{
  "id": 8703307,
  "title": "US Federal Register Caught Using Chinese AI Model for Document Search",
  "url": "https://urgent.news/2026/09/20/us-federal-register-caught-using-chinese-ai-model-for-document-search",
  "topic": "ai",
  "section": "AI",
  "published": "2026-09-20T14:03:00.000Z",
  "source": {
    "name": "Futurism",
    "slug": "futurism",
    "url": "https://futurism.com/artificial-intelligence/us-federal-register-chinese-ai-qwen-search-interface"
  },
  "original_language": "en",
  "account": "The US Federal Register, the official journal of the US federal government, has been caught using a Chinese AI model for document search. This development comes despite the country's top tech billionaires expressing fears about the capabilities of Chinese AI. The Federal Register offers a \"search mode\" option on its interface, which defaults to \"semantic\" but also allows users to choose between two options - \"Qwen3:0.6B\" and \"Qwen3:0.6B\". These options are from Alibaba Cloud's Qwen models, which are much cheaper to run compared to models developed by US-based labs. One user noted that \"even the United States government is unable to afford the cost per token charged by American AI companies\". The two Qwen options were removed from the interface a day after being discovered. The inclusion of these models does not suggest any malicious intent, but rather shows a preference for low-cost, capable AI.",
  "summary": "\"Even the United States government is unable to afford the cost per token charged by American AI companies.\" The post US Federal Register Caught Using Chinese AI Model for Document Search appeared first on Futurism .",
  "key_points": [
    "US Federal Register uses Chinese AI model for document search",
    "Qwen3:0.6B models from Alibaba Cloud offered as options",
    "Qwen options removed a day after discovery"
  ],
  "editors_take": null,
  "illustration": null,
  "coverage": {
    "outlets": 1,
    "also_reported_by": []
  },
  "ai_generated": true,
  "disclaimer": "Summaries, key points and the editor’s take are written by software from other outlets’ reporting and may contain errors — always check the linked original."
}