{
  "id": 5052130,
  "title": "Creeping AI: One in three new web pages weren't generated by a human",
  "url": "https://urgent.news/2026/09/02/creeping-ai-one-in-three-new-web-pages-werent-generated-by-a-human-5052130",
  "topic": "ai",
  "section": "AI",
  "published": "2026-09-02T08:08:34.000Z",
  "source": {
    "name": "The National Business",
    "slug": "the-national-business",
    "url": "https://www.thenationalnews.com/news/2026/09/02/one-in-three-new-web-pages-shows-signs-of-ai-generation/"
  },
  "original_language": "en",
  "account": "A recent Pew Research Centre study found that more than one-third of new web pages published after the release of ChatGPT in 2022 contain elements generated by AI. This figure, at 35%, is significantly higher than the 9.6% observed for pages created before ChatGPT's launch. Before the AI tool's release, signs of AI involvement were virtually non-existent, with only around 1% of web pages exhibiting such characteristics.\n\nThe shift towards AI-generated content has been observed across various online domains, but the impact appears to be greater on commercial websites (9.3%) compared to educational (.org - 4.6%) and government (.edu/.gov - 1%) sites. This disparity suggests that institutions like schools and government agencies are still cautious about incorporating AI-generated content into their official web pages.\n\nThe Pew researchers also analyzed linguistic patterns in AI-generated text and found notable changes since 2023. For instance, the use of em dashes has roughly doubled, and the employment of Oxford commas has increased by 63%. Additionally, phrases like \"align with\", \"delve\", \"interplay\", \"showcase\", and \"testament\" have more than doubled in usage, suggesting a distinct AI writing style. Negative parallelism, a writing technique that describes something by contrasting it with what it is not, has nearly tripled in prevalence.\n\nWhile these linguistic patterns can indicate the likelihood of AI involvement in content creation, AI detection tools are not infallible. They sometimes misclassify human-written content as AI-generated, and vice versa. Furthermore, as AI-generated text becomes more prevalent, there's a concern that it might influence the way humans write, potentially altering the linguistic landscape of the internet.",
  "summary": "You've probably read something online recently and wondered: was this written by AI? There's a good chance it was. New research from the Pew Research Centre, in the US, examined hundreds of thousands of web pages to measure how often they showed signs of AI generation or editing. Our chart of the week shows that web pages published after ChatGPT was released in November 2022 exhibit significant…",
  "key_points": [],
  "editors_take": null,
  "illustration": null,
  "coverage": {
    "outlets": 2,
    "also_reported_by": [
      {
        "outlet": "The National UAE",
        "title": "Creeping AI: One in three new web pages weren't generated by a human",
        "url": "https://urgent.news/2026/09/02/creeping-ai-one-in-three-new-web-pages-werent-generated-by-a-human",
        "published": "2026-09-02T08:08:34.000Z"
      }
    ]
  },
  "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."
}