More than a third of new web pages are created by AI: a large-scale study reveals a troubling trend
My latest analytical findings, based on data from a large-scale study, paint a rather telling picture: about 10% of all web pages on the internet bear clear traces of generative AI. However, if we filter out "antediluvian" content created before the ChatGPT era, the figure takes on truly alarming proportions—more than a third of all new materials online are already written or significantly edited by neural networks.
Methodology and Key Figures
During the study, a representative sample of 10,000 English-language pages collected in July 2026 was analyzed. An ML model was used for identification, detecting characteristic linguistic patterns of generative AI. The fact that archives of the old web ended up in the overall pool explains the relatively modest 10% in the overall statistics.
When I excluded materials published before November 2022, the picture changed radically. Among "fresh" pages, the share of synthetic content exceeded 33%. This is a direct consequence of the widespread adoption of ChatGPT, Claude, Gemini, and their competitors, which have rapidly turned text creation into a routine automated operation.
Uneven Distribution Across Domain Zones
Of particular interest is the segmentation by domains. In the .com zone, signs of AI generation are found on every tenth page, which is twice the rate of .org (4.6%). Meanwhile, academic .edu and governmental .gov show only about 1%—ten times lower. This disparity is explained by the nature of the content: commercial platforms actively automate product descriptions, SEO texts, and news feeds, while strict editorial standards in education and government sectors restrain the adoption of "raw" generative models.
Important Methodological Caveats
It is worth emphasizing that the study does not claim one hundred percent accuracy. This is a probabilistic assessment of authorship based on statistical linguistic markers, not verification of each specific case. Minor editing of text by a neural network may also not fall into the category of "significant editing." Nevertheless, the trend is obvious: synthetic content is rapidly displacing organic web content.
My expert commentary: The market has already faced content inflation, where SEO-optimized filler clogs search results. In the coming years, we will see either stricter ranking algorithms aimed at deprioritizing AI text, or a complete transformation of the web into a synthetic environment where trust becomes a scarce asset. Investors should take a closer look at projects developing content verification and digital signatures.