Training eines Sprachmodells in der Justiz

Language models offer great potential for generative use in the justice system, for example in the preparation of procedural facts and the drafting of court decisions. However, training a specific AI model for the judiciary based on German court files and decisions involves the processing of personal data within the meaning of the GDPR and the use of material protected under the German UrhG.
This work addresses these challenges, comprehensively outlines the legal framework for training such a language model for use in the justice system, and provides answers to pressing legal questions.