One-Third of Webpages Post-ChatGPT Show AI Authorship, Pew Study Finds
· Technology · TechCrunch
A third of webpages published after ChatGPT's November 2022 launch exhibit signs of AI authorship, per a Pew Research study analyzing over 490,000 English-language pages. The report found 35% of post-ChatGPT pages in a filtered dataset used AI detection tools like Open Pangram, with .com URLs 10x more likely to show AI use compared to .edu or .gov domains. This aligns with Cloudflare's report that bot traffic surpassed human web activity in 2023, highlighting bot-generated content's growing dominance. The study also noted increased use of em dashes, Oxford commas, and phrases like 'It's not X, it's Y' as AI markers. Researchers acknowledge detection tools' limitations but affirm results are directionally accurate.
Why it matters
The findings reveal AI-generated content's rapid spread, raising concerns for digital trust, media integrity, and workforce shifts as industries adapt to automated writing. Journalists, educators, and content platforms may need new verification systems to differentiate human and AI work.
Read the original report — TechCrunch
Join us on Telegram
Breaking news the moment it lands. At 10,000 members we ship the Android app.