Ethical Governance Framework for AI Deployment in Education and Public Administration: An Empirical Documentary Analysis

Artificial intelligence (AI) is being deployed rapidly in schools, universities and government agencies, yet it remains unclear whether existing governance instruments convert ethical principles into enforceable mechanisms. This empirical study used a structured documentary content analysis of 12 governance instruments issued between 2019 and 2025: four education-specific, four oriented towards public administration and four cross-sector. Each instrument was coded against ten ethical governance dimensions derived from the literature, using a three-point scale (0 = absent, 1 = principle stated, 2 = mechanism specified). Instruments scored a mean of 13.3 out of 20 (SD = 2.96); 39.2% of coded cells specified a mechanism and 5.8% were absent. Human oversight (M = 1.50) was the best-covered dimension, whereas inclusion, equity and access (M = 1.17) was the weakest. Education-specific instruments scored lowest on accountability and on monitoring and enforcement (both M = 0.75), while public-administration instruments scored lowest on inclusion (M = 0.75). Total scores did not differ significantly across instrument types, H (2) = 1.49, p =.476, but coverage of risk management and of monitoring were positively associated, ρ =.62, p =.032. Building on these gaps, the paper proposes a five-layer Ethical Governance Framework for AI Deployment in Education and Public Administration (EGF-AIEPA) and sets out implications for policymakers in Nigeria and comparable contexts. Findings are exploratory because of the small purposive sample and single-coder design.