Llm writing changes reading effort by word and length differences
The Cognitive Cost of LLM Writing: A Reading-Time Comparison with Human Text
Computation and Language
Summary
Reading text created by large language models (LLMs) can feel different than reading text written by humans. The authors created a model to predict how long people spend reading each word, based on eye movement data. They found LLM-generated text is usually shorter overall but each word may take a bit more effort to read. This is because LLM texts use longer, less common words and more variety in vocabulary. Their study shows that while LLM writing reduces how much you read, it may increase the mental effort for each word.
What this means in practice
- •For content designers: Adjust text length and word choice in AI-generated materials to optimize reader cognitive load based on predicted reading times.
- •For user experience teams: Evaluate AI-generated content for potential increases in word-level cognitive effort when designing digital reading experiences.
Authors
Kamer Cekini, Elisabetta Biondi, Chiara Boldrini, Andrea Passarella, Marco Conti
Abstract
Although linguistic differences between human-written and Large Language Model (LLM)-generated texts have been extensively studied, their implications for the cognitive effort involved in reading remain less explored. In psycholinguistics, reading time is commonly used as an indirect measure of cognitive processing effort. Building on this perspective, we investigate whether human-written texts and their LLM-generated rewrites exhibit differences in predicted reading time that may reflect differences in processing demands for readers. To this end, we introduce RoBERTa-RT, a model trained on human eye-tracking data to predict word-level reading times. We evaluate the model on both internal and external test sets, construct a paired dataset of 496 human-written articles and their LLM-generated rewrites, and apply RoBERTa-RT to compare their predicted reading times. Our findings show that LLM-generated rewrites tend to be shorter and have slightly lower predicted total reading times, while exhibiting higher predicted reading times per word. These differences are associated with systematic linguistic changes, including longer and less frequent words, greater lexical diversity, and higher contextual predictability. From a cognitive perspective, our findings suggest that LLM-generated writing may redistribute reading demands: reducing the amount of text to process while increasing the predicted processing effort associated with individual words. This highlights the importance of considering not only the linguistic characteristics of machine-generated content, but also its potential implications for human cognitive processing.