A recent Pew Research study reveals that over one-third of web pages published since the launch of ChatGPT in November 2022 exhibit signs of AI authorship or significant editing. This finding aligns with other research showing AI-generated content has become a major part of the web’s newest material. The study analyzed nearly half a million English-language web pages spanning five years using Common Crawl data, focusing on those likely created or heavily modified by AI technologies like ChatGPT.

To detect AI involvement in content creation, Pew employed the technology from Open Pangram, which scans for linguistic and stylistic markers indicative of AI writing. Their results indicated that about 10% of a random sample of 10,000 pages from July 2026 displayed notable AI authorship. However, this percentage rose sharply to 35% when considering only pages published after ChatGPT’s debut, reflecting the rapid adoption of AI in generating online text.

Domain-specific trends showed that websites with a .com extension were about ten times more likely to contain AI-authored content compared to .edu or .gov domains, which had around 1% AI authorship. Additionally, .org domains had a lower rate, at 4.6%. Despite the limitations of AI-detection tools, which can sometimes misclassify content, the overall data indicates a strong directional trend towards AI shaping a significant portion of new web content.

This surge in AI-generated web pages coincides with reports from Cloudflare that bot web traffic has overtaken human web traffic on the internet. Pew’s research highlights how the web is increasingly becoming a space where AI not only produces content but also consumes it, as bots interact with pages created by other AI systems. This evolution presents new challenges and opportunities for understanding the future dynamics of online information.