A new Pew Research study says signs of AI authorship are now visible across a large share of the newer web. According to TechCrunch, Pew found that 35% of English-language webpages in its sample that were published after ChatGPT’s November 2022 release showed signs of being written or substantially edited by artificial intelligence. The finding is not a measurement of all web content, and it is not a direct audit of who wrote each page. TechCrunch reports that Pew used the Common Crawl web archive to collect nearly half a million English-language webpages spanning roughly the past five years, including material from before and after ChatGPT’s release. Pew then used Open Pangram’s technology to identify pages likely written or heavily edited by AI. The headline number came after Pew narrowed the lens to pages published after ChatGPT became widely available. In a random sample of 10,000 webpages collected in July 2026, around 10% showed “significant signs of AI authorship,” according to TechCrunch’s summary of the study. Pew noted that this broader sample would include older pages created before today’s AI writing tools existed, so it filtered out pre-ChatGPT pages to estimate the AI signal in newer material. After that filter, the rate rose to 35%. That is the central result: among the sampled English-language pages published since ChatGPT’s release, more than one-third carried signals that Pew’s detector associated with AI-written or heavily AI-edited text. Pew’s domain-level findings suggest the pattern is uneven. TechCrunch reports that .com URLs showed signs of AI authorship at about 10 times the rate of .edu and .gov domains, both of which were around 1%. The study also found a 4.6% rate for .org domains. The method has an important limitation. TechCrunch notes that Open Pangram, like other AI-detection systems, can incorrectly label human-written pages as AI-written. That means the study should not be read as a definitive page-by-page classification system. Pew’s claim is better understood as a large-scale signal about how much AI-generated or AI-edited writing may now be present in newer web pages. The report lands against a broader shift in how the web is produced and consumed. TechCrunch also points to Cloudflare’s recent report that bot web traffic had overtaken human web traffic, while distinguishing that issue from Pew’s focus. Cloudflare’s claim concerns what is browsing the web; Pew’s study concerns what those browsers may be reading. For publishers, search teams, data companies, and model builders, the practical takeaway is narrower but consequential: a substantial share of sampled newer English-language webpages showed signs of AI authorship. The precise rate remains dependent on detection methods, but Pew’s study adds another data point to the view that AI-generated and AI-edited text is becoming a measurable part of public web content. Who benefits: Publishers and domain owners using AI tools to produce or edit pages can scale output more easily. The study does not identify specific companies or products benefiting from that shift. Who's exposed: Readers, researchers, and systems that assume recently published webpages are human-authored face more uncertainty. The exposure is especially relevant for workflows that ingest public web text without tracking provenance or AI-authorship signals.