AI model reveals insights into human reading habits

Researchers from Aalto University, the Hong Kong University of Science and Technology, City University of Hong Kong, and the National University of Singapore have used artificial intelligence to study human reading patterns. By training an AI model to maximize comprehension under constraints of limited time, memory, and visual processing capacity, the team discovered that the AI developed reading behaviors strikingly similar to those of humans.

Without being programmed to mimic human eye movements, the AI exhibited ‘resource rationality,’ a concept where attention is managed like a budget. The model spent less time on predictable words and was more likely to skip them, while it frequently backtracked when encountering difficult or ambiguous text.

In experiments involving 39 adults, the AI mirrored human responses to time pressure: when time was limited, it prioritized covering more text, whereas with more time, it focused on difficult passages and rereading. While researchers caution that this does not prove the human brain performs identical calculations, the model provides a framework for how reading behaviors emerge from the goal of maximizing understanding with limited cognitive resources. This research could eventually lead to personalized reading technologies and customizable applications.