A study reported by TechCrunch says roughly one third of web pages published since ChatGPT launched show signs of AI authorship. The finding suggests generative AI is no longer a side tool for the web; it is shaping a large share of newly published material.
That shift has practical consequences for search, media, education, and model training. If more online text is generated or heavily edited by AI systems, readers may face more repetitive content, and future models may train on material influenced by earlier models.
The result depends on how the study detects AI authorship, so it should be read as evidence of a broad pattern rather than a perfect census. Even with that caveat, the scale is hard to ignore.