Guozhen AIGlobal AI field notes and model intelligence

Realtime AI News

DHS orders DeepSeek removed from federal civilian IT systems

The Department of Homeland Security ordered US civilian agencies to remove DeepSeek products and DeepSeek-based third-party services from their IT systems, and barred buying them. DHS cited data flows to China, possible malware and intelligence risks, extending a wider Trump administration crackdown on Chinese technology.

Published
美国土安全部下令联邦民事机构从IT系统中移除DeepSeek
Image source: deepseek.com

The US Department of Homeland Security ordered civilian federal agencies on October 19 to remove DeepSeek products, along with third-party services built on the Chinese AI model, from government IT systems.

The notice also bars civilian agencies from buying DeepSeek products or third-party services developed using DeepSeek, according to Reuters' account of the document published that day.

DHS cited “multiple threat vectors.” It said use of the product “may result in the transfer of sensitive data to the PRC if used by US government employees or vendors,” and warned that agency use “could produce actionable intelligence on US persons and/or government agencies.” The models themselves, it added, “may contain malware or other security vulnerabilities.”

The order was not improvised. In a July memo, DHS had already told federal agencies to monitor traffic flowing in and out of government websites to address potential DeepSeek-related risk; that memo was made public on October 19 as well.

Several agencies had moved first: the Commerce Department and the Navy banned DeepSeek on government devices, and states including New York, Tennessee and Virginia took similar steps. The Chinese Embassy in Washington did not immediately comment.

The move is part of a broader push by the Trump administration against Chinese technology companies. Reuters reports that the Federal Communications Commission will vote on October 29 to bar Chinese labs from testing electronics — smartphones, cameras, computers — destined for the US market, a category in which 82% of US electronics were tested in China in 2025.

The signal for the AI industry is that model-level scrutiny is moving from device bans toward procurement and supply-chain exclusion. Vendors selling AI services to US civilian agencies must now trace whether DeepSeek sits anywhere in their stack, not just whether an employee installed the app.

What to watch next: compliance deadlines and enforcement guidance for agencies, whether the Defense Department and its contractors extend similar rules, and the outcome of the October 29 FCC vote, which could widen restrictions across Chinese AI and hardware supply chains.

Why it matters

The order pushes DeepSeek restrictions from devices into procurement and supply chains, forcing AI vendors serving US civilian agencies to audit where the model appears in their stack. Because DHS names third-party services too, commercial products built on open models may be swept in, shifting compliance costs onto integrators and resellers.

DeepSeekUS PolicyData SecurityDHS
Back to AI Daily

Nearby Updates

All

10/20, 04:09

AWS launches TorchNeuron to move PyTorch workloads onto Trainium

AWS released TorchNeuron on Monday, a tool that lets PyTorch applications run on its Trainium chip after a one-line configuration change instead of a code rewrite. Amazon is also pushing changes upstream to make PyTorch less tied to Nvidia hardware, a direct challenge to CUDA's long-standing software moat.

10/20, 04:19

Circle and Proof demo AI agents proving identity and paying in USDC

Circle and Proof demonstrated a flow in which an AI agent presents a verified credential over Proof's x401 identity protocol and pays in USDC over x402 from a Circle Agent Wallet. The pitch is to give merchants a way to verify who an agent acts for and get paid, rather than blocking agents outright.

10/20, 03:28

Hackers breach OpenCode Console, leak API keys and run up user bills

OpenCode, an open-source AI coding agent, said attackers broke into an older system and exposed API keys for 1.4% of users on its OpenCode Console platform, with some developers reporting hundreds or thousands of dollars in unauthorized charges. Anthropic and Google blocked the keys and OpenAI is issuing refunds, while the incident fueled debate over safety guardrails that stop defenders from analyzing malicious code.

10/20, 03:27

Germany signs industrial AI declaration with NVIDIA as Huang backs four German startups

German Chancellor Friedrich Merz and NVIDIA CEO Jensen Huang discussed Germany's industrial AI roadmap in Berlin, the same day Berlin signed a joint declaration of intent with NVIDIA on physical AI. Huang also announced planned investments in four German startups and a Siemens partnership on autonomous factories.