As AI systems become embedded in consequential decisions across finance, healthcare, energy, legal justice, and defence, explainability has shifted from a technical desideratum to an organisational and regulatory imperative. Yet what counts as sufficient explanation varies by domain, stakeholder, and decision context — and the tension between model performance and interpretability forces trade-offs that are rarely made explicit.
This panel, sponsored by Accenture, brings together some experts from high-stakes industries to examine what governing AI responsibly looks like in practice: who bears accountability when a system fails, how organisations decide when an explanation is good enough, and whether the commercial adoption of XAI advances or dilutes the ethical imperatives that motivated the field. The session aims to surface actionable insights for researchers, practitioners, and policymakers alike.
