Editable XAI: Toward Bidirectional Human-AI Alignment with Co-Editable Explanations of Interpretable Attributes

TL;DR

This paper introduces Editable XAI, a novel approach allowing users to modify explanations and rules to enhance understanding and alignment with AI models, demonstrated through a neural-symbolic system and user study.

Contribution

It presents CoExplain, a system enabling bidirectional human-AI alignment by making explanations editable and incorporating user-written rules into neural network models.

Findings

Editable XAI improves user understanding and model alignment.

CoExplain is easier to use with fewer edits and less time.

User study shows enhanced control and comprehension.

Abstract

While Explainable AI (XAI) helps users understand AI decisions, misalignment in domain knowledge can lead to disagreement. This inconsistency hinders understanding, and because explanations are often read-only, users lack the control to improve alignment. We propose making XAI editable, allowing users to write rules to improve control and gain deeper understanding through the generation effect of active learning. We developed CoExplain, leveraging a neural network for universal representation and symbolic rules for intuitive reasoning on interpretable attributes. CoExplain explains the neural network with a faithful proxy decision tree, parses user-written rules as an equivalent neural network graph, and collaboratively optimizes the decision tree. In a user study (N=43), CoExplain and manually editable XAI improved user understanding and model alignment compared to read-only XAI. CoExplain was easier to use with fewer edits and less time. This work contributes Editable XAI for bidirectional AI alignment, improving understanding and control.