Automation Bias and Legal Reasoning in Law Students: Experimental Dataset on Generative AI Errors

<p dir="ltr"><b>This dataset contains experimental data collected from 360 law students to examine automation bias and legal reasoning when using generative artificial intelligence systems that may provide erroneous legal information.</b></p><p dir="ltr">The study includes <b>2,880 student-case observations</b> generated under controlled experimental conditions. Participants were assigned to different AI-assistance conditions designed to evaluate how exposure to accurate or erroneous AI-generated legal recommendations influences legal reasoning, error detection, reliance on AI, and decision-making.</p><p dir="ltr">The dataset supports research on <b>automation bias, generative AI errors, legal reasoning, trust in artificial intelligence, human-AI interaction, and legal education</b>.</p><p dir="ltr">The files include participant-level and case-level information required for reproducible analysis of the experimental design.

The dataset does not contain personally identifying information.</p><p dir="ltr"><b>Participants:</b> 360 law students<br><b>Observations:</b> 2,880 student-case observations<br><b>Research area:</b> Artificial intelligence and legal education<br><b>Study design:</b> Controlled experimental study</p>