More than a third of webpages created after ChatGPT's launch show signs of AI writing, according to a new Pew Research Center study that analyzed over half a million English-language webpages from the last five years.
The study, published last week, used an AI detection tool called Open Pangram to analyze a random sample of 10,000 webpages collected in July 2026. Researchers found that 10% of all pages in the sample showed signs of AI authorship - but that share jumped to 35% when they looked only at pages created after ChatGPT debuted in 2022.
The trend began with ChatGPT's launch and accelerated as other chatbots, including Claude and Gemini, entered the market. The findings add to a growing body of research suggesting that artificial intelligence is shaping a large part of the modern web.
Where AI writing appears most
AI-written text is not evenly distributed across the web. The study found that machine-generated content is most common on ".com" pages and remains far less prevalent on ".edu" and ".gov" domains.
Researchers cautioned that AI detection models are not foolproof, but said they can expose patterns associated with machine-generated writing, including specific punctuation, words, and phrases. ChatGPT and similar tools have made it easy for anyone to produce text at scale, and the data suggests many website operators are doing exactly that.
Chatbots are becoming a daily tool
The study arrives alongside other Pew research showing chatbots have become increasingly common in Americans' lives. A separate Pew study released in June found that over half of US adults now report using chatbots, and roughly 38% of employed adults said they use them for work.
Young adults under 30 are more wary of AI's potential negative impacts on society than older adults, according to the same data.
Why this matters for writers
For writers, the practical takeaway is direct: the web is filling up with machine-generated text, and readers are increasingly exposed to it. That doesn't make AI tools useless - but it raises the bar for work that aims to stand out. Original reporting, distinct voice, and verifiable facts are becoming more valuable, not less, as automated content grows.
Writers who use AI in their workflow should also understand how detection tools work and where their own output might carry machine-written patterns. Knowing the difference between drafting with AI and publishing unedited AI text is becoming a professional skill in itself. Resources for AI for Writers can help clarify those boundaries.
Your membership also unlocks:
