Researchers find 35% of new websites are AI-generated by mid-2025
Updated
Updated · 404 Media · Apr 27
Researchers find 35% of new websites are AI-generated by mid-2025
12 articles · Updated · 404 Media · Apr 27
A study by Stanford, Imperial College London, and the Internet Archive analyzed 33 months of web data using Pangram v3 AI-detection software.
The research shows AI-generated sites have made the internet less semantically diverse and more positive, but not more dishonest or less sourced.
Researchers plan ongoing monitoring to track AI's evolving influence on website content, aiming to understand impacts by category and language, and to inform future internet policy and design.
Studies show the web is now more positive but less diverse. Is AI creating a sanitized echo chamber for humanity?
As publishers block web archives over AI fears, is the internet’s digital memory being permanently erased?
One study finds no AI-driven lies, another finds rampant misinformation. Which reality of AI's impact is true?
If users despise 'AI slop,' why are social media giants building an AI-centered internet?
With fragmented global AI laws, can we truly stop the cross-border flood of deepfakes and misinformation?
The 'AI Supercycle' is rebuilding the internet. Who will ultimately control this new digital infrastructure?