arxiv.org web signal

Why is AI-generated language so hard to understand?

By Alexis Dufresne Published October 30, 2026 at 12:12 UTC

Summary

Ben Williamson @benpatrickwill.bsky.social: Maybe this explains why AI use disclosures on journal manuscripts claiming an LLM was used for "editing and language improvement only" usual… →
Jake Browning @jake-browning.bsky.social amplified

Ben Williamson @benpatrickwill.bsky.social

Maybe this explains why AI use disclosures on journal manuscripts claiming an LLM was used for "editing and language improvement only" usually function as warnings that 8000 words of unintelligible text are to follow arx…
View on Bluesky →

Originally reported by arxiv.org

Original headline: Why is AI-generated language so hard to understand?