The United States Department of Homeland Security has ordered federal civilian agencies to remove Chinese artificial intelligence company DeepSeek’s products and related third-party services from their IT systems, citing concerns that sensitive government information could be transferred to China.
The directive, published on Monday, also prohibits federal civilian agencies from acquiring DeepSeek products or third-party services developed using its AI models. The restrictions reflect the department’s concerns about the potential security risks associated with the technology’s use in government operations.
What Has the US Government Ordered?
The Department of Homeland Security has directed US civilian government agencies to remove DeepSeek products from their existing information technology systems. The order also covers third-party services that use DeepSeek’s artificial intelligence models, extending the restriction beyond applications directly supplied by the Chinese company.
The directive addresses both the continued use of existing products and the acquisition of new ones. Federal civilian agencies are therefore being instructed to discontinue affected services while avoiding future purchases of technology developed using DeepSeek models.
The available notice does not specify a deadline for completing the removal or explain how individual agencies will implement the restrictions.
Why Is DeepSeek Being Restricted?
The Homeland Security Department has identified multiple potential threat vectors associated with DeepSeek models. Its principal concern is that sensitive information could be transferred to the People’s Republic of China if the technology is used by US government employees or vendors.
The notice warns that the use of DeepSeek products could create risks involving government data, particularly when employees or contractors interact with AI-powered applications during official work.
However, the available information does not establish that sensitive US government information has actually been transferred to China through DeepSeek. The directive is based on the security concerns identified by the department rather than a publicly detailed incident of confirmed data exposure.
Which AI Services Are Covered?
The restrictions apply to DeepSeek products as well as third-party services that use or have been developed using the company’s AI models. This means the directive is not limited to applications carrying the DeepSeek name.
A service incorporating DeepSeek technology may also fall within the scope of the order, even when it is supplied by another company. The department has specifically included such services in both the removal requirement and the prohibition on future acquisition.
The notice does not identify individual third-party vendors or provide a detailed list of affected products. It also does not specify how agencies should determine whether a particular service incorporates DeepSeek technology.
FCRF Launches CP-FRM to Build India’s Next Generation of Fraud Risk Professionals
Who Will Be Affected by the Decision?
The directive applies to US civilian government agencies and addresses the use of DeepSeek technology by government employees and vendors. It requires affected agencies to remove the covered products and services from their IT environments and restricts further acquisition.
The decision may also affect third-party technology providers whose services depend on DeepSeek models and are supplied to federal civilian agencies.
The available information does not indicate that the order extends to private individuals, businesses outside the federal civilian government or the general public. Its stated scope concerns the federal civilian agency IT systems covered by the Homeland Security Department’s notice.
What Are the Government’s Main Security Concerns?
The department’s warning centres on the possibility that information processed through DeepSeek-powered systems could be exposed to security risks, including potential transfer to China.
This concern extends to third-party applications because government employees may interact with AI models indirectly through other software products. The inclusion of such services indicates that the department is addressing potential exposure through both direct and indirect use of DeepSeek technology.
The notice describes the risks as multiple threat vectors but does not provide a detailed technical assessment in the available information. It also does not identify a particular vulnerability, cyberattack or confirmed data breach as the basis for the directive.
What Does the Order Mean for Government AI Use?
The directive establishes a restriction on the use and procurement of a specific Chinese AI technology within US federal civilian government systems. It places the responsibility on affected agencies to remove covered services and avoid acquiring additional products developed using DeepSeek models.
The decision also brings third-party AI integrations within the scope of government technology restrictions. Agencies must therefore consider not only the names of products they use but also whether those products rely on DeepSeek models.
The available notice does not outline a replacement technology, implementation schedule or broader policy governing other AI developers.
The420 Insight: AI Adoption Brings Data Security Under Greater Scrutiny
The US decision highlights how the underlying AI models used in government software can become a security concern, even when the technology is supplied through third-party services. By restricting both DeepSeek products and applications developed using its models, the directive places attention on the origins of AI technology and the potential exposure of sensitive information. The order reflects a government response to identified risks, although the available details do not establish that a data transfer or security breach has occurred.
About the author — Ayesha Aayat writes on cybercrime, digital safety, and emerging online threats. Her work focuses on public awareness, legal clarity, and technology-driven risks.
Follow for daily updates on cybercrime, corporate fraud, DFIR, hacking, investigations, and digital forensics
