{
  "id": 2433077,
  "title": "The Dead Internet Theory May Be Coming True, Pew Research Findings Show",
  "url": "https://urgent.news/2026/08/21/the-dead-internet-theory-may-be-coming-true-pew-research-findings-show",
  "topic": "ai",
  "section": "AI",
  "published": "2026-08-21T20:36:40.000Z",
  "source": {
    "name": "CNET",
    "slug": "cnet",
    "url": "https://www.cnet.com/tech/services-and-software/dead-internet-theory-pew-research/"
  },
  "original_language": "en",
  "account": "The Dead Internet Theory, which posits that the internet essentially died in 2016 and is now dominated by AI-generated content, may be gaining credence, according to a recent Pew Research study. Conducted by the renowned research organization, the study analyzed nearly half a million English webpages from the past five years, including pages from before ChatGPT's public release in November 2022. The study then cross-referenced these pages with Open Pangram, an AI detection tool, to determine whether the content was authored by artificial intelligence.\n\nOut of the pages analyzed, 10% displayed \"significant signs of AI authorship,\" according to the study. The trend is believed to have begun in 2022 with the launch of ChatGPT, and has continued with the emergence of other AI bots such as Claude and Google's Gemini. The researchers observed that AI-generated content is most prevalent on .com domains, with around 1 in every 10 pages, while other domains like .org, .edu and .gov have a significantly lower rate of AI-written material - approximately 4.6%, 1%, and 1% respectively.\n\nAI-written text is characterized by certain linguistic patterns that deviate from typical human writing. For instance, AI-generated content is more likely to employ em dashes, Oxford commas, as well as phrases like \"delve\" and \"interplay.\" Additionally, AI bots often repeat key terms from prompts, produce repetitive explanations, and exhibit a robot-like tone instead of a natural one. Furthermore, AI bots are known to acknowledge the potential inaccuracy of their responses, which could be a helpful indicator when assessing the credibility of online sources.\n\nHowever, it should be noted that AI detection models are not infallible. They can sometimes misclassify human-written text as AI-generated or vice versa. Despite this, the Pew Research findings suggest that a significant portion of the internet content could be AI-generated, making it crucial for users to exercise caution and verify information from multiple, reliable sources.",
  "summary": "More of what you’re reading online could be written by an AI bot.",
  "key_points": [],
  "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."
}