Urgent.News

What's breaking now, across thousands of outlets.

AI

DeepSeek banned from U.S. federal civilian agencies under new DHS order

The US Department of Homeland Security (DHS) has issued an order prohibiting federal civilian agencies from utilizing technology products from Chinese AI lab DeepSeek, as well as removing any existing implementations. The order encompasses all DeepSeek and DeepSeek-affiliate products, along with third-party products and services developed using DeepSeek's technology.

Compliance may prove challenging as it remains unclear which third-party products and services have been developed using DeepSeek's technology. Furthermore, the order is ambiguous regarding whether "developed using" implies ongoing access to DeepSeek models via API or refers to any app developed with the aid of DeepSeek. This ambiguity also raises questions about AI gateways, such as OpenRouter and Vercel AI Gateway, which provide DeepSeek's models through APIs.

Some federal agencies, including the Defense Information Systems Agency (DISA), the US Navy, NASA, and the US House of Representatives, have already restricted access to DeepSeek on their IT networks. Concerns surrounding DeepSeek's iOS mobile app arose shortly after its release in early 2025 due to potential plaintext information leaks, links to servers run by Chinese tech firms, and the lack of SSL pinning.

However, these issues are primarily associated with mobile security problems common to many apps, especially those integrating third-party advertising SDKs from outside US-based companies. Despite these concerns, Katie Moussouris, founder and CEO of Luta Security, argues that the order extends beyond necessary national security measures, particularly in banning third-party products and services developed using DeepSeek technologies.

The order's lack of clarity regarding "developed using" and the potential implications for software developed with DeepSeek tools pose practical verification challenges.

Written by urgent.news from The Register Science's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.

This story

This is one outlet's version. Read the fullest account.

Read the original at reuters.com →

More in AI

How to audit MCP tool schema token bloat and sync CLAUDE.md with AGENTS.md locally

If you connect 4–5 MCP servers to Claude Code or OpenAI Codex (GitHub, Playwright, Postgres, Notion), their tools/list JSON schemas silently consume 15,000–35,000 tokens on every turn before you even…

  • 4–5 MCP servers can consume up to 35,000 tokens per turn
  • CtxWise v1.0 tool inspects and fixes schema token bloat
  • CtxWise syncs CLAUDE.md with AGENTS.md locally

Your AI Agent Crashes on 429 Errors and You're Burning Money

Your AI Agent Crashes on 429 Errors and You're Burning Money Handling rate limits the right way saves you from unexpected bills and keeps your app reliable.

  • API rate limits exceeded, causing 429 errors
  • Errors lead to lost work and increased costs
  • Python script provides retry function with logging

More from Monday 19 October →