Personalized Rationales: Tailoring Natural Language Explanations to User Expertise in Healthcare and Education

Gurrapu et al. (2023) argue that rationalization—generating natural language justifications for model outputs—makes explanations more accessible to non-technical users. However, current rationalization approaches typically produce generic explanations, not accounting for the varied expertise of end-users in healthcare (doctors, patients) or education (teachers, students). This research would develop user-adaptive XAI models that dynamically adjust their explanations’ language and detail, using user profiles or context detection. For example, a diagnostic system might explain its reasoning in layperson’s terms for patients, but use more technical reasoning or citations for clinicians. Similarly, in education, explanations could be tailored for novice vs. advanced students. This extension brings together rationalization, user modeling, and adaptive NLP—offering a new level of personalization in explainable AI and significantly enhancing comprehension and trust.

References:

1. Rationalization for explainable NLP: a survey. Sai Gurrapu, Ajay Kulkarni, Lifu Huang, Ismini Lourentzou, Laura J. Freeman, Feras A. Batarseh (2023). Frontiers Artif. Intell..