Between 2023 and July 2025, artificial intelligence (AI) governance in the USA, European Union, and China shifted from programmatic statements to actionable instruments. The USA moved from Executive Order 14110 to three July 2025 executive orders on data-center permitting, export promotion, and procurement neutrality. The European Union completed the AI Act, initiated staged application in 2025, and issued a code of practice for general-purpose AI. China consolidated domestic controls on public-facing generative AI and launched a Global AI Governance Action Plan with United Nations-centered cooperation, standards work, and capacity-building. The UK continued a regulator-led, assurance-first model. This essay compares these trajectories and distils implications for libraries: stronger accountability in procurement and vendor management; lawful, well-described training data; the publication of assessment artifacts; and AI literacy as a core service. The analysis highlights convergence on safety, transparency, and inclusion, alongside divergence in regulatory technique and international posture, which will shape library strategy.
Leo S. Lo (2025) studied this question.