese government could influence companies to moderate content that is deemed offensive or sensitive to China. As Chinese companies become more and more mainstream we could see important historical events and free speech becoming absent from the general public. Moreover, the Chinese government has the power to decide what is offensive and what isn’t.

In extreme cases, this censorship could lead to herd mentality, preventing the general public from thinking as individuals and being subject to what the majority or the government thinks. Restricting free speech is the same as suppressing human creativity, many famous innovative thinkers of the past thought outside of the box, breaking social and scientific norms. Chinese censorship hinders people from learning about the truth of history and suppresses diverse thinking. The restriction of DeepSeek in Korea could be a stepping stone in minimizing government censorship on the internet.