Deciding Fast and Slow: The Role of Cognitive Biases in AI-assisted Decision-making

Several strands of research have aimed to bridge the gap between artificial intelligence (AI) and human decision-makers in AI-assisted decision-making, where humans are consumers AI model predictions ultimate high-stakes applications. However, people's perception understanding often distorted by their cognitive biases, such as confirmation bias, anchoring availability name a few. In this work, we use knowledge from field science account for biases human-AI collaborative decision-making setting, mitigate negative effects on performance. To end, mathematically provide general framework through which researchers practitioners can understand interplay accuracy. We then focus specifically bias commonly encountered collaboration. implement time-based de-anchoring strategy conduct our first user experiment that validates its effectiveness decision-making. With result, design time allocation resource-constrained setting achieves optimal collaboration under some assumptions. We, then, second shows with explanation effectively de-anchor improve performance when has low confidence is incorrect.