On the occasion of the first anniversary of China’s artificial intelligence (AI) model “DeepSeek,” its security policies and privacy protection issues are facing increasing scrutiny from countries around the world.
According to its own privacy policy, DeepSeek stores a large amount of personal data (such as requests to the AI program or uploaded documents) on computers within China.
As reported by Reuters, as of January 2026, more than ten countries have taken substantial restrictive actions against DeepSeek, with the global review list continuing to grow.
The Trump administration is weighing measures to block the company’s access to U.S. technology. Secretary of Defense Pete Hegseth has received proposals from Congress to designate DeepSeek as a “facilitator for Chinese military entities.”
Republican senators have also called on the Department of Commerce to assess the potential vulnerabilities that its open-source model poses to U.S. data security.
Germany:
Regulatory authorities have requested Apple and Google to remove the program from their app stores citing data security vulnerabilities.
Netherlands and Czech Republic:
Both have banned government officials from using it on work devices. Netherlands pointed out that the company originates from a country with “offensive cyber programs.”
Belgium:
The National Cyber Security Center (CCB) is conducting a risk assessment to prevent its potential use in social engineering attacks against EU officials.
France, Italy, and Spain:
Privacy regulators of these countries have initiated investigations into DeepSeek’s data collection practices and compliance with the General Data Protection Regulation (GDPR).
Switzerland:
As a non-EU member country closely linked to the European economy, the Swiss Federal Data Protection and Information Commissioner (FDPIC) remains vigilant, evaluating whether DeepSeek’s algorithms might induce users to disclose sensitive banking confidentiality data during conversations.
Australia and Taiwan:
Both have implemented government bans, considering it a serious national security risk.
India:
The Ministry of Finance has instructed employees to avoid using it to prevent leakage of government confidential documents.
South Korea:
Previously suspended its download for failure to comply with personal information protection regulations.
Japan:
While not issuing a formal ban, it has administratively directed core companies to strictly filter its access to prevent leakage of semiconductor and other technological secrets.
Compared to AI tools like ChatGPT from the United States, governments of various countries stress three core reasons when reviewing DeepSeek:
Legal accountability inequality:
In case of violations by U.S. companies, the EU can enforce effective prosecution through cross-border regulations such as the DSA; however, DeepSeek rooted in the Communist Chinese legal environment makes it nearly impossible for Western regulatory agencies to conduct on-site audits or cross-border accountability.
Data sovereignty inequality:
Under the PRC’s National Intelligence Law, companies like DeepSeek must cooperate with government intelligence work. This means that all conversations, codes, or files uploaded by users could be accessed by PRC authorities at any time on a legal basis.
Algorithm transparency inequality:
DeepSeek has been found to have preset positions on specific politically sensitive terms. Western governments believe it is not just a technical product but also a tool for the PRC to export “thought censorship” globally and conduct “cognitive warfare.”