AI-generated content accounts for over one-third of new web pages

Research from the Pew Research Center indicates a significant rise in AI-generated content on the web. An analysis of 500,000 English pages suggests that while only 10% of the total sample shows signs of AI intervention, this figure rises to 35% for content created since the launch of ChatGPT in late 2022.

Researchers used machine learning to identify linguistic patterns, specific punctuation, and vocabulary common to models like ChatGPT, Claude, and Gemini. Certain terms, such as ‘delve’, ‘interplay’, and ‘testament’, have seen their usage more than double in these contexts.

The rise of AI writing has also sparked debates regarding authorship and ethics. While human ghostwriters have long been used by public figures to craft memoirs and opinion pieces, the use of AI by individuals—such as investor Stanley Druckenmiller—has raised questions about transparency. Some professional ghostwriters argue that they provide a level of diplomacy, interviewing, and emotional intelligence that current large language models cannot replicate.