Operationalizing the NIST AI Risk Management Framework for AI-Driven Autism Care Systems: A Governance-Centric Technical Architecture

Abstract

Artificial intelligence–driven systems are increasingly deployed in autism spectrum disorder care to support behavioral monitoring, escalation prediction, and caregiver decision-making. While these systems offer substantial clinical and operational benefits, they also introduce risks related to safety, privacy, bias, accountability, and trust—particularly when applied to vulnerable pediatric populations. The NIST Artificial Intelligence Risk Management Framework provides high-level guidance for managing AI risks, yet practical operationalization within real-world autism care systems remains limited. This study proposes a governance-centric technical architecture that embeds NIST AI Risk Management Framework principles directly into the design, deployment, and lifecycle management of AI-driven autism care systems. The framework translates abstract governance functions into concrete technical controls, audit artifacts, and operational workflows aligned with reinforcement learning, IoT-based behavioral monitoring, and caregiver decision support. Through a structured risk taxonomy, control catalog, and implementation blueprint, this research demonstrates how trustworthy AI principles can be transformed into deployable, auditable, and scalable autism care solutions.