Evaluating Explanations Through LLMs: Beyond Traditional User Studies

TL;DR

This paper investigates using Large Language Models to replicate human user studies for evaluating explainable AI tools, aiming to make the process more scalable and cost-effective.

Contribution

It demonstrates that LLMs can replicate key findings of human user studies in XAI evaluation, highlighting their potential as scalable evaluation tools.

Findings

LLMs can replicate most conclusions from human studies

Different LLMs show varying levels of alignment

Factors like LLM memory and output variability influence results

Abstract

As AI becomes fundamental in sectors like healthcare, explainable AI (XAI) tools are essential for trust and transparency. However, traditional user studies used to evaluate these tools are often costly, time consuming, and difficult to scale. In this paper, we explore the use of Large Language Models (LLMs) to replicate human participants to help streamline XAI evaluation. We reproduce a user study comparing counterfactual and causal explanations, replicating human participants with seven LLMs under various settings. Our results show that (i) LLMs can replicate most conclusions from the original study, (ii) different LLMs yield varying levels of alignment in the results, and (iii) experimental factors such as LLM memory and output variability affect alignment with human responses. These initial findings suggest that LLMs could provide a scalable and cost-effective way to simplify qualitative XAI evaluation.