ARUM

ARUM explores how legal norms can be translated into structured, machine-executable representations and how this process can be automated using large language models.

Together with Rulemapping Group GmbH, which has developed the open-source “Rulemap” standard, the project builds on an existing approach to formally model legal rules and administrative processes. Traditionally, this work has been carried out manually by trained legal experts. With recent advances in AI, a new opportunity emerges. Large language models can potentially take over parts of this formalization process and turn legal text into executable structures. ARUM investigates how far this automation can go and how it can make the digital transformation of legal systems faster, more scalable, and more accessible.