Catalogue: AI in the Public Sector clear
2 practices not yet assessed for evidence strength.
Most practices in: France (17)
Since 2021 France's VIGINUM service has tracked foreign disinformation networks with AI tools; in 2025 it open-sourced D3lta, an LLM-based detector for …
France's tax authority scans aerial imagery with AI to find undeclared pools and extensions. In 2023 it flagged ~140,000 unreported pools, yielding …
France's public employment service rolled out ChatFT to nearly 40,000 advisors from November 2024, saving up to 3 hours/week each, while a …
France's data-protection authority ran a 2023-24 regulatory sandbox pairing four public bodies — including France Travail and RATP — with AI projects, …
France's 2019 Grand Débat National used a seven-firm consortium's NLP pipeline to classify and synthesise nearly 2 million online contributions and 16,000+ …
LLaMandement is DGFiP's fine-tuned Llama 2 model that classifies and summarises amendments to France's annual Finance Bill. Officials report it cuts amendment-processing …
The European Ombudsman is piloting the EU Commission's GPT@EC generative-AI tool for legal research and drafting support, explicitly barred from taking decisions …
Nantes Métropole became the first French metropolitan authority to publish the source code of algorithms it uses — water and transport low-income …
DINUM's Albert API (open-source Llama/Mistral, SecNumCloud) provides sovereign generative AI to French civil servants. A France Services pilot found 71% ease-of-use; 10,000 …
In February 2026 Etalab launched a public MCP server exposing data.gouv.fr's full catalogue — ~74,000 datasets/APIs, 390,000 files — to AI assistants …
France's open-data portal uses Albert, the state's open LLM, to auto-draft dataset descriptions and suggest metadata tags, and exposes its catalogue via …
Founded by three independent researchers after France's 2016 transparency law went largely unenforced, ODAP catalogues government algorithms. Its Nov 2025 inventory logged …
Researchers launched a citizen-led observatory in Nov 2024 to track which French agencies disclose their algorithms as a 2016 law requires — …
France's 2016 law requiring agencies to disclose algorithmic decision rules has gone almost entirely unenforced, the Ombudsman found in Nov 2024, naming …
France Travail built an AI tool, trained on 60,000 cases across 26 variables, to classify job seekers as 'suspect' or 'non-suspect' for …
In February 2026, France's data.gouv.fr began exposing its dataset catalogue via a Model Context Protocol (MCP) server, letting AI assistants such as …
Since 2010, France's CNAF has scored roughly 32 million people in benefit households from 0 to 1 for fraud-control targeting. Lighthouse Reports/Le …
Open full copilot Grounded in cited practices — always check the sources.
