Abstract:For synergistic interactions between humans and artificial intelligence (AI) systems, AI outputs often need to be explainable to people. Explainable AI (XAI) systems are commonly tested in human user studies. However, whether XAI researchers consider potential cultural differences in human explanatory needs remains unexplored. We highlight psychological research that found significant differences in human explanations between many people from Western, commonly individualist countries and people from non-Western, often collectivist countries. We argue that XAI research currently overlooks these variations and that many popular XAI designs implicitly and problematically assume that Western explanatory needs are shared cross-culturally. Additionally, we systematically reviewed over 200 XAI user studies and found that most studies did not consider relevant cultural variations, sampled only Western populations, but drew conclusions about human-XAI interactions more generally. We also analyzed over 30 literature reviews of XAI studies. Most reviews did not mention cultural differences in explanatory needs or flag overly broad cross-cultural extrapolations of XAI user study results. Combined, our analyses provide evidence of a cultural bias toward Western populations in XAI research, highlighting an important knowledge gap regarding how culturally diverse users may respond to widely used XAI systems that future work can and should address.

What problem does this paper attempt to address?

The main problem that this paper attempts to solve is the problem of cultural bias in current Explainable AI (XAI) research. Specifically, the author focuses on whether XAI systems fully consider the differences in people's needs for explanations in different cultural backgrounds during the design and evaluation process. The paper points out that most existing XAI research is tested on individuals from Western, Educated, Industrialized, Rich, and Democratic (WEIRD) countries, and it is assumed that these results can be generalized to people in other cultural backgrounds around the world. However, this approach ignores the significant differences between culture and the need for explanations, especially the differences between individualistic cultures (such as Western countries) and collectivist cultures (such as non - Western countries). ### Summary of Main Problems: 1. **The Impact of Cultural Differences on the Need for Explanations**: - Existing XAI research ignores the fact that there may be significant differences in people's needs for explanations in different cultural backgrounds. - For example, individuals from individualistic cultures are more inclined to explanations of internal factors (such as personal beliefs, desires, etc.), while individuals from collectivist cultures are more inclined to explanations of external factors (such as social norms, situations, etc.). 2. **The Problem of Over - Generalization**: - Most XAI user studies are limited to samples from WEIRD countries, but the conclusions are often widely generalized to all populations. - This approach may lead to "hasty generalization", that is, the scope of the conclusion exceeds the scope supported by actual evidence, thus ignoring cultural diversity. 3. **Lack of Consideration of Cultural Diversity**: - The research finds that most XAI user studies do not consider cultural differences and do not report the cultural background information of the samples. - At the same time, the literature review also fails to point out these problems, further exacerbating the existence of cultural bias. ### Goals of the Paper: - **Reveal Cultural Bias**: Through a systematic analysis of existing XAI user studies, reveal the problem of cultural bias therein. - **Provide Improvement Suggestions**: Suggest that future research should pay more attention to cultural diversity to ensure that the explanations of XAI systems can meet the needs of users in different cultural backgrounds. - **Promote Cross - Cultural Communication**: Emphasize the importance of considering cultural differences in XAI research to promote broader cross - cultural communication and understanding. Through these efforts, the paper aims to fill an important knowledge gap in XAI research and provide guidance for future research to ensure that XAI systems are more universal and fair on a global scale.

Cultural Bias in Explainable AI Research: A Systematic Analysis

Cultural Bias in Explainable AI Research: A Systematic Analysis

On the Importance of User Backgrounds and Impressions: Lessons Learned from Interactive AI Applications

Initial evidence for biased decision-making despite human-centered AI explanations

Towards Human-centered Design of Explainable Artificial Intelligence (XAI): A Survey of Empirical Studies

On the Influence of Cognitive Styles on Users' Understanding of Explanations

Unjustified Sample Sizes and Generalizations in Explainable AI Research: Principles for More Inclusive User Studies

Evaluating the Influences of Explanation Style on Human-AI Reliance

Designing Theory-Driven User-Centric Explainable AI

Towards Human-centered Explainable AI: A Survey of User Studies for Model Explanations

On Two XAI Cultures: A Case Study of Non-technical Explanations in Deployed AI System

Expl(AI)ned: The Impact of Explainable Artificial Intelligence on Users' Information Processing

The Who in XAI: How AI Background Shapes Perceptions of AI Explanations

Towards a Praxis for Intercultural Ethics in Explainable AI

Interdisciplinary Expertise to Advance Equitable Explainable AI

On the Influence of Explainable AI on Automation Bias

The Drawback of Insight: Detailed Explanations Can Reduce Agreement with XAI

Who Made That Decision and Why? Users' Perceptions of Human Versus AI Decision-Making and the Power of Explainable-AI

Selective Explanations: Leveraging Human Input to Align Explainable AI

Can Explainable AI Explain Unfairness? A Framework for Evaluating Explainable AI

A Multidisciplinary Survey and Framework for Design and Evaluation of Explainable AI Systems