Over a Third of New Webpages Show AI Authorship Signs, Pew Study Finds

More than one-third of webpages published after ChatGPT’s launch in November 2022 show signs of AI authorship, according to a new Pew Research study released Thursday. The report, which analyzed nearly half a million English-language webpages, underscores the growing influence of generative AI on the web’s content landscape.

Study Methodology and Key Findings

Pew used the Common Crawl web archive to collect a random sample of webpages from the past five years, starting two years before ChatGPT’s release. In a July 2026 snapshot, around 10% of all sampled pages showed “significant signs of AI authorship,” but when Pew filtered out older pages published before AI tools existed, that figure jumped to 35%.

The detection was performed using Open Pangram’s technology, which identifies linguistic patterns typical of AI-generated text. While such tools can misclassify content, Pew notes the data is “directionally correct” at scale.

Domain Disparities and Linguistic Tells

The study also revealed significant differences by domain type. URLs with a .com domain showed AI authorship signs at nearly 10 times the rate of .edu or .gov domains, both of which had around 1% AI-authored content. .org domains had a 4.6% rate.

Pew also observed an increase in linguistic markers often associated with AI writing, such as em dashes, Oxford commas, and phrases like “it’s not X, it’s Y,” across the sampled pages.

Implications for Web Content Quality

The findings arrive shortly after Cloudflare reported that bot web traffic had overtaken human traffic, a milestone reached sooner than expected. This suggests a growing ecosystem where bots are both generating and consuming content, raising questions about the future of web authenticity and trust.

For readers, this means critical evaluation of online sources is more important than ever. For publishers and platforms, the study highlights the need for transparency and robust detection methods to maintain credibility.

Conclusion

The Pew study provides concrete evidence of AI’s pervasive role in shaping the modern web. As AI tools become more integrated into content creation, distinguishing between human and machine authorship will be a key challenge for platforms, regulators, and users alike.

FAQs

Q1: How did Pew detect AI authorship?
Pew used Open Pangram’s technology to analyze linguistic patterns in webpages, identifying features like em dashes, Oxford commas, and specific phrasing that are common in AI-generated text.

Q2: What does this mean for web content quality?
The prevalence of AI-authored content raises concerns about accuracy, originality, and trustworthiness. It underscores the need for readers to verify information from multiple sources and for platforms to implement robust content moderation.

Q3: Are all AI-written pages harmful?
Not necessarily. AI can assist in drafting content efficiently, but the lack of human oversight can lead to errors or low-quality information. The study highlights the scale of AI’s influence rather than judging individual pages.