Consorci AOC's Algorithmic Transparency Reports — Catalonia Publishes Plain-Language Sheets for Its Public-Sector AI Services

Consorci AOC, the shared digital-services body of Catalan public administrations, began an algorithmic transparency initiative in 2023 after a study of regulations, guidelines and international initiatives. Its model puts a transparency report on each service's homepage and on the entity's transparency portal.
Four reports have been published: conversational chatbots (26 May 2023), video identification for idCAT Mòbil (2 February 2024), automation of energy-poverty reports (2 May 2024, a deterministic algorithm rather than AI) and generative-AI chatbots (26 November 2024).
Each sheet explains the problem addressed, how the system works, who is affected, risks (discrimination bias, security, availability), the fundamental-rights safeguards in place, appeal routes and a responsible contact and supplier. The methodology draws on the NIST AI Risk Management Framework and the EU Charter of Fundamental Rights, with six rights families (equality, data protection, security, transparency, accountability, social impact).
The Diputació de Barcelona's Local.ia portal highlights the approach as a model for local authorities. Evidence is limited to the published reports: no uptake or user-comprehension data is published, and coverage is four systems, so it demonstrates disclosure practice rather than measured impact.

Read the full analysis: https://www.aoc.cat/en/projecte-innovacio/transparencia-en-lus-dalgorismes-dintelligencia-artificial-a-laoc/