BlueLab: preparing law students for responsible AI use

BlueLab is a practice-oriented training environment integrated into the Legal Analytics course for third-year students of the Bachelor of the European Law School. “Students step into the role of sustainability compliance officers and work on realistic cases from the blue economy, the part of the economy built around the sea and its resources,” Nanda begins to explain. “They are given a case study of a company dealing with maritime sustainability obligations, along with a package of EU directives, treaty provisions, and corporate sustainability reports.”  

Via the BlueLab dashboard, students analyse this material using AI tools, identify the relevant requirements, compile a register of obligations, and write a client-style memorandum with an action plan. The BlueLab dashboard is a new development that integrates document processing tools with large language models (LLMs) and a curated legal knowledge base. Given a compliance scenario, the system identifies relevant legal provisions and supports students in generating structured compliance outputs grounded in those provisions. 

Marcos: “Students therefore use AI not as a tool to answer their questions, but as a scaffolding instrument. At every step, the dashboard asks them to link their assertions to a source text. The goal is to let students learn the actual workflow of compliance work in an environment where they can experiment, make mistakes, and learn to verify what AI tells them instead of taking it at face value. This verification workflow forms the core of responsible AI use, which is one of our priorities within the project.” 

Nanda and Marcos observed a gap between legal education and the ESG compliance labour market, where AI has become indispensable. Students must therefore be aware of regulations regarding AI and learn how to apply AI responsibly in their future work. According to Nanda, this is important in the Netherlands because the country is economically heavily dependent on the sea. “There is currently a great deal of fragmented documentation. Therefore, AI support is needed to be able to do the work. On top of that, new treaties are being concluded, such as the Biodiversity Beyond National Jurisdiction Agreement (BBNJ Agreement), which give this an extra boost.” 

According to Marcos, students are currently studying environmental and sustainability law as theory and learning what the rules entail. “What is missing is the professional layer on top of that: how to manage the workflow of due diligence, risk analysis, and client reporting in practice when dealing with thousands of pages of regulations and disclosure documents. BlueLab was developed to bridge precisely that gap. How do you verify the AI-generated outputs against authoritative sources? How do you document reasoning and account for your conclusions? These are skills that students will develop while working with BlueLab.”

The researchers hope that the BlueLab project will not only impact student development and the ESG compliance labour market. They also hope to spark something closer to home: more collaboration and dialogue between the various faculties of Maastricht University. Marcos: “At the Maastricht level, we can also look at the integration of AI into and for education, to create AI-supported education systems that are not only responsible but also legally compliant.” Nanda adds: “The goal is also to foster dialogue between different faculties. This is a compliance project within the legal sector, but many domains face a similar problem. We will openly publish all teaching materials and project documentation so that other educators can reuse and adapt them. The project also brings together the Faculty of Law and the Department of Advanced Computing Sciences, with the Responsible Data Science master's as one of our dissemination targets. By sharing more, talking, and developing these kinds of interdisciplinary programmes, we can also promote collaboration between faculties. Shashank Chakravarthy, a software engineer from the Brightlands Institute for Smart Society (BISS) at UM, also supports the project's technical development and long-term maintenance of the dashboard.”