Beyond Model Readiness: Institutional Readiness for AI Deployment in Public Systems

Technical viability is insufficient for successful AI deployment: Both case studies described in the paper—an image-based anthropometric screening tool and a speech-analysis system for early learning risk identification—reached technically sound stages but failed to advance to broader rollout due to institutional barriers rather than model performance issues, highlighting a critical gap in how AI readiness is typically assessed.

Institutional readiness requires five key dimensions: The paper introduces the Institutional Alignment Readiness (IAR) framework covering institutional and operational compatibility, data ecosystem maturity, human oversight capacity, fiscal sustainability, and regulatory alignment readiness—factors that existing responsible AI frameworks often overlook in favor of model-centric evaluation.

Resource-constrained public institutions face acute deployment gaps: The framework is specifically designed for resource-limited settings where the gap between what is technically possible and what can be responsibly deployed is most severe, suggesting that responsible AI deployment requires tailored approaches beyond one-size-fits-all evaluation tools.

Deployment readiness extends beyond technical teams: Successful AI implementation in public systems requires attention to approvals, data arrangements, human oversight structures, operational capacity, ongoing funding, and legal clarity—organizational factors that developers alone cannot control or ensure.

IAR complements rather than replaces existing AI evaluation tools: The framework is positioned as an addition to established responsible AI and model evaluation frameworks, not a replacement, suggesting that comprehensive AI governance requires both technical assessment and institutional readiness evaluation.