Dr. Philipp Spitzer and co-authors from Carnegie Mellon University and the University of Bayreuth have received the Best Paper Award from ACM Transactions on Interactive Intelligent Systems (TiiS) for their paper “Imperfections of XAI: Phenomena Influencing AI-Assisted Decision-Making.” The work examines how incorrect AI explanations affect human decision-making, with a central finding: they can degrade it. As AI-assisted decision-making becomes more prevalent, the quality of explanations provided by these systems is increasingly critical. The paper examines imperfect explanations in explainable AI (XAI) systems, analyzing both their behavioral effects and the human-centric factors that mediate them. The research demonstrates that users are influenced by incorrect explanations in ways that can worsen their decision-making performance.

The project was initiated through a collaboration between Carnegie Mellon University and the Karlsruhe Institute of Technology, led by Philipp Spitzer and Katelyn Morrison. This resulted in a research stay at the CMU Human-Computer Interaction Institute and eventually led to a joint effort across KIT, CMU, and the University of Bayreuth. The team continues to investigate the effects of incorrect explanations to further consolidate findings in this area. The paper was co-authored by Katelyn Morrison, Niklas Kühl, Adam Perer, Michelle Feng, and Violet Turri. If you are interested and want to give it a full read, follow the link: https://dl.acm.org/doi/full/10.1145/3750052