Context: On 20 January, Chinese AI startup DeepSeek launched its flagship DeepSeek-R1 model, which it claimed rivals technology developed by ChatGPT-maker OpenAI while requiring far less money to make. One of its previous models, DeepSeek-V3, only cost $6 million worth of computing power from Nvidia H800 chips, it said. The platform has received a lot of media attention since its launch, especially after its app topped download charts and caused technology stocks in the U.S. to crash.
What’s new: Several South Korean governmental bodies are seeking – or have already successfully done so – to block DeepSeek over data collection practice concerns. The country’s Ministry of Environment today blocked access to the AI service from its internet-connected computers, while its finance ministry has unveiled plans to also block access to DeepSeek. Both are citing concerns over its data collection practices.
Direct impact: This is a first in South Korea and it comes right after the government launched an initiative to restrict access to DeepSeek to prevent any potential leaks of critical information through GenAI services. The government has not blocked any other AI services.
Wider ramifications: The blocks by South Korea’s government agencies join several others issued in Italy, Taiwan, and Australia. DeepSeek is also under a lot of pressure in the U.S., where its rival OpenAI has accused the company of infringing its IP by distilling the knowledge out of its models and breaching its terms of use (although there is currently no formal litigation that we know of).
In a statement today, South Korea’s environment ministry said:
“The intelligence authorities have asked for caution in using DeepSeek, as its personal information collection system is not yet clearly known.”
Meanwhile, an official at the finance ministry stated:
“Due to multiple technical concerns raised about DeepSeek from home and abroad, we plan to block access for the service on PCs connected to external networks.”
The country’s Ministry of Unification is also considering similar restrictions on AI services, including DeepSeek, while the Ministry of the Interior and Safety sent letters to government agencies, and metropolitan and provincial governments, urging caution when using AI services like DeepSeek and ChatGPT. The letter reportedly emphasized the importance of not sharing personal details and refraining from blindly trusting the outcomes provided by these services.
Outside of government action, several major tech companies in South Korea are also bringing in internal bans against DeepSeek. Kakao on Tuesday prohibited the use of DeepSeek for business purposes, becoming the first major tech company to do so. LG Uplus, Samsung Electronics, SK Group and LG Electronics, which are also all developing their own generative AI services, have banned such programs on company computers without prior authorization.
DeepSeek under international pressure
South Korea was not the first to bring in bans against DeepSeek. Italy’s privacy regulator Garante per la Protezione dei Dati Personali (GPDP, frequently referred to as “il [= the] Garante”) issued an immediate ban on DeepSeek, citing data privacy and lack of transparency (January 30, 2025 Garante press release). The move mirrored the one that it made in March 2023 against OpenAI’s ChatGPT, when the platform was temporarily banned over GDPR violations. However, that one was allowed back online a month later following compliance improvements.
A day after Italy issued its ban, Taiwan’s Ministry of Digital Affairs (MODA) also prohibited government employees from using the service, citing concerns that it could expose sensitive data to China. MODA described DeepSeek as a Chinese information and communications technology product, warning that data leaks could threaten national security.
In Australia, the Department of Home Affairs ordered government agencies to remove all DeepSeek products, applications, and web services from official devices and prevent future installations on Tuesday (February 4, 2025 Australian DHA Directive). The use of DeepSeek products, applications and web services poses “an unacceptable level of security risk” to the Australian Government, it said, adding that any “legitimate business reasons” to download DeepSeek must be time-limited and must be “necessary for the carrying out of national security or regulatory functions”.
In the U.S., OpenAI has alleged there is evidence that DeepSeek distilled the knowledge out of OpenAI’s models, which is in breach of its terms of use and infringes on its IP. It said in a statement:
“We are aware of and reviewing indications that DeepSeek may have inappropriately distilled our models, and will share information as we know more. We take aggressive, proactive countermeasures to protect our technology and will continue working closely with the US government to protect the most capable models being built here.”
The company has not (publicly) filed a lawsuit against DeepSeek.
The nominee for U.S. Commerce Secretary Howard Lutnick has also accused DeepSeek of stealing U.S. technology and circumventing U.S. export controls to acquire Nvidia chips. During his confirmation hearing before the U.S. Senate last week, he alleged that the Chinese AI platform had been able to develop its AI models at a significantly lower cost by purchasing large quantities of Nvidia chips and exploiting data from Meta’s open platform.
Mr. Lutnick stated:
“I take a very jaundiced view of China. They only think about themselves and seek to harm us, and so we need to protect ourselves. We need to drive our innovation, and we need to stop helping them. Meta’s open platform lets DeepSeek rely on it. Nvidia’s chips – which they bought tons of, and they found their ways around [export controls] – drive their DeepSeek model. It’s got to end.”