Taiwan's vTaiwan and Join platforms used AI-assisted opinion clustering to let citizens petition government (13,853 proposals by 2022, 10M+ visitors) — but …
Taipei, Taiwan
Good practice
Evidence: Descriptive / self-reported
Top 25%73/100
The UK Cabinet Office trialled Microsoft 365 Copilot with 20,000+ civil servants across departments including Companies House and the DWP, reporting self-reported …
Germany's Bundesnetzagentur, Hesse's digitalisation ministry and the federal data-protection commissioner simulated an EU AI Act sandbox with two medical-device start-ups, producing a …
Wiesbaden, Germany
Good practice
Evidence: Descriptive / self-reported
Top 8%73/100
Kazakhstan's Orleu teacher-training agency, with UNESCO IITE and UNESCO-ICHEI, ran a free national generative-AI micro-certification for teachers: by Nov 2025, 324,000 had …
Bedfordshire ran a published evaluation comparing its human-tutor Studiosity writing service with a new AI-augmented version. The 2024 AI trial cut average …
Luton, United Kingdom
Good practice
Evidence: Descriptive / self-reported
Top 8%73/100
Finland's education ministry and Opetushallitus published national AI recommendations (31 March 2025) after consulting 172 organisations and youth 'Digiraati' councils, covering legal …
Helsinki, Finland
Good practice
Evidence: Descriptive / self-reported
Top 25%73/100
Since 2020, Estonia's Riigikogu has used HANS, an AI speech-recognition tool built with Tallinn University of Technology, to auto-transcribe plenary debates — …
Croatia's Ministry of Justice deployed ANON, an AI pseudonymisation tool integrated with the national eSpis system, to automatically redact personal data from …
Zagreb, Croatia
Good practice
Evidence: Observational / pre–post
Top 25%73/100
B’Odogwu (Nigeria’s Unified Customs Management System) applies AI risk-profiling and ML pattern detection across 34 commands. Between Oct 2024 and May 2025 …
Hungary requires all businesses to report invoices to NAV in real-time XML since 2021; ML risk-scoring targets suspicious companies before returns are …
Budapest, Hungary
Good practice
Evidence: Observational / pre–post
Top 25%73/100
Romania's ANI automatically cross-references procurement officials' identities with registered bidders before contracts are signed; over 100,000 procedures worth RON 72 billion screened, …
Bucharest, Romania
Good practice
Evidence: Observational / pre–post
Top 25%73/100
Malawi national programme deploying onecourse (onebillion) AI-adaptive tablets for Standards 1–4 literacy and numeracy across all 6,000 public primary schools by 2029. …
Lilongwe, Malawi
Good practice
Evidence: Descriptive / self-reported
Top 8%73/100
Kazakhstan approved joint ministerial AI-in-education standards (2025–2029) on ethics, data protection and academic integrity for 4.1 million pupils; AI entered school subjects …
Jamaica's 2022 pilot developed accessible digital storybooks tested in five schools—four for deaf, blind and intellectually disabled students. UNICEF's 10-country ADT programme; …
Kingston, Jamaica
Good practice
Evidence: Descriptive / self-reported
Top 8%73/100
From April 2022, Japan's MEXT made "Information I" mandatory for all ~3M high school students—covering programming, data analysis, networks and AI; the …
Tokyo, Japan
Good practice
Evidence: Observational / pre–post
Top 5%73/100
Scottish Government grant programme funding voluntary peatland restoration by landowners; 90,000 ha on path to recovery and 14,860 ha restored in 2024–25 …
Edinburgh, United Kingdom
Good practice
Evidence: Observational / pre–post
Top 5%73/100
National programme paying contractors to eliminate invasive wilding conifers from ~180,000 ha/year; wilding conifers remove up to 40% of catchment water. NZD …
Wellington, New Zealand
Good practice
Evidence: Observational / pre–post
Top 25%73/100
Kazakhstan's Digital Family Card uses predictive scoring across 100+ socio-economic indicators from 20+ agencies to proactively identify and serve vulnerable families. Since …
Astana, Kazakhstan
Good practice
Evidence: Descriptive / self-reported
Top 25%73/100
STF's Victor AI classifies Extraordinary Appeals by 'general repercussion' theme in ~5 seconds vs 44 minutes manually, handling ~50,000 cases/year. Built with …