PHAX: A Structured Argumentation Framework for User-Centered Explainable AI in Public Health and Biomedical Sciences

One-size-fits-all explanations don't work in healthcare: PHAX demonstrates that AI explanations need to be tailored to different audiences—clinicians, policymakers, and patients—rather than providing identical justifications to everyone. The framework uses "user modeling" to adapt explanations based on audience expertise, recognizing that a doctor and a patient need different levels of technical detail to understand the same AI recommendation.

Structured argumentation makes AI reasoning more transparent: Rather than treating AI explanations as black boxes, PHAX uses formal argumentation to break down AI decisions into logical chains of reasoning. This approach allows stakeholders to see not just what the AI decided, but why it made that decision and what assumptions underpin that reasoning.

Medical jargon is a barrier to trust: The paper highlights medical term simplification as a key use case, showing that translating complex medical language into plain language is essential for patient understanding and trust. By modeling simplification decisions as argument chains, PHAX helps ensure that explanations remain accurate while becoming accessible to non-experts.

Interactive dialogue builds accountability: PHAX enables two-way conversations between users and AI systems rather than one-directional explanations. This interactive capability allows clinicians, patients, and policymakers to ask follow-up questions and challenge AI recommendations, creating a more accountable decision-making process.

Transparency requires both technical rigor and communication skill: The framework bridges formal reasoning methods with communicative demands, suggesting that responsible AI in healthcare cannot rely on technical accuracy alone—it must also prioritize how information is presented and discussed with diverse stakeholders.