February 14, 2025 - Last month, we published an article warning about the privacy risks associated with DeepSeek, a China-based artificial intelligence chatbot. That article, titled Use of DeepSeek May Be the Best Way to Become a Victim of Identity Theft, outlined how DeepSeek collects excessive user data and poses a serious threat to personal privacy.

Now, South Korea’s National Intelligence Service (NIS) has released a security bulletin that not only confirms the dangers we previously reported but also reveals even more alarming details about DeepSeek’s operations. Their findings paint a picture of a platform deeply embedded in surveillance and data exploitation, with ties to the Chinese government, meaning the Chinese Communist Party, and commercial entities.

The NIS has issued a warning to government agencies about DeepSeek’s data collection practices, highlighting the following concerns:

1. Excessive Data Harvesting – Unlike many other AI services, DeepSeek does not just log chat history; it collects keyboard input patterns that can be used to uniquely identify individuals. This means that even anonymized data could potentially be linked back to specific users.
2. Storage on Chinese Servers – Data from South Korean users, an presumably users in other nations as well, is stored on Chinese servers, where the Chinese Communist Party (CCP) has legal access to all stored information upon request. This significantly raises concerns about government surveillance and foreign data exploitation.
3. No Opt-Out for Data Sharing – DeepSeek’s terms and conditions allow the company to retain user data indefinitely and share it with third-party advertisers without any opt-out option for users. This contradicts common privacy standards and leaves users powerless to control their own information. Notably, this also means that DeepSeek is operation well outside of laws in some states that specifically require companies to provide opt-out options for data sharing and tracking.
4. Propaganda & Misinformation – DeepSeek’s responses vary depending on the language used, pushing Chinese Communist Party narratives on historical and political topics. For example, when asked about the origins of kimchi, DeepSeek tells Korean users that it is a significant Korean dish, but in Chinese, it claims that kimchi originated in China. Likewise, inquiries in English receive a neutral response.
5. Steep Decline in South Korean Usage – Due to these privacy concerns, the number of Korean users has dropped by over 50% in the past week. At its peak on January 28, DeepSeek had 191,000 users in South Korea, but after a warning from the Korean Interior Ministry, that number has plummeted to just 74,000.
6. Bans by Government and Businesses – In response to these revelations, multiple South Korean government agencies, banks, schools, and large corporations have banned DeepSeek from their networks, reflecting growing fears over national security and personal privacy.

If you are considering using DeepSeek, these latest findings make it clear that you should think twice. The AI is not just collecting casual chat history; it is monitoring how you type, storing that data indefinitely, and making it available to foreign entities with no oversight.

With no transparency in its data retention policies and clear evidence of propaganda manipulation, DeepSeek represents a severe risk to personal privacy and digital security. The risks of identity theft, government surveillance, and data exploitation are simply too high.

For those who value their privacy, it is advisable to avoid DeepSeek entirely and use AI platforms with clearer policies on data protection and user control.

As new information continues to emerge, we’ll keep a close watch on DeepSeek and other AI platforms that pose privacy risks. Stay informed and stay safe online.

by Jim Malmberg

Note: When posting a comment, please sign-in first if you want a response. If you are not registered, click here. Registration is easy and free.

Follow ACCESS