Drawing on 26 UNDP Artificial Intelligence Landscape Assessments (AILA) completed between 2024 and 2026, with 10 more underway, this report examines what AI readiness means once adoption is already underway. The countries analysed reflect the set of national AILA engagements completed at the time of writing, undertaken in response to government demand and UNDP country-level programming priorities across diverse regional, institutional, and digital maturity contexts. They are not presented as a statistically representative sample, and the report does not rank countries, aggregate readiness scores, or provide a country-by-country review. Instead, the report uses this evidence base to identify recurring patterns in how AI adoption is moving through public systems, markets, institutions, and wider ecosystems.

UCL Partners helps accelerate digital health, patient safety and health innovation across the NHS through transformation, adoption and innovation support programmes.

The central finding is straightforward. Many of the most consequential AI governance decisions are not being made in policy frameworks. They are being made in procurement offices, vendor contracts, software update cycles, and the informal decisions of civil servants and institutional teams navigating AI use without clear guidance, approval processes, or accountability pathways. Getting governance right means embedding it in these operational settings, not treating it as a separate layer above them.

A new European Commission study examines how Secondary Publication Rights and copyright exceptions could support a more open, reusable, and competitive European research system.

This paper examines how governments are translating international principles for the use of AI into practice across public institutions. By reviewing global experiences, seven practice areas were identified. These demonstrate how governments have taken concrete steps to operationalize principles and highlight several key takeaways as well as providing some cautionary tales from early adopters. Each practice area is linked to relevant principles and evidence of good practice, showcased through diverse country examples. To support implementation, the paper outlines a set of suggested activities, presented in annexed tables, which offer practical techniques governments may consider to operationalize responsible AI, reduce fragmentation, and build institutional capacity across the AI lifecycle.