← Back to CiudadanIA summarises citizen letters to the Spanish President with small language models on local government infrastructure

Source proof for CiudadanIA summarises citizen letters to the Spanish President with small language models on local government infrastructure

Source-bound proof

Verified source excerpts for every supported field

Each colour maps a published value to the exact source passage used to support it. Only bounded excerpts are public; administrators can inspect the complete captured source.

12 fields supported

Title

CiudadanIA summarises citizen letters to the Spanish President with small language models on local government infrastructure

derived · high
…s that directly affects how quickly and effectively citizens receive responses. CiudadanIA automates the summarisation step using small language models. Rather than relying exclusively on large cloud-based AI systems, the project co…
…h fine-tuned compact models (specifically Gemma-3-4B, quantised for efficiency) to run on local government infrastructure. This design decision was driven by two requirements: citizen letters contain p…

Description

The Spanish Prime Minister's Office uses fine-tuned compact models (Gemma-3-4B, quantised) running on local government infrastructure to summarise citizen letters sent through the 'Write to the President' service, supporting routing and response. Government staff reviewed outputs for hallucination and tone distortion, and the project produced an open-source LLM-as-a-judge evaluation workbench. Humans stay in the loop; the source lists impacts qualitatively, without figures.

derived · high
…ng exclusively on large cloud-based AI systems, the project combined cloud with fine-tuned compact models (specifically Gemma-3-4B, quantised for efficiency) to run on local government infrastructure. This design decision was driven by two requirements: citizen letters contain p…
…ddressed to the Spanish President through the “Write to the President” service. Small language models generate summaries of citizen correspondence, enabling faster routing and response. The project also produced open-source tools for testing AI prompts and models a…
…tively refine prompts and model behaviour. To support this evaluation at scale, the programme developed an open-source LLM evaluation workbench that enables systematic testing of different prompts, models, and configuration…
…espondence, public consultations, or similar high-volume text processing tasks. The project keeps human-in-the-loop, maintaining oversight over all final decisions that affect how citizen letters…

Company

Spanish Prime Minister's Office

quote · high
CiudadanIA: AI-Powered Citizen Correspondence Processing for the Spanish Prime Minister’s Office - OECD.AI Going Digital DPP Blog Live data Policies and initiatives Priority is…

Country

Spain

classification · high
…ags AI-enabling Ecosystem Privacy Generative AI Trustworthy AI AI in Government About the policy initiative Spain Category: AI policy initiatives, programmes and projects Initiative type: AI use cases/p…

Industry

Government & Public Sector

classification · high
…ype: AI use cases/projects in the public sector Status: Active Start Year: 2025 Target Sectors: Social & welfare issues Inclusive development Public sector OECD AI Principles: 1.2 Respect for the rule of law, human rights and democrati…

Problem

The Spanish President's 'Write to the President' service receives thousands of citizen letters, each of which needs reading, summarising, categorising and routing to the appropriate government department, a labour-intensive process that affects how quickly citizens receive responses.

derived · high
…through the official “Escribe al Presidente” (Write to the President) service. This channel receives thousands of letters from citizens raising concerns, making requests, and sharing opinions on policy matters. Each letter requires reading, summarising, categorising, and routing to the app…
…zens raising concerns, making requests, and sharing opinions on policy matters. Each letter requires reading, summarising, categorising, and routing to the appropriate government department, a labour-intensive process that directly affects how quickly and effectively citizens receive responses. CiudadanIA automates the summarisation step using small language models. Rather…

Solution

CiudadanIA automates the summarisation step using small language models. The project combined cloud with fine-tuned compact models (Gemma-3-4B, quantised for efficiency) that run on local government infrastructure, because the letters contain personal and sensitive information and the government's AI sovereignty objectives call for reduced dependency on third-party systems. Government staff who process the correspondence gave feedback on outputs to refine prompts and model behaviour, and the project keeps a human in the loop.

derived · high
…government infrastructure. This design decision was driven by two requirements: citizen letters contain personal and sensitive information, and the government’s AI sovereignty objectives call for reduced dependency on third-party systems. The fine-tuning and evaluation process was rigorous. Government staff who proc…
…aracterising respectful requests as demands), and omission of relevant content. This feedback was used to iteratively refine prompts and model behaviour. To support this evaluation at scale, the programme developed an open-source LLM…
…espondence, public consultations, or similar high-volume text processing tasks. The project keeps human-in-the-loop, maintaining oversight over all final decisions that affect how citizen letters…

Business value

The source lists impacts qualitatively, without figures: faster processing of citizen correspondence, improved service quality, reduced staff workload on routine summarisation, and a reusable open-source evaluation methodology for other government AI deployments.

derived · high
…ning (SFT) together with diverse reinforcement learning for better performance. Impacts: faster processing of citizen correspondence, improved service quality, reduced staff workload on routine summarisation, and reusable evaluation methodology applicable to other government AI deployments. AI Tags AI-enabling Ecosystem Privacy Generative AI Trustworthy AI AI in Govern…

AI capabilities

Generative AI

classification · high
…e overview CiudadanIA is an AI initiative within the PresidencIA programme that applies generative AI to improve the processing of citizen correspondence addressed to the President of Spain through the official “Escribe al Presidente…

Technology

Gemma-3-4B

classification · high
…ud-based AI systems, the project combined cloud with fine-tuned compact models (specifically Gemma-3-4B, quantised for efficiency) to run on local government infrastructure. This design decision was driven by…

Headline outcome

derived · high
…ning (SFT) together with diverse reinforcement learning for better performance. Impacts: faster processing of citizen correspondence, improved service quality, reduced staff workload on routine summarisation, and reusable evaluation methodology applicable to other government AI deployme…

Use case type

Document processing

classification · high
…s that directly affects how quickly and effectively citizens receive responses. CiudadanIA automates the summarisation step using small language models. Rather than relying exclusively on large cloud-based AI systems, the project co…
Capture details
Captured
16 Oct 2026, 05:19 UTC
Extractor
fetch-strip@1
Snapshot hash
644bd151059ab2fdc1f1eee67f2736ccf5a0a5659ba0a893e73c0a00c739daa6