The New Reality of the Digital World: 35% of Online Content Shows Signs of AI!

The New Reality of the Digital World: 35% of Online Content Shows Signs of AI!

The launch of ChatGPT in November 2022 marked an irreversible turning point in the world of digital content creation and marketing. A recent data analysis published by the U.S.-based independent research organization Pew Research Center reveals the scale of this transformation in striking numbers: 35% of web pages published after the launch of ChatGPT show detectable signs of AI involvement.

Pew Research examined approximately 490,000 English-language web pages published between January 2021 and July 2026.

The Difference Between the Web as a Whole and the New Generation of Pages

In a random sample of 10,000 web pages from July 2026, researchers found that approximately 10% showed signs of AI-generated content. However, this sample also included older web pages created before generative AI became widely popular.

When the analysis was narrowed down to pages created after the launch of ChatGPT, the picture changed dramatically:

• AI usage in new-generation pages: 35%
• Commercial websites (.com): 10%
• Non-profit / civil society organizations (.org): 4.6%
• Academic (.edu) and government (.gov) websites: ~1%

The need for commercial websites to produce product-focused content, guides, SEO articles, and explanatory materials appears to have placed .com websites at the forefront of AI-assisted content creation.

AI's Digital Footprints: Linguistic Signals

In addition to the Open Pangram algorithm, researchers examined increases in linguistic patterns that are frequently associated with AI-generated writing. Some of the most striking changes observed between 2023 and 2026 include:

• Colons and em dashes: Their usage nearly doubled.
• Oxford comma: Usage increased by 63%.
• Characteristic words: Words such as “delve,” “interplay,” and “testament” appeared more than twice as frequently.
• Comparative sentence structures: Patterns such as “not only X, but also Y” increased threefold.

Important note: Pew Research emphasizes that the use of an em dash or certain words in a piece of writing does not prove that the text was written by AI. However, when observed at scale, these statistical shifts clearly indicate a broader transformation in the linguistic structure of the internet.

What Does This Mean for SEO and Content Marketers?

The rapid transformation of the web has two critical implications for search engines—particularly Google—and for users:

1. The Risk of “Model Collapse”

Future AI models may struggle to find sufficient amounts of human-generated data and could increasingly be trained on pages that were themselves written by AI. This could potentially amplify information pollution, reduce linguistic diversity, and encourage increasingly superficial content.

2. Quality and the Value of E-E-A-T

Although AI tools enable faster and cheaper content production, first-hand human experience (Experience), expertise (Expertise), authoritativeness (Authoritativeness), and trustworthiness (Trustworthiness) have become more valuable than ever for achieving sustainable success in search engines.