South Korea Flags Chinese AI App DeepSeek Over Data

South Korea’s intelligence agency has raised alarms over the Chinese artificial intelligence app DeepSeek, citing concerns about excessive data collection and national security risks. The National Intelligence Service (NIS) recently issued a formal advisory to government agencies, urging them to implement security precautions when dealing with the app, per Reuters.

According to Reuters, the NIS highlighted that DeepSeek differs from other generative AI platforms by not only storing user chat records but also collecting keyboard input patterns, which could potentially identify individuals. The agency warned that this data is transmitted to Chinese servers linked to companies such as volceapplog.com.

As a result of these concerns, several South Korean government ministries have taken steps to block access to the app. South Korea now joins Australia and Taiwan in restricting or warning against DeepSeek due to security risks, according to Reuters. The intelligence agency further alleged that DeepSeek grants advertisers unrestricted access to user data while also storing South Korean users’ information on Chinese servers. The NIS pointed out that under Chinese law, the government has the authority to access such data upon request.

Read more: DeepSeek Faces Global Scrutiny as Governments Ban AI Model Over Security Concerns

In addition to security concerns, the NIS also flagged inconsistencies in the app’s responses to culturally and politically sensitive topics. The agency noted that DeepSeek provides different answers depending on the language used. For instance, when asked about the origin of kimchi in Korean, the app affirmed that the dish is Korean. However, when posed the same question in Chinese, it claimed that kimchi originated in China. Reuters corroborated these discrepancies in DeepSeek’s responses.

The debate over kimchi’s origins has long been a contentious issue between South Koreans and Chinese social media users. Such discrepancies in AI-generated responses have only added fuel to the ongoing cultural dispute, per Reuters.

Beyond cultural questions, the NIS also accused DeepSeek of censoring responses related to politically sensitive events. When asked about the 1989 Tiananmen Square crackdown, the app reportedly suggested shifting the conversation by responding, “Let’s talk about something else.”

Source: Reuters