Explainable AI does not provide the explanations end-users are asking for

TL;DR

This paper argues that current explainable AI methods often fail to meet end-user needs and suggests that transparency and validation are more effective for building trust in AI systems.

Contribution

The paper critically examines the limitations of explainable AI in real-world deployment and advocates for transparency and validation as more reliable trust-building approaches.

Findings

XAI techniques often do not meet end-user expectations

Transparency and validation are more effective for trust

Limitations of XAI in organizational deployment

Abstract

Explainable Artificial Intelligence (XAI) techniques are frequently required by users in many AI systems with the goal of understanding complex models, their associated predictions, and gaining trust. While suitable for some specific tasks during development, their adoption by organisations to enhance trust in machine learning systems has unintended consequences. In this paper we discuss XAI's limitations in deployment and conclude that transparency alongside with rigorous validation are better suited to gaining trust in AI systems.