DeepSeek and privacy: Why Korea hit pause on China's AI disruptor

By Chyung Eun-ju and Joel Cho

When DeepSeek burst into the global AI scene earlier this year, it shook up not just the tech world, but also the global market. However, just as it has been a lurking issue with other AI technologies, it sparked concerns over data privacy, security and regulatory compliance. Developed by the Chinese startup founded by Liang Wenfeng, DeepSeek’s chatbot gained, almost overnight, a massive international user base, rising to the top of app stores and drawing comparisons to OpenAI’s ChatGPT.

Its sudden rise, however, also came with a price, as various countries held regulatory concerns over this new platform. Korea has recently fallen into one of the countries that have put DeepSeek under regulatory scrutiny, suspending new downloads due to concerns over how it processes user data.

We believe that DeepSeek’s success and broad public acceptance were driven by its ability to deliver high-performance AI at a cost significantly lower than its Western competitors. But, just as has occurred for its competitors, questions about how DeepSeek handles user data have emerged and gained public attention, as users are becoming more and more conscious of the importance of data protection for preserving their privacy.

Recent reports suggested that DeepSeek shares data with ByteDance, the Chinese tech giant behind TikTok, raising concerns of possible data transfers to Chinese state-linked entities. These revelations, coupled with China’s data protection laws, have led regulators to question whether DeepSeek can adequately protect user privacy. These concerns have led the Personal Information Protection Commission (PIPC) of Korea to decide on the temporary removal of DeepSeek from app stores within the country until its data practices could be examined further. While existing users can still access the AI model, new downloads have been blocked. The decision mirrors actions taken by several other countries, including Italy, Australia and Taiwan, all of which have restricted or banned DeepSeek at some level.

Several countries have imposed restrictions on DeepSeek due to concerns over data security and transparency. Italy’s Data Protection Authority (Garante) limited DeepSeek’s processing of user data, citing insufficient transparency. Taiwan blocked government departments from using DeepSeek services, and Australia banned the AI tool on all government devices, citing national security risks. These governments argue that DeepSeek’s unclear data handling practices pose potential privacy threats, particularly given China’s National Intelligence Law, which allows government access to corporate data. Unlike Western AI companies that face regulatory scrutiny but, in most cases, remain permissible, Chinese AI firms like DeepSeek are often seen as direct security risks, leading to outright bans due to the stricter data protection regulations being adopted around the globe.

Korea’s data protection law, like that of the European Union’s General Data Protection Regulation (GDPR), emphasizes principles such as data minimization, purpose limitation, and transparency, such so that both laws are seen as having similar levels of protection. If DeepSeek is found to be transferring user data in ways that violate any of the principles provided by these Korean laws, it could face more severe regulatory action. For a country like Korea, which has established itself as a leader in tech innovation while maintaining strong data protection standards, DeepSeek’s case serves as a critical test of its regulatory framework.

According to a press release by the PIPC, published on Feb. 17, “DeepSeek has temporarily suspended its service in Korea as of February 15, 2025, to enhance compliance with the Personal Information Protection Act (PIPA). The company plans to resume service once necessary improvements are made.”

Deepseek’s rise is representative of China’s efforts to lead the AI race, independently from Western technology. However, its rapid expansion into global markets has reignited concerns over data sovereignty, as China’s National Intelligence Law provides the prerogative of the Chinese government to require companies and citizens to cooperate with national defense and intelligence services, as the law provides that national intelligence institutions may “request that relevant organs, organizations, and citizens provide necessary support, assistance, and cooperation” and that “All organizations and citizens shall support, assist, and cooperate with national intelligence efforts in accordance with law [...]”.

The suspension of DeepSeek has drawn attention from Korea’s leading AI firms, Naver and Kakao. Naver has responded by enhancing its AI model, HyperCLOVA X, improving efficiency and reducing operational costs to better compete with DeepSeek’s low-cost, high-performance approach. While committed to developing its own AI, Naver remains open to collaboration with external providers.

Kakao, on the other hand, has aligned itself with OpenAI, integrating multiple AI models rather than focusing solely on in-house development. However, Kakao’s leadership has expressed concerns over DeepSeek’s security and reliability, suggesting that additional investments in safety measures might negate its cost advantages.

Korea is increasingly recognized as an "AI middle power," alongside Japan, the EU and the Middle East. The Boston Consulting Group (BCG) notes that as U.S.-China competition intensifies, mid-tier AI nations like Korea have a unique opportunity to expand their influence.

One of Korea’s key advantages lies in its dominance of AI-related semiconductor manufacturing, with Samsung and SK hynix leading the global memory chip market. However, there are concerns that this advantage may not be sustainable, as countries like the U.S. and EU members invest heavily in domestic semiconductor production to reduce reliance on foreign supply chains.

However, as observed with the cautionary measures adopted in regard to DeepSeek, Korean companies also face the challenge of regulatory constraints on AI development. While Korean firms have strong AI model design capabilities, data accessibility remains a major hurdle due to strict privacy laws. Some industry leaders have proposed allowing select AI companies greater access to domestic datasets to support innovation while maintaining strong oversight, but for this to be successfully implemented, the regulations in force regarding data protection must be observed, or else the same risks and concerns raised in regard to DeepSeek will echo for any other company processing data within Korean jurisdiction.

As Korea’s AI industry adapts to these developments, the DeepSeek case underscores the ongoing debate over AI governance, data privacy and the balance between innovation and regulation. While DeepSeek’s return to the Korean market remains uncertain, its influence on AI policy and industry competition is undeniable.

Chyung Eun-ju ([email protected]) is a tech research associate at Donghyun ASP. She earned both her bachelor's in business and master's in marketing from Seoul National University. Joel Cho ([email protected]) is a practicing lawyer specializing in IP and digital law.