Observed Signal · Aug 20, 2026 · Technical Release · Source: techcrunch · Impact: 3/5 · Sentiment: Negative
Study: One-Third of Post-ChatGPT Pages Show AI Authorship
A Pew Research Center study released August 20, 2026, analyzed Common Crawl data covering nearly half a billion English-language pages from the past five years and scanned them with a detection tool (Open Pangram) to estimate AI-written content. Across the full archive roughly 10% of pages showed significant signs of AI authorship or heavy AI editing; when pages published before ChatGPT’s November 2022 launch are excluded, that share rises to about 35%. Pew also examined a random July 2026 sample of 10,000 pages, which likewise found about 10% with strong AI-authorship signals. The distribution is uneven across domains—about 10% of .com pages, 4.6% of .org pages, and roughly 1% for both .edu and .gov—highlighting the growing prevalence of chatbot-assisted content on the open web.
A credible, large-scale study (Pew) showing widespread AI authorship affects content quality, publisher inventory composition, brand safety, ad measurement and bot/fraud considerations for the ad ecosystem.
Track Pew Research Center Signals & Market Shifts in Real-Time
Polaris7 autonomous intelligence agents track regulatory filings, primary sources, executive changes, and deal flow 24/7. Create your free Explorer workspace to monitor these entities.
Key Takeaways & Evidence Grounding
- Pew Research Center published "How much of the internet is written with AI?" on 2026-08-20.
- The study used Common Crawl data covering nearly half a billion English-language webpages from the past five years.
- Pew scanned pages with Open Pangram's detection tool and a random July 2026 sample of 10,000 pages also found ~10% with strong AI signals.
- About 10% of analyzed pages showed significant AI authorship; excluding pages published before ChatGPT (Nov 2022) raises the share to ~35%.
- Domain breakdown: ~10% of .com pages, 4.6% of .org pages, and ~1% for both .edu and .gov pages.
Connected Companies & Entities
8 Entities mapped“Over one-third of webpages published after the release of ChatGPT show signs of being written by AI, according to a new study from Pew Resea...”
“Pew then used Open Pangram’s technology to detect how many were likely written or heavily edited by AI....”
“It also arrives shortly after internet infrastructure provider Cloudflare reported that bot web traffic had overtaken human web traffic....”
“Cloudflare reported that bot web traffic had overtaken human web traffic — a milestone that was reached sooner than the company had estimate...”
“Sarah has worked as a reporter for TechCrunch since August 2011....”
“She joined the company after having previously spent over three years at ReadWriteWeb....”
Ontology Mapping & Concepts
Related Market Signals & Shifts
Recent verified developments and strategic activity across this market segment.
One in Three New Websites Are AI-Generated
A study by researchers in the US and UK finds that roughly 35% of newly created websites are AI-generated, just three years after ChatGPT's launch. The team analyzed web snapshots in collaboration with the Internet Archive and used the Pangram AI-detection tool to classify content. The researchers report that AI-written pages tend to be steriler and more positive in tone but did not find an increase in factual inaccuracies compared with pre-AI content. The article cites Netcraft data showing about 1.5 billion websites worldwide (March 2026) and Bitkom estimates of roughly 60 million new sites per year. The research team plans to continue monitoring trends and to build a continual data pipeline with the Internet Archive.
2017 Human Text Classified as AI-Generated
A persoenlich.com editorial by Christian Beck describes how a 2017 blog post he wrote was flagged as AI-generated by detectors. He reports an initial TextGuard result of 76% AI-generated and later notes TextGuard rated the same text 61% AI while Pangram judged it 100% human. The article summarizes many LinkedIn comments showing widespread mistrust of AI text detectors, with examples such as Martin Luther King Jr.'s speech being labeled highly AI-generated and genuine AI text sometimes labeled human. The piece criticizes detector reliability, highlights how detectors favor highly structured journalistic style, and warns about commercial practices that profit from humanizing content. Beck concludes that, for now, AI text detectors are unreliable.
Pangram: Over 40% of LinkedIn Long Posts AI-Generated
Pangram Labs, maker of an AI-text detector, analysed social-media posts users actually encountered across platforms and found a high prevalence of AI-generated text. The study reports that more than 40% of LinkedIn posts longer than 250 words were classified as fully AI-generated. Pangram claims a 99.98% detection accuracy, supported by independent studies at the University of Maryland and the University of Chicago. Across the sampled platforms the average AI share was 13.8%, with longer posts more likely to be AI-generated; on X Pangram classified 25% of content as fully AI-generated and 23% as AI-assisted. LinkedIn has publicly commented on the rise of so-called "AI Slop" and said it will respond to the trend.
Track Real-Time Market Signals & Shifts
Set up custom watchlists to receive automated, evidence-grounded executive digests whenever material signals or shifts occur across your tracked landscape.