“Less is More”: Challenging the Completeness Assumption in Explanations

Much XAI work, as catalogued by Rong et al. (2022) and Naveed et al. (2024), equates more information or higher fidelity to better explanations. However, Nimmo et al. (2024) and Kim et al. (2024) suggest that user preferences and cognitive load are often overlooked, and that explanation completeness might actually hinder understanding. This idea proposes a systematic exploration of selective explanations: providing users with only the most actionable or relevant information based on their context, role, or expressed needs, rather than a comprehensive breakdown. The research would experimentally compare user-centered, minimal explanations to traditional “full” explanations across various domains and user types. The work would help redefine what “good” explanations mean in practice and could lead to new metrics for “explanation sufficiency,” potentially challenging a core XAI assumption.

References:

1. User Characteristics in Explainable AI: The Rabbit Hole of Personalization?. Robert Nimmo, Marios Constantinides, Ke Zhou, D. Quercia, Simone Stumpf (2024). International Conference on Human Factors in Computing Systems.
2. Design Opportunities for Explainable AI Paraphrasing Tools: A User Study with Non-native English Speakers. Yewon Kim, Thanh-Long V. Le, Donghwi Kim, Mina Lee, Sung-Ju Lee (2024). Conference on Designing Interactive Systems.
3. An Overview of the Empirical Evaluation of Explainable AI (XAI): A Comprehensive Guideline for User-Centered Evaluation in XAI. Sidra Naveed, Gunnar Stevens, Dean Robin-Kern (2024). Applied Sciences.
4. Towards Human-Centered Explainable AI: A Survey of User Studies for Model Explanations. Yao Rong, Tobias Leemann, Thai-trang Nguyen, Lisa Fiedler, Peizhu Qian, Vaibhav Unhelkar, Tina Seidel, Gjergji Kasneci, Enkelejda Kasneci (2022). IEEE Transactions on Pattern Analysis and Machine Intelligence.
5. Human-centered evaluation of explainable AI applications: a systematic review. Jenia Kim, Henry Maathuis, Danielle Sent (2024). Frontiers Artif. Intell..