Keystroke Dynamics Against Academic Dishonesty in the Age of LLMs

TL;DR

This paper introduces a keystroke dynamics-based method to detect academic dishonesty involving AI-assisted writing, achieving up to 85.72% accuracy, thereby improving integrity verification in online education.

Contribution

It presents a novel keystroke pattern analysis approach using a modified TypeNet architecture to distinguish genuine from AI-assisted writing in academic settings.

Findings

Keystroke dynamics differ significantly between genuine and assisted writing.

The proposed detector achieved up to 85.72% accuracy in condition-specific scenarios.

The study provides insights into user-AI interaction patterns in educational contexts.

Abstract

The transition to online examinations and assignments raises significant concerns about academic integrity. Traditional plagiarism detection systems often struggle to identify instances of intelligent cheating, particularly when students utilize advanced generative AI tools to craft their responses. This study proposes a keystroke dynamics-based method to differentiate between bona fide and assisted writing within academic contexts. To facilitate this, a dataset was developed to capture the keystroke patterns of individuals engaged in writing tasks, both with and without the assistance of generative AI. The detector, trained using a modified TypeNet architecture, achieved accuracies ranging from 74.98% to 85.72% in condition-specific scenarios and from 52.24% to 80.54% in condition-agnostic scenarios. The findings highlight significant differences in keystroke dynamics between genuine and assisted writing. The outcomes of this study enhance our understanding of how users interact with generative AI and have implications for improving the reliability of digital educational platforms.