Typing Behavior in Human-LLM Interaction: Keystroke Dynamics Reveal Cognitive Effort During Prompting

2026-06-26

AI Summary

This study addresses the real-time assessment of cognitive load and collaborative experience during human interaction with large language models (LLMs). Through a user study, subjective cognitive load was evaluated using the NASA-TLX scale, while keystroke dynamics—including keystroke count, typing speed, and pause duration—were analyzed for the first time as behavioral proxies of real-time cognitive effort. The findings demonstrate that task difficulty significantly influences keystroke behavior: higher task difficulty leads to increased keystrokes, slower typing, and longer pauses, all of which effectively reflect elevated cognitive load. However, these behavioral indicators do not predict users’ perceived usefulness of the model’s outputs. This work offers a novel perspective and a set of behavioral metrics for understanding cognitive processes in human–LLM interaction.

Abstract

As Large Language Models (LLMs) become increasingly integrated into daily routines, understanding how users interact with these systems is crucial for effective human-AI collaboration. This work investigates keystroke dynamics as a behavioral measure of user mental effort and perceived output usefulness in human-LLM interaction. We conducted a user study (N = 36) to examine how task difficulty (easy vs. hard) and device type (desktop vs. mobile) influence typing behavior and workload (NASA-TLX) during interactions. Our results indicate that hard tasks led to significantly more keystrokes, slower typing, increased pauses, and higher self-reported workload. Device type had weaker effects, with mobile use slightly reducing input length and typing speed. While keystrokes captured differences in cognitive effort, they did not predict perceived LLM output usefulness. These findings highlight the potential of keystroke dynamics as real-time indicators of cognitive effort during LLM prompting, while also showing their limitations in capturing perceived collaboration success.

Problem

• keystroke dynamics
• cognitive effort
• human-LLM interaction
• prompting
• typing behavior

Innovation

• keystroke dynamics
• cognitive effort
• human-LLM interaction
• prompting behavior
• behavioral sensing