Towards Explainability as a Functional Requirement: A Vision to Integrate the Legal, End-User, and ML Engineer Perspectives

The rapid advancement and integration of artificial intelligence (AI) in various sectors have accentuated the importance of explainability in AI systems. This vision paper presents an exploration into the multifaceted nature of AI explainability, consolidating insights from three critical perspectives: legal, end-user, and ML engineer. We work towards a comprehensive taxonomy by creating a detailed classification that integrates various viewpoints and establishes explainability as an essential functional requirement, despite its usual treatment as a non-functional requirement.Our taxonomy will guide towards the interdependencies and distinct requirements of each perspective. We aim to foster a more profound understanding of AI explainability, encouraging a more holistic approach to AI system development. This research will advance the explainable AI discussion, providing key insights for policymakers, developers, and users, and promoting the development of AI systems that are technically sound, trustworthy, understandable, and legally compliant.