EXPLAINABILITY REQUIREMENTS FOR AI DECISION-MAKING IN REGULATED SECTORS

The increasing deployment of artificial intelligence (AI) systems in regulated sectors such as finance, public
administration, healthcare, pharmaceuticals, and cloud-based business intelligence has intensified regulatory and
societal demands for explainability in automated decision-making. As AI-driven decisions increasingly affect
legal rights, economic outcomes, and public trust, explainability has emerged as a foundational requirement for
accountability, transparency, and lawful governance rather than a purely technical feature. This article examines
the explainability requirements for AI decision-making in regulated sectors, with a specific focus on how legal
frameworks, regulatory instruments, and ethical principles shape expectations for interpretable and transparent
AI systems. Drawing on recent scholarship and regulatory discourse, the study situates explainable AI (XAI)
within evolving governance regimes, including sector-specific compliance obligations and cross-jurisdictional
regulatory initiatives. The analysis highlights how explainability functions as a mechanism for enabling human
oversight, supporting auditability, and facilitating reason-giving obligations in high-stakes decision contexts.
Furthermore, the paper synthesizes theoretical and applied perspectives on XAI to clarify the distinction
between technical interpretability, meaningful explanations for affected stakeholders, and legally actionable
transparency. By integrating insights from law, policy, and AI research, this article proposes a structured
understanding of explainability requirements that aligns technical design choices with regulatory expectations.
The findings underscore that explainability must be context-sensitive, proportionate to risk, and embedded
across the AI lifecycle to ensure trustworthy and compliant AI deployment in regulated environments