Explainable AI: roles and stakeholders, desirements and challenges

Abstract

IntroductionThe purpose of the Stakeholder Playbook is to enable the developers of explainable AI systems to take into account the different ways in which different stakeholders or role-holders need to “look inside” the AI/XAI systems.MethodWe conducted structured cognitive interviews with senior and mid-career professionals who had direct experience either developing or using AI and/or autonomous systems.ResultsThe results show that role-holders need access to others (e.g., trusted engineers and trusted vendors) for them to be able to develop satisfying mental models of AI systems. They need to know how it fails and misleads as much as they need to know how it works. Some stakeholders need to develop an understanding that enables them to explain the AI to someone else and not just satisfy their own sense-making requirements. Only about half of our interviewees said they always wanted explanations or even needed better explanations than the ones that were provided. Based on our empirical evidence, we created a “Playbook” that lists explanation desires, explanation challenges, and explanation cautions for a variety of stakeholder groups and roles.DiscussionThis and other findings seem surprising, if not paradoxical, but they can be resolved by acknowledging that different role-holders have differing skill sets and have different sense-making desires. Individuals often serve in multiple roles and, therefore, can have different immediate goals. The goal of the Playbook is to help XAI developers by guiding the development process and creating explanations that support the different roles.

When to choose ranked area integrals versus integrated gradient for explainable artificial intelligence – a comparison of algorithms

“…(3) Using abstractions to simplify explanations: High-level patterns are the basis for describing a big plan's little steps. Automating the discovery of abstractions has long been a challenge and understanding the discovery and sharing of abstractions in learning and explanation are at the frontier of XAI research today (Kuppa and Le-Khac, 2020;Hoffman et al, 2018).…”