Study Identifies Six Linguistic Patterns That Make AI Text Hard to Read
Understand the story
Researchers from New York University and the University of Texas at Austin conducted a study on the readability of AI-generated text. The research team had 14 models generate three types of text: everyday conversation, programming, and argumentative essays, analyzing 120 conversations, 45 programming sessions, and 980 essays in total, and comparing them with human-written text. Through this comparative analysis, the study identified six linguistic structures that AI overuses, which make AI-generated text hard to read, including patterns such as rare word combinations, noun stacking, and using nouns as verbs. The study was reported by The Decoder on October 27, 2026.
AI overview of reporting · Updated 3 h前
Reporting timeline
Explore different aspects of the story through its reporting.
- The DecoderResearchers find six patterns that make AI text so exhausting to read
Research from New York University and the University of Texas at Austin had 14 models generate everyday conversation, programming, and argumentative essay text, analyzing 120 conversations, 45 coding sessions, and 980 articles in total, and comparing them with human text, identifying six language structures that AI overuses, such as rare word combinations, noun stacking, and verbing nouns.
Story heat trend
Current heat9·Comparable peak10（10/27 03:00）·Comparable change over 24 hours –
The trend compares the same continuously observed sources, which may cover fewer sources than the current heat score. Hover or select the chart to inspect each hour; use the left and right arrow keys to navigate.