A new AI authorship study from Pew Research reveals that over one-third of web pages published after ChatGPT’s release show significant signs of being written or substantially edited by artificial intelligence. The research, released on Thursday, provides compelling evidence of how AI has rapidly transformed content creation across the internet.
The Pew Research AI authorship study analyzed nearly half a million English-language web pages from the past five years, using Common Crawl’s web archive. Researchers employed Open Pangram’s detection technology to identify pages likely written or heavily edited by AI.
What the AI Authorship Study Found
In a random sample of 10,000 web pages collected in July 2026, approximately 10% showed significant signs of AI authorship. However, Pew researchers noted that this snapshot included older pages that predate AI writing tools, making the statistic less revealing.
By filtering out pages published before ChatGPT’s November 2022 launch, the picture becomes far more dramatic. According to the AI authorship study, over one-third (35%) of newer web pages display clear indicators of AI involvement in their creation or editing process.
Domain Variations in AI Authorship
The AI authorship study found significant differences across domain types:
.com domains: Approximately 10% showed AI authorship signs
.org domains: Only 4.6% rate of AI authorship
.edu domains: Roughly 1% AI-authored
.gov domains: Approximately 1% AI-authored
This stark contrast suggests that commercial websites have adopted AI writing tools far more aggressively than educational institutions or government agencies.
AI Detection Methodology and Limitations
Pew’s research relied on AI detection technology from Open Pangram, which analyzes writing patterns to identify machine-generated content. The researchers acknowledge that detection tools can sometimes misclassify human-written content as AI-generated.
Despite these limitations, the scale of the data provides directional accuracy about the growing prevalence of AI on the web. The findings corroborate other studies examining how much of the internet’s newer content is now produced by artificial intelligence.
Broader Context: The Rise of Bot Traffic
The Pew AI authorship study arrives shortly after Cloudflare reported that bot web traffic had surpassed human web traffic—a milestone reached earlier than the company had projected. This creates an interesting dynamic where bots may increasingly be reading web pages written by other bots.
While Cloudflare’s data focused on who or what is browsing the web, Pew’s research examines what is being browsed. Both studies paint a picture of an internet increasingly dominated by automated systems on both the consumption and production sides.
Linguistic Markers of AI Authorship
Beyond detection scores, Pew found that certain linguistic patterns associated with AI writing have increased over the years. These include:
Em dashes (—) used more frequently
Oxford commas appearing with greater regularity
Specific phrasing patterns like “it’s not X, it’s Y”
These stylistic signatures have become more common across the web as AI writing tools proliferate. While none of these markers alone proves AI authorship, their increased frequency aligns with broader trends in AI content generation.
What This Means for Internet Users
The prevalence of AI-generated content has significant implications:
For readers: Understanding that much of what you read online may be AI-authored helps you evaluate information more critically. AI content often requires human verification for accuracy, especially on complex topics.
For content creators: The AI authorship study highlights the competitive pressure to adopt AI tools while maintaining quality and originality. Many websites may turn to AI to scale content production, potentially affecting the diversity of voices and perspectives online.
For search engines and platforms: Google and other platforms face challenges in distinguishing human from AI content. The ongoing development of detection tools and ranking algorithms that reward human expertise and original research remains crucial.
The Changing Landscape of Web Publishing
Since ChatGPT’s launch, the web has undergone rapid transformation. The Pew Research AI authorship study provides concrete evidence that AI is now a primary driver of web content creation, particularly in commercial domains.
As AI writing tools become more sophisticated and widespread, distinguishing between human and machine authorship may become increasingly difficult. The study’s findings suggest that we have already crossed a threshold where a significant portion of the web’s new content is machine-generated.
For publishers, the challenge lies in balancing efficiency gains with maintaining trust and quality. For readers, developing media literacy skills to evaluate content regardless of its origin becomes ever more important.
The full impact of this shift on information quality, search engine results, and online discourse remains to be seen. However, Pew’s research makes clear that AI has fundamentally changed how content is created on the internet—and that change is accelerating.
