Participate
The repeated-decision experiment
Sixty short decisions with an AI advisor, in two task families. About 12–15 minutes.
Each trial has the same shape: you answer, rate your confidence, see the AI's recommendation, and answer again. Depending on the condition you are assigned to, the AI may show a confidence level and you may or may not see the correct answer after each trial.
Current design
Reliance & confidence — 2 (AI accuracy: 0.70, 0.90) × 3 (confidence display: none, calibrated, miscalibrated); explanation none; immediate feedback. Conditions are assigned at random and balanced across participants. Block order is counterbalanced.
Checking collection status…
Two ways this page runs
In live collection your responses are stored on the research server and contribute to the study. In demo mode nothing leaves your browser: you experience the exact same experiment and can download your own responses at the end. The status badge above tells you which one this deployment is in.