An open source AI tool to help people and organisations interpret and apply EU legislation more efficiently.
EurLexGPT is an open source AI tool being developed in the OIVA project to help organisations interpret and apply EU legislation more efficiently. It is a specialised language model trained on EU legal texts and designed to provide reliable, up-to-date information about European regulations and compliance requirements.
In the framework of the OIVA project, EurLexGPT is being developed through a process of first transforming EU legal documents into a machine-readable, vector form, which can then be used by AI to help businesses and organizations make sense of EU law. The technology is designed to serve various use cases and applications, which will be demonstrated in OIVA’s pilot areas.
EurLexGPT builds on insights from earlier Sitra-funded pilots and research.
If EU legislation were a vast library, EurLexGPT would be the librarian: The laws, regulations, guidance documents and court decisions already exist, just like books on library shelves. But finding the right information, understanding how different sources connect, and getting reliable answers can be difficult and time-consuming. EurLexGPT helps people navigate that library. It can find relevant legal information, explain connections between rules, and point users to the sources they need.
European legislation is complex and constantly evolving. Businesses, public authorities, auditors, and technology developers often face challenges in understanding how legal requirements apply in practice. Regulatory compliance can also be time-consuming, costly, and difficult to manage, especially when legislation changes frequently. At the same time, European businesses are struggling with declining competitiveness in the global markets, a fact closely tied to their regulatory burden, as identified by the European Commission.
EurLexGPT aims to bridge this gap by making legal information easier to access, interpret, and use. It supports the transformation of regulatory texts into formats that can be understood not only by people but also by digital systems and automated services. It also contributes to the Commission target to support competitiviness by reducing the administrative burden on European businesses in its 2024-2029 work program: instead of spending resources on regulation compliance, companies will have more time to innovate and grow.
Built on European values
A key objective of EurLexGPT is to support a European approach to trustworthy AI. EurLexGPT is designed in accordance with EU standards, including strong privacy, security, and ethical safeguards, demonstrating strong technological sovereignty in the form of a European model of AI that citizens and organisations can trust in.
The vision for EurLexGPT is to offer a common digital infrastructure for AI-enabled application of EU regulations in Europe, for Europe.
Open, public good for Europe
EurLexGPT is being developed as an open resource for the European ecosystem. The project’s vision is that organisations should not need to separately build and maintain their own versions of EU legal knowledge. Or be reliant on digital giants or expensive licenses. Instead, EurLexGPT can serve as shared digital infrastructure that supports innovation, compliance, and public value across Europe.
The software, documentation, and AI models developed in the project will be released openly, enabling others to reuse, improve, and build upon the results, tailored to their specific needs. As such, EurLexGPT offers an input source for anyone wishing to develop AI applications as an integrated part of their services. The vision is for EurLexGPT to enable an explosive growth in regulatory technology innovation.
1. A comprehensive EU legal database
The system uses a continuously updated database containing core EU legal documents, including treaties, legislation, and case law from official EU sources.
2. A specialised AI large language model
An open-source large language model is further trained using EU legal material so that it can better understand legal terminology, concepts, and regulatory contexts.
3. Intelligent retrieval of legal information
EurLexGPT uses Retrieval-Augmented Generation (RAG) technology, which allows it to search the legal database and use relevant source material when generating responses. This helps improve accuracy and reliability.
4. Open interfaces for innovation
The system will provide open APIs that allow pilot projects, developers, and regulatory technology (RegTech) companies to build their own services on top of EurLexGPT.
Funded by the European Union. Views and opinions expressed are however those of the author(s) only and do not necessarily reflect those of the European Commission. Neither the European Union nor the granting authority can be held responsible for them.