The US Department of Homeland Security (DHS) made public on Monday an order requiring the removal of products and services from the Chinese AI company DeepSeek from the information technology systems of civilian federal agencies. The measure also prohibits new acquisitions of these products and of third-party services developed based on the model's implementations.
The order, signed by Homeland Security Secretary Markwayne Mullin and dated July 24, was disclosed through a notice in the Federal Register. It applies to civilian executive agencies such as Treasury, State and Justice, but excludes national security and classified systems, which follow separate rules.
What the US government alleges
According to the DHS, DeepSeek models present “multiple threat vectors.” The agency states that use may result in the transfer of sensitive data to China if used by government employees or vendors, generate actionable intelligence on US persons and agencies, and that the models themselves may contain malware or security vulnerabilities. These are the reasons presented by the government; there is, so far, no independent public proof that all DeepSeek products contain malicious code.
The measure arrives at a moment when DeepSeek has gained international attention for efficient open-weight models, such as the recent V4.1-Flash version. Previous restrictions already existed in some agencies and US states, but the current order is the broadest civilian federal action against the Chinese lab.
Limited scope
The ban does not extend to private or business use in the United States. Ordinary users and private companies are not subject to it. Open-weight models remain available in public repositories, and non-governmental third-party services are not automatically affected.
The Chinese Embassy in Washington and DeepSeek itself were contacted by outlets such as the BBC and Reuters, but there was no immediate response at the time of the reports.
Context of technological tension
The decision occurs in the context of criticism from the American administration of Chinese AI companies and growing concerns about the software supply chain of artificial intelligence. Other countries, such as South Korea, had already restricted government access to DeepSeek in 2025.
For the geek and technology public, the measure reinforces the debate on the use of open models of Chinese origin in sensitive environments, while not preventing development and experimentation outside the civilian federal government scope.
Sources
- BBC: US bans DeepSeek AI from government use
- Reuters: US banning Chinese AI DeepSeek from government agency IT systems
- Bloomberg: US Moves to Bar China’s DeepSeek From Federal Agency Systems
Transparency: This content was created, edited or reviewed with the assistance of artificial intelligence. The information was cross-checked with public posts on X and sources available on the internet. Consult the original sources to verify the full context.
By GeekikiBot
