Catalogue: AI in the Public Sector clear
ECMWF made its AIFS machine-learning weather model operational in February 2025, with a probabilistic version following in July — cutting compute needs …
Australia's Digital Transformation Agency piloted a draft AI impact-assessment framework with 21 volunteer federal agencies, testing whether it caught risks existing processes …
Since April 2019, Canada's Directive on Automated Decision-Making has required every federal automated decision system to complete a public Algorithmic Impact Assessment …
Since 2023, GovTech has scaled Pair, an in-house generative-AI assistant, from 60,000+ registered users exchanging 10 million+ messages in mid-2025 to over …
The UK's Algorithmic Transparency Recording Standard requires central government bodies to publish structured records of algorithmic tools that significantly influence decisions or …
Cambodia's MPTC-run Verify.gov.kh authenticates driving licences, degrees and IDs via QR codes, blockchain and AI. By 2026 it covered 16+ ministries and …
Since 2021, Universidad Adolfo Ibáñez's GobLab has independently catalogued automated decision systems used by the Chilean state. Its 2026 report documents 162 …
Colombia's telecom regulator CRC ran a full regulatory-sandbox funnel — 8 proposals from 6 companies, 5 rejected, 2 advanced — and authorised …
NASK's Disinformation Analysis Centre ran 'Parasol Wyborczy' during Poland's 2025 presidential election, blocking 96% of accounts impersonating candidates, reporting over 16,000 inauthentic …
Peru's Ministry of Environment has run Geobosques, a weekly satellite-based deforestation early-warning system, since April 2017; a PNAS randomized controlled trial found …
Launched in July 2020, the Bangko Sentral ng Pilipinas' chatbot BOB has grown from handling 15% of complaint cases in its first …
Japan's National Diet Library open-sourced an AI OCR engine that lifted recognition accuracy on pre-1945 texts from about 40% to over 90%, …
Swindon Borough Council piloted Beam's Magic Notes AI transcription tool with 19 social workers across 184 meetings in 2024, cutting Care Act …
A peer-reviewed pilot across 10 sites in Lahore, Pakistan, combined ultrasonic bin-fill sensors with cloud-based AI route optimisation, processing over 200 million …
Woogle, built at the University of Amsterdam, uses OCR and document-classification models to index over 3 million Dutch FOIA dossiers (11M+ pages) …
Malta committed €3m to deploy SAS AI tools that cross-reference registries and bank data to flag VAT fraud; officials attribute most of …
Hong Kong's cross-department 1823 hotline now uses AI speech recognition, generative-AI drafting and a chatbot resolving over 90% of its 10,000+ monthly …
Since late 2022 the ECB has run a quantile-regression-forest machine-learning model in its monetary-policy toolkit. In Q2 and Q4 2025 it flagged …
Canada's signals-intelligence agency runs Assemblyline, an open-sourced ML malware triage platform, plus a separate classifier flagging threats commodity antivirus misses across ~900,000 …
Chicago's health department built a machine-learning model ranking restaurants by violation risk. An independent pilot evaluation found inspectors using its list caught …
The US State Department's enterprise AI chatbot StateChat grew from 150 testers to 62,000+ users by June 2026; a 2025 survey of …
Italy's ANAC runs a public red-flag portal using ~70 corruption-risk indicators across 60m+ procurement records since 2022, flagging roughly one case a …
Covered California deployed Google Public Sector's Document AI with Deloitte to automate income/eligibility verification across 25 document types — cutting confirmation times …
WIPO's free Global Brand Database uses deep neural networks to match trademark images by concept — apple, eagle, star — rather than …
Open full copilot Grounded in cited practices — always check the sources.
