Beyond XAI:Obstacles Towards Responsible AI
It addresses challenges in making AI systems more transparent and accountable for researchers and practitioners, but is incremental as it builds on existing XAI critiques.
The paper identifies limitations in current explainable AI (XAI) methods and evaluation strategies, and discusses how these issues impact broader responsible AI concerns such as privacy, fairness, and contestability.
The rapidly advancing domain of Explainable Artificial Intelligence (XAI) has sparked significant interests in developing techniques to make AI systems more transparent and understandable. Nevertheless, in real-world contexts, the methods of explainability and their evaluation strategies present numerous limitations.Moreover, the scope of responsible AI extends beyond just explainability. In this paper, we explore these limitations and discuss their implications in a boarder context of responsible AI when considering other important aspects, including privacy, fairness and contestability.