AI is taking over the web: a third of new pages are created by neural networks
The scale of generative models' penetration into the web space has turned out to be far more serious than commonly assumed. My analysis of recent data shows: roughly one in ten web pages on the internet bears the imprint of synthetic text. However, if we look only at content published after ChatGPT's debut in November 2022, the picture becomes radically different — more than a third of such pages show clear traces of AI generation or deep editing.
Methodology and Key Figures
During the study, a random sample of 10,000 English-language pages collected in July 2026 via Common Crawl was analyzed. For identification, a machine learning model tuned to detect characteristic linguistic patterns of generative AI was used. The fact that the overall pool included archival materials created long before the era of modern neural networks explains why the overall share of 10% looks modest. After clearing the sample of "prehistoric" content, we see explosive growth: synthetic content occupies dominant positions in the new web.
The Commercial Sector — Epicenter of Generation
The distribution of AI content across domain zones confirms the economic logic. In the .com zone, signs of AI are detected in about 10% of pages — twice as much as in .org (4.6%). Academic (.edu) and government (.gov) resources turned out to be the "cleanest" — there, the share of synthetic content barely reaches 1%. This is explainable: commercial platforms actively automate the production of SEO texts, product descriptions, and news feeds, while institutional sites maintain stricter control over authorship.
Important Caveats
It is worth emphasizing that this is a probabilistic estimate, not precise detection. The methodology relies on statistical linguistic markers rather than checking edit history. This means that some "synthetic" pages may have been created by a human with a characteristic style, while others may have only been edited by AI. Nevertheless, the trend is obvious: neural networks have become the main "content generator" of the modern internet.
My conclusion: We are witnessing not just a trend, but a fundamental shift in the web ecosystem. The market is already oversaturated with uniform machine-generated text, which will inevitably lead to content devaluation and intensified competition for quality. Under these conditions, search and ranking algorithms will be forced to evolve to distinguish live analytics from streamed generation. The question is not whether this will stop AI, but who will adapt faster to the new realities.