Kevin Bauer and Michael Kosfeld, together with Oliver Hinz and Rebecca Heigl, have received the AIS Senior Scholars Best IS Publication Award for their paper “Feedback Loops in Machine Learning: A Study on the Interplay of Continuous Updating and Human Discrimination.”
The study shows that machine learning systems not only influence decisions but also shape the future data on which they are trained. When algorithmic predictions affect outcomes, such as credit approvals, they can generate feedback loops that alter the learning environment itself. Using extensive simulations combined with secondary empirical data, the authors provide evidence that continuous model updating can reduce discrimination over time and improve economic efficiency. However, persistent human decision-making in the loop may obstruct these self-correcting dynamics and can even cause initially unbiased models to become discriminatory. The findings highlight the critical role of human oversight in determining whether adaptive AI systems ultimately promote fairness or reinforce bias.
The paper previously received the 2024 Best Paper Award from the Journal of the Association for Information Systems.
More about the publication is available here.