The true purpose of this simulation was not to test your legal knowledge, but to study
Automation Bias — how we trust AI systems under pressure.
Different participants received different AI assistant styles: some confident and direct, others cautious and hedged, and some had the option to verify claims against source documents.
The AI's advice was sometimes
incorrect. We are interested in whether participants followed the AI regardless, and whether access to sources changed behaviour.
Your anonymised data contributes to research on human-AI decision-making. Thank you.
SurveyCircle Users: Redeem Survey Code with one click
Prolific participants — enter this code to confirm completion:
CBE40X6N