The PJ anonymization tool for Italian legal documents was presented by Daniele Licari at the 16th International Conference on Statistical Analysis of Textual Data – Naples July 6-8, 2022 (JADT2022).
This tool facilitates the dissemination of judicial decisions which is not only a valuable source of decision support for judges and jurists, but also increases public trust in the judicial system. It allows people to personally verify their rights and the principle of equality. It also promotes data sharing and advanced analytics.
In our work, we have argued and demonstrated that the current manual approach to anonymization used by the Italian Justice System is inadequate both to preserve privacy and to handle the enormous amount of legal data.
In JADT2022 conference, we discussed and presented the first solution aimed at automatic anonymization of the Italian National Legal Archive (Archivio Giuridico Nazionale). The solution achieves very high performance in identifying entities to be anonymized (>99% for Person and ID Code entities) and supports several anonymization methods that can be applied to the text depending on the purpose of anonymization.
This work, carried out in collaboration with the Genoa’s Trial Court, can result in significant manual labor savings and minimize the time required to create a privacy-compliant (GDPR-compliant) legal data publication system.
- Licari D, Romano MF, Comandé G. Automatic Anonymization of Italian Legal Textual Documents using Deep Learning, 2022, JADT2022 proceedings