Researchers propose an Inverse Theory of Mind (IToM) pipeline to better understand user behavior in dynamic interfaces. The IToM pipeline infers users' beliefs, preferences, and decision-making traits from observed interactions, reconstructing each user's decision context and applying counterfactual reasoning to produce evidence-grounded belief statements. This approach is evaluated on the OPeRA dataset against ground-truth personality assessments and shows promising results in next action prediction, shopping attitude alignment, Big Five personality inference, and held-out category prediction. --- Why it matters: This research matters because it tackles a fundamental challenge in AI: understanding human behavior in dynamic environments. The proposed pipeline can improve content recommendation systems and adaptive interfaces by providing more accurate user personas, enabling better decision-making and personalization. Source: https://arxiv.org/abs/2608.11354