Postdoctoral Researcher Poslon at the Workshop on Explainable Artificial Intelligence

Postdoctoral Researcher Luka Poslon from the Catholic University of Croatia participated in the international scientific workshop Philosophy of Explainable AI – New Directions, held from 7 to 9 July 2026 at Aarhus University, Denmark. Together with Johannes Gebert, PhD, from the High-Performance Computing Center at the University of Stuttgart, he presented a paper entitled Towards Responsible AI in Transplant Medicine: Normative Criteria for Epistemic Justification. The workshop was organized as part of the Towards Responsible Explainable AI Technologies (TREAT) research project at Aarhus University, which explores how to develop ethical and trustworthy explainable artificial intelligence (XAI) systems and establish theoretical and practical guidelines for their responsible use.

The presentation addressed the ethical and philosophical challenges of applying artificial intelligence in transplant medicine, with a particular focus on XAI. The authors argued that the mere ability of AI systems to provide explanations is not sufficient to establish trust. Rather, such explanations must be grounded in scientifically justified and transparent epistemic criteria.

The paper introduced a normative framework for evaluating the epistemic justification of AI in kidney transplantation, based on the principles of fairness, accountability, and equity. The proposed criteria are intended to support the more responsible development and deployment of AI systems in healthcare decision-making, particularly in the allocation of scarce medical resources, where every decision carries significant ethical implications.

Participation in this international workshop contributes to the ongoing global scientific discussion on the responsible use of artificial intelligence in healthcare and highlights the Catholic University of Croatia's active involvement in interdisciplinary research at the intersection of philosophy, bioethics, and emerging digital technologies.