Narrative-Driven Explanations: Leveraging Qualitative User Narratives

The open-ended qualitative studies by Panigutti et al. (2022) and Ding et al. (2024) reveal that user perceptions of explanations are nuanced and context-dependent, shaped by personal narratives and lived experiences. Rather than relying solely on quantitative metrics or static user profiles, this research proposes collecting brief user narratives—either via in-situ prompts or post-hoc interviews—about how they interpreted and acted on specific AI explanations. These narratives would be analyzed (possibly with NLP techniques) to identify themes, misconceptions, or recurring needs, which would then inform iterative improvement of explanation strategies. This approach fuses human-computer interaction (HCI) and cognitive psychology with XAI, creating a richer feedback loop for explanation design. The result could be explanations that are better calibrated to real-world user contexts and mental models.

References:

1. Understanding the impact of explanations on advice-taking: a user study for AI-based clinical Decision Support Systems. Cecilia Panigutti, Andrea Beretta, F. Giannotti, Dino Pedreschi (2022). International Conference on Human Factors in Computing Systems.
2. Unveiling user responses to AI-powered personalised recommendations: a qualitative study of consumer engagement dynamics on Douyin. Liang Ding, G. Antonucci, M. Venditti (2024). Qualitative Market Research.