The US orders DeepSeek AI removed from federal civilian systems

On 19 October, the US Department of Homeland Security ordered federal civilian agencies to remove DeepSeek products and third-party services built using the Chinese AI model from their IT systems. The department also prohibited agencies from acquiring these products and services.

The decision marks another escalation in the technology competition between Washington and Beijing.

What is the US government worried about?

The department cited several potential security risks:

Sensitive information could be transferred to China through government employees’ or contractors’ use of the systems.
AI tools could potentially be used to obtain intelligence about US individuals or government agencies.
Software vulnerabilities or embedded malicious code could expose government systems to attack.

A July memorandum, made public on Monday, also instructed agencies to monitor traffic to and from government websites to address potential risks associated with DeepSeek. Some US agencies and individual states had already restricted its use.

It is important to distinguish the government’s security assessment from independently established proof that every DeepSeek deployment contains malware or transfers sensitive data. The announcement describes the risks cited by the department; it does not establish that every suspected outcome has occurred.

What does this mean for Europe?

European companies face a related question: should they allow employees to use AI services developed in countries whose governments may have different legal obligations and security interests?

The answer will vary by application. A public-facing tool used to brainstorm marketing ideas has a different risk profile from an AI system processing defence documents, medical records or confidential government information.

Organisations should assess data handling, contractual terms, access permissions, model hosting and security controls—not rely solely on a product’s country of origin.

The bigger picture: AI is becoming part of national security policy. As governments restrict particular tools, the global AI market may split into different technology ecosystems, increasing the costs of doing business across borders.