Study Explains Why AI Text Is Hard to Read: Overuse of High Working-Memory Structures

Linguists at NYU and UT Austin compared outputs from 14 major LLMs with human text, finding AI models overuse language patterns that raise working-memory load. This explains why AI-generated content often feels hard to read despite being factually correct.

2026-10-27 ~ 2026-10-27 · 2 related posts

• NYU and UT Austin study finds AI models overuse language patterns linked to reading difficulty — andrewaltair · 2026-10-27
• NYU/UT Austin Study: LLMs Overuse Six Working-Memory-Heavy Structures That Make Text Exhausting — xiaohu · 2026-10-27