Understanding When People use AI on Different Cognitive Tasks

Read full text Summarise with VIRAT
Authors: Sunny Yu, Ilia Sucholutsky, Katherine M. Collins, Anonymous , Dan Jurafsky, Robert Hawkins, Ahmad Jabbar
Cited by: 0

Summary

No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.

Summarise with VIRAT Free to try with your Science Catalogs account.

Abstract

Large language models (LLMs) have the potential to boost human productivity by speeding up task completion---provided users know when to offload cognitive work to them. For each task, users have the choice to do it themselves or ask an AI for help. We can examine the choice of how to complete the task under the lens of resource rationality, assessing whether people choose the option that best conserves their cognitive resources (e.g., where those resources may be time or mental effort). Are people well-calibrated about when they would use AI for a task? How does the rate of AI use in naturalistic settings compare with people’s stated preferences of how they would like to complete a task? We hypothesize that across cognitive tasks, people use AI significantly more than they state they would, and that this gap is more salient for easy tasks. We also hypothesize that people who are averse to thinking are more likely to use AI for help when completing different tasks.

Fields of Study & Subjects

Human-Automation Interaction and Safety Decision-Making and Behavioral Economics Information Retrieval and Search Behavior

Publication details

Publication date
1 January 2026
Language
EN
Document type
Preprint