Digital twins, which rely on surrogate models to replicate physical systems in real-time, face a significant challenge when operating conditions change, causing these models to degrade. This degradation, known as concept drift, can be particularly problematic when models need to account for aleatoric uncertainty. To address this issue, researchers have proposed a continual validation, updating, and decision-making framework that utilizes robust model predictive control1. This framework is particularly relevant in additive manufacturing, where digital twins can be used to optimize production processes. The framework enables digital twins to adapt to changing conditions and maintain their fidelity, even in the presence of uncertainty. This is crucial in industries where digital twins are used to make critical decisions, as inaccurate models can have significant consequences. The development of such a framework has significant implications for the effective deployment of digital twins in various fields, so what matters to practitioners is the potential to enhance the reliability and accuracy of digital twins in complex, dynamic environments.