Six linguistic patterns that make AI text exhausting to read
In brief
- Advanced AI models reproduce complex linguistic patterns that are hard to read.
- Training on code and math datasets favors accuracy over clarity.
- Smaller models generate more natural text than larger ones.
A joint study by New York University and the University of Texas at Austin finds that advanced language models systematically reproduce linguistic patterns that are hard for humans to parse. The researchers attribute this issue to training on code and math datasets, where accuracy is prioritized over clarity. Smaller models, in contrast, produce more human-like text.
This summary gives you the essentials. The full article, its figures and quotes are on The Decoder.
Read the original articleAI-generated summary of the source, checked by rules. How we write · Report an error