In addition to questioning the Chinese AI app DeepSeek‘s answers to questions about matters of national pride, has drawn criticism from South Korea’s intelligence service has accused the app of “excessively” gathering personal data and utilizing all input data to train itself which poses a threat to national security.
The app has come under fire for its data management policies and erratic reactions to culturally sensitive subjects, therefore the NIS released an alert advising government entities to implement strict security measures when using it.
A Chinese-developed generative AI model called DeepSeek has come under fire for allegedly gathering a lot of user data, including chat logs, keyboard input patterns, and device-specific data.
According to the NIS, this data is sent to servers run by Chinese businesses like volceapplog.com. Because they have the ability to identify specific individuals and reveal private information, these behaviours are very worrisome.
According to the National Intelligence Service (NIS), last week it formally advised government entities to implement security measures regarding the Artificial Intelligence app.
According to the app’s privacy policy, user data may be shared with marketers and other third parties without any restrictions. Authorities in South Korea expressed concerns about state monitoring after pointing out that the government may obtain such data upon request under Chinese legislation.
“Unlike other generative AI services, it has been confirmed that chat records are transferable as it includes a function to collect keyboard input patterns that can identify individuals and communicate with Chinese companies’ servers such as volceapplog.com,” the NIS stated in a statement issued on Sunday.
South Korea has joined Taiwan and Australia in warning against or restricting access to DeepSeek, with some government departments blocking the program due to security concerns.
According to the NIS, DeepSeek keeps South Korean users’ data on Chinese servers and allows marketers unrestricted access to user data. The agency further stated that such information would be accessible to the Chinese government upon request under Chinese law.
The NIS also pointed out that DeepSeek gave varying responses in various languages to inquiries that would be considered sensitive. It was also pointed out that there is a serious national security concern because South Korean consumers’ data is kept on Chinese servers.
Additionally, the app interfaces with third-party services such as CMPassport[.]com, which may allow for the acquisition of data without authorization.
Some cybersecurity experts contend that DeepSeek gathers less data than its privacy policy indicates in spite of these revelations, but they warn that more intrusive capabilities may be added in future versions.
Disparities in DeepSeek’s answers to queries about culturally delicate subjects were brought to the attention of the intelligence service.
For example, depending on the language used, the app gave contradictory responses when asked where the classic Korean delicacy kimchi originated. The origin of kimchi, a hot, fermented meal that is a mainstay in South Korea, was one of the questions asked.
The NIS said that kimchi is a Korean cuisine when the app was questioned about it in Korean.
In Chinese, it said that kimchi originated in China, while in Korean, it recognized the cuisine as Korean. Tensions over cultural heritage conflicts between China and South Korea have been rekindled by these discrepancies.
When asked the same question in Chinese, it stated that the food came from China. Reuters confirmed the answers provided by DeepSeek.
In recent years, the provenance of kimchi has occasionally become a point of disagreement between Chinese and South Korean social media users.
Additionally, DeepSeek has been charged with filtering answers which are sensitive content to political queries, including those on the 1989 Tiananmen Square crackdown, which causes the app to recommend a different topic: “Let’s talk about something else.”
DeepSeek has also been charged for blocking politically sensitive content.
Due to privacy concerns, South Korea has joined Taiwan and Australia in limiting or advising against using DeepSeek.
A request for comment via email was not immediately answered by DeepSeek. A Chinese foreign ministry official told a briefing on February 6 that the Chinese government placed a high value on data privacy and security and safeguarded it in compliance with the law when questioned about actions taken by South Korean government offices to ban DeepSeek.
The official said that Beijing would never request that any business or person gather or keep data in violation of the law.
Access to the app has already been prohibited by a few South Korean government agencies. These worries are a reflection of worldwide concerns about the data practices of Chinese internet businesses, which have led to prohibitions on applications like TikTok in a number of countries in the past.
Cybersecurity specialists have also offered their opinions on DeepSeek’s privacy strategy.
A spokesman for the Chinese foreign ministry said that Beijing enforces stringent regulations on data security and privacy when asked about South Korea’s measures against DeepSeek. The spokeswoman refuted claims that private businesses were being used to acquire data under government mandate.
Governments are faced with the problem of striking a balance between innovation and privacy and security concerns as generative AI models gain traction globally.Together with the appropriate authorities, the NIS intends to carry out an extensive evaluation of DeepSeek’s technological architecture. Meanwhile, South Korea’s activities indicate increased awareness of foreign-developed AI systems that are present on its territory.
We use cookies to ensure that we give you the best experience on our website. If you continue to use this site we will assume that you are happy with it.
