Reproduce “Scientific Explanations in Health Sciences: Causality, Trust, and Epistemic Adequacy” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Martina Mattioli, Marcello Pelillo
Mattioli and Pelillo propose three philosophically grounded design principles for medical Explainable AI (XAI) to address the conceptual gaps in current research, advocating for explanations that support intervention, are stakeholder-specific, and rigorously assess trust. The paper highlights that existing XAI methods often provide limited causal insight and may fail to foster justified confidence in clinical settings.
Key insights
- Medical explanations must explicitly support intervention by providing causal insights rather than just correlations, aligning with medicine's action-oriented nature.
- Explanations in healthcare require stakeholder-specific approaches, adapting content (causal, statistical, narrative) to the distinct epistemic needs of clinicians, patients, and regulators.
- Trust in medical AI systems must be rigorously assessed to distinguish between superficial plausibility and genuine faithfulness, fostering justified reliance consistent with the fiduciary nature of medical relationships.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.