In early 2019, UNHCR’s Division of International Protection embarked on a project to use machine learning to automate manual annotations of legal documents in Refworld. Refworld is UNHCR’s public database of case laws and other legal documents — a vital resource for legal aid workers, lawyers, judges, and UNHCR staff who rely on it for their legal research. These annotations enhance navigation and searchability between documents to assist in building legal arguments for asylum-seekers and people of concern.
With a $50K grant from Splunk Inc. in 2022, The Hive, the innovation lab at USA for UNHCR, conducted a feasibility study for automating the extraction of case law citations from legal documents. Machine learning was used to identify case law entities within the documents to improve the website’s user navigation experience.