Kris Luyten is a researcher affiliated with the Intelligible Interactive Systems research unit at Hasselt University (UHasselt) – Expertise Centre for Digital Media (EDM). He conducts research in uncovering new and interesting ways to interact with complex computing systems and computing systems using some type of embedded intelligence. Within this scope he works on, amongst others, user interface software and technologies for intelligible systems, interface design tools, engineering processes for building intelligible systems, better ways for human-robot interaction, and building intelligent and distributed user interfaces.
Modern decision-making increasingly depends on Artificial Intelligence (AI) systems, particularly in critical and time-sensitive contexts. However, reliance on single-model approaches often results in overtrust or undertrust, negatively impacting decision quality. Our research seeks to introduce a paradigm shift by employing multiple machine learning models, such as Rashomon sets and model multiplicity, to foster appropriate trust levels. By allowing users to adjust their reliance on multiple AI systems based on the context, this approach aims to improve transparency, fairness, and decision effectiveness.