[Paper Review] Explainable AI is Responsible AI: How Explainability Creates Trustworthy and Socially Responsible Artificial Intelligence

Stephanie Baker, Wei Xiang|arXiv (Cornell University)|Dec 4, 2023

Ethics and Social Impacts of AI12 citations

TL;DR

The paper argues that Explainable AI (XAI) is the foundational technology for Responsible AI (RAI), enabling fairness, robustness, privacy, security, and transparency across contexts, with a novel framework and real-world use cases.

ABSTRACT

Artificial intelligence (AI) has been clearly established as a technology with the potential to revolutionize fields from healthcare to finance - if developed and deployed responsibly. This is the topic of responsible AI, which emphasizes the need to develop trustworthy AI systems that minimize bias, protect privacy, support security, and enhance transparency and accountability. Explainable AI (XAI) has been broadly considered as a building block for responsible AI (RAI), with most of the literature considering it as a solution for improved transparency. This work proposes that XAI and responsible AI are significantly more deeply entwined. In this work, we explore state-of-the-art literature on RAI and XAI technologies. Based on our findings, we demonstrate that XAI can be utilized to ensure fairness, robustness, privacy, security, and transparency in a wide range of contexts. Our findings lead us to conclude that XAI is an essential foundation for every pillar of RAI.

Motivation & Objective

• Ground the role of XAI within Responsible AI (RAI) frameworks across multiple pillars (fairness, robustness, transparency, accountability, privacy, safety).
• Propose a framework that positions explainability as the foundation of RAI rather than a single pillar.
• Synthesize evidence showing how XAI supports RAI pillars and enables improvement in real-world use cases.
• Demonstrate through sector-focused use cases (generative AI, healthcare, transportation) how XAI directly supports RAI in practice.

Proposed method

• Conduct a broad-scoping review of the literatures on XAI and RAI.
• Propose a novel framework with explainability as the foundation of RAI.
• Rigorous exploration showing XAI underpins fairness, robustness, privacy, safety, transparency, and accountability.
• Illustrate real-world use cases in generative AI, healthcare, and transportation to demonstrate practical impact.

Figure 1 . Comparison of RAI frameworks (a) as commonly defined in the literature with explainability as a pillar, and (b) as proposed by this work with explainability as the foundational concept.

Experimental results

Research questions

• RQ1 How can Explainable AI be conceptualized as the foundation rather than a pillar of Responsible AI?
• RQ2 What evidence from XAI literature demonstrates its contribution to fairness, robustness, privacy, safety, and transparency in RAI?
• RQ3 How can XAI be applied to real-world use cases to enhance trustworthiness and social responsibility of AI systems?
• RQ4 What recommendations arise for future AI research and deployment when XAI is treated as foundational to RAI?

Key findings

• XAI can support and enhance all RAI pillars: fairness, robustness, privacy, safety, transparency, and accountability.
• XAI provides transparency that improves interpretability and understanding, aiding accountability in sectors like finance and law.
• XAI can be combined with privacy-preserving techniques to meet privacy requirements without compromising other RAI needs.
• The authors propose that XAI should be foundational to RAI, not merely a separate pillar, and illustrate this with real-world use cases.
• The work fills a gap by linking XAI and RAI in a unified framework and highlighting practical applications in healthcare, transportation, and generative AI.

Figure 2 . Overview of the key contents of this paper