As artificial intelligence (AI) performs increasingly complex cognitive tasks, human decision-makers must closely collaborate with AI to monitor, interpret, evaluate, and refine its results (augmented decision-making). Augmented decision-making can place high metacognitive demands on human decision-makers, who must monitor and control not only their own cognitive processes but must also assess the accuracy of AI results and act upon these assessments. To understand the role of individuals’ metacognitive ability in augmented decision-making, we will analyze data from a large-scale experiment on medical decision-making. In this experiment, N = 117 medical students and medical doctors made confidence judgments for 150 diagnoses based on CT scans in one unsupported trial without AI support and two AI-supported trials. After determining which measure of metacognitive accuracy is best suited for assessing decision-makers’ metacognitive abilities, we will examine the association between metacognitive accuracy in unsupported trials and the decision-making performance in AI-augmented trials. We expect that high metacognitive competencies (1) positively predict performance in AI-augmented decision-making and (2) foster adaptive reliance on AI advice. This research will provide insights into the contributions of metacognition to successful AI-augmented decision-making.