Subscribe to our Newsletter

Subscribe to our Newsletter

AI for Justice

Intelligent assistant to support judges in drafting sentences by locating texts of previous judgments and jurisprudence.

Centre de Telecomunicacions i Tecnologies de la Informació (CTTI)

El Centre de Telecomunicacions i Tecnologies de la Informació (CTTI) is the public body that integrates all the information technologies and telecommunications services of the Government of Catalonia. It is responsible for designing, building, coordinating, and deploying technological projects to provide solutions to all departments and various bodies of the Public Administration, being in charge of managing and evolving all the IT services of the Ministry of Justice of the Government of Catalonia according to their needs and demands.

The Secretariat for Justice Administration of the Ministry of Justice of the Government of Catalonia includes among its functions the modernization of the justice administration in Catalonia through the renovation of judicial infrastructures, information systems, and the organization of the judicial office to achieve a more open, agile, efficient and quality justice.


The time dedicated by judges to search for precedents and legal foundations to write a sentence is high and an impediment to being more agile in the drafting of sentences, causing a high response time and a sense of slow justice.

Currently, the location of paragraphs of interest of legal or procedural references, of any documentary information or the entirety of the magistrates’ own judgments or resolutions previously issued, which may be applicable or reusable in the drafting of new judgments or resolutions, becomes a manual and costly process that causes magistrates to spend a lot of time locating, remembering, and reviewing previous judgments, relying on their memory, both for their content and the date they were issued.

In Catalonia, the judge usually performs a manual search in the procedural management application for justice (, in the archive of the magistrate’s own judgments, or through the Judicial Documentation Center (CENDOJ), depending on the General Council of the Judiciary (CGPJ), as well as other repositories or jurisprudential databases. The search is more or less effective depending on what the magistrate remembers at that time from similar cases or parts of an old judgment that may be applicable to the current one.

The Challenge is to reduce the time dedicated in redacting sentences by locating reusable information within judicial documents, such as sentences issued in any instance and jurisdiction, in order to speed up and facilitate the drafting of new sentences by judges.

Currently, the location of paragraphs of interest (headings, legal foundations, dispositive part or judgment), legal or procedural references or the entirety of the judges’ previous sentences, that are applicable or reusable in the drafting of new sentences, becomes a manual and costly process that causes judges to spend a lot of time locating, remembering and reviewing previous judgments, based on memory, both of their content and of the date they were issued.

The solution we want to obtain is an assistant to judges and magistrates that, by entering an open text or prompt, can quickly, accurately, and relatedly obtain all applicable historical information for the sentence being drafted. In the same way, the assistant must provide the applicable legal foundations and sentence draft for a better management of mass litigation cases with hundreds of very similar claims that has also similar resolutions.


Applications closed on the 24th of January 2024 at 17:00 CET.



CT is a company specialized in data analytics and artificial intelligence with over 7 years of experience in the sector. It provides tailor-made solutions to enterprises across diverse industries and has participated in several EU-funded AI projects.

The solution is designed to automate and optimize several key aspects of a judge’s workflow, from the initial acquisition of legal documents to the final drafting of rulings. Here’s how the solution aligns with and addresses the challenge:

1. Streamlined Legal Document Processing: The proposed system begins by systematically collecting and refining legal documents from various judicial databases and judges’ private archives, preparing them for in-depth analysis. This directly addresses the challenge of time-consuming manual searches for precedents and legal foundations, thereby speeding up the drafting process.

2. Advanced Knowledge Integration and Analysis: By leveraging vectorization and Knowledge Graphs, the system provides a higher level of explainability for outputs and mitigates potential misinformation.

3. Personalized Recommendations and Contextual Understanding: The system employs clustering techniques and judge profiling to organize case law into relevant groups and tailor its output to individual judges’ historical preferences. This personalized approach ensures that the assistance provided is not only relevant but also respects each judge’s unique decision-making patterns.

4. Comprehensive Language Support: Recognizing the bilingual nature of the Catalan judicial system, the solution is designed to support both Catalan and Spanish, ensuring accessibility and usability for all judges in the region.

5. GDPR Compliance and Data Privacy: The solution is fully GDPR compliant, addressing the critical requirement for privacy and security in handling sensitive judicial information. Anonymizing algorithms and security measures are implemented to protect personal data.

6. Efficiency and Quality Improvement: By automating the search and analysis of legal documents and providing structured, personalized assistance, the solution promises a significant reduction in the time required to draft rulings. This not only enhances efficiency but also contributes to the uniformity and quality of sentences, ensuring fairness, equity, and consistency in the application of the law.