The Governance–adoption asymmetry of artificial intelligence: A systematic review of educational, labour, and cybersecurity policy gaps

Artificial intelligence (AI) has emerged as one of the most transformative technologies of the twenty-first century. This paper presents a systematic multidimensional review examining the dual nature of AI across three critical domains: education, employment, and cybersecurity. Unlike prior reviews that aggregate evidence without an interpretive structure, this analysis introduces the Governance–Adoption Asymmetry (GAA) Framework, an original five-dimensional analytical tool comprising Adoption Velocity, Governance Readiness, Distributional Equity, Evidence Maturity, and Augmentation–Displacement Ratio. Drawing upon 68 sources identified through systematic screening from an initial pool of 412 records, the analysis reveals that cybersecurity exhibits the highest governance–adoption asymmetry (GAA score: 3.8/5.0), followed by education (3.2/5.0) and employment (2.7/5.0). In education, AI tutoring systems produce statistically significant learning gains in controlled settings. In employment, McKinsey’s November 2025 estimate that 57% of U.S. work hours are technically automatable contrasts with the Budget Lab at Yale’s finding of no significant AI-driven employment decline. In cybersecurity, AI-powered defences save USD 1.88 million per breach even as AI-generated phishing has surged 1,265%. Seven time-bound, actor-specific policy recommendations emerge from the framework analysis.