Research projects
Project managers: Axel Adrian, Stephanie Evert
Collaborators:
- Nathan Dykes
- Philipp Heinrich
- Michael Keuchen
- Thomas Proisl
- numerous student assistants
Start: April 2020
End: March 2022
Acronym: LeAK 1
Funding: Bavarian State Ministry of Justice (StMJ)
Research objective:
A demonstrator for automatic anonymization and semi-automatic pseudonymization was developed and evaluated. In particular, judgments of local courts in residential tenancy law and traffic law were processed.
Only very few court decisions are currently published in Germany, mainly because they have to be anonymized manually at great expense. However, a large number of legally relevant documents are indispensable for legal practice and legal science as well as for the development of digital services in business, politics and administration. Automated processes for anonymizing these texts are necessary so that large volumes of court decisions can be made available in a time- and resource-efficient manner in compliance with the applicable data protection regulations. The judgments made available in this way could, for example, be used as training data for the development of processes based on artificial intelligence (AI), which will make the work of courts and authorities easier in the future.
The use of legal field-specific software solutions for the automatic anonymization of court judgments is a novelty in the field of e-justice and legal tech. Pioneering work is also being done in the project by systematically combining results from different areas of law for the first time and comprehensively anonymizing and pseudonymizing court judgments. This will make urgently needed data available on a large scale for the digital transformation of processes in state institutions. In addition, not only the private sector legal tech community and existing start-ups will benefit from the large database of anonymized and pseudonymized judgments, but also the digitalization of the state and administration as well as the innovative strength of Germany as a business location as a whole.
Further links and literature:
https://www.linguistik.phil.fau.de/projects/leak-anger/
Project managers: Axel Adrian, Stephanie Evert
Collaborators:
- Philipp Heinrich
- Michael Keuchen
- Numerous student assistants
Start: April 2023
End: March 2024
Acronym: LeAK 2
Funding: Bavarian State Ministry of Justice (StMJ)
Research objective:
An improved demonstrator for automatic anonymization and semi-automatic pseudonymization was developed and evaluated. In particular, OLG judgments in traffic accident and traffic insurance cases, as well as banking and insolvency cases, capital investment cases, cost cases and arbitration cases were examined. Furthermore, deanonymization experiments and usability studies for the self-developed user interface were carried out.
In Germany, only very few court decisions are currently published, mainly because they have to be anonymized manually at great expense. However, a large number of legally relevant documents are indispensable for legal practice and legal science as well as for the development of digital services in business, politics and administration. Automated processes for anonymizing these texts are necessary so that large volumes of court decisions can be made available in a time- and resource-efficient manner in compliance with the applicable data protection regulations. The judgments made available in this way could, for example, be used as training data for the development of procedures based on artificial intelligence (AI), which will make the work of courts and authorities easier in the future.
The use of legal field-specific software solutions for the automatic anonymization of court judgments is a novelty in the field of e-justice and legal tech. Pioneering work is also being done in the project by systematically combining results from different areas of law for the first time and comprehensively anonymizing and pseudonymizing court judgments. This will make urgently needed data available on a large scale for the digital transformation of processes in state institutions. In addition, not only the private sector legal tech community and existing start-ups will benefit from the large database of anonymized and pseudonymized judgments, but also the digitalization of the state and administration as well as the innovative strength of Germany as a business location as a whole.
Further links and literature:
https://www.linguistik.phil.fau.de/projects/leak-anger/
Project managers: Axel Adrian, Stephanie Evert
Collaborators:
- Bao Minh Doan Dang
- Philipp Heinrich
- Michael Keuchen
- Melanie Rosa
- Julian Werner
- Leonardo Zilio
- Numerous student assistants
Start: December 2022
End: December 2025
Acronym: AnGer
Funding: Federal Ministry of Education and Research (BMBF), Bavarian State Ministry of Justice (StMJ)
AnGer is part of the research cluster „Anonymization for the secure handling of data“ (https://www.forschung-it-sicherheit-kommunikationssysteme.de/foerderung/bekanntmachungen/anonymisierung)
Research objective:
The technical possibilities for automatic anonymization and pseudonymization are being researched on a very broad scale and a demonstrator for automatic anonymization and pseudonymization is being developed, which can be used to process judgments from a wide range of instances and different areas of civil law. In addition, various other experiments are being carried out, in particular deanonymization experiments.
Only very few court decisions are currently published in Germany, mainly because they have to be anonymized manually at great expense. However, a large number of legally relevant documents are indispensable for legal practice and legal science as well as for the development of digital services in business, politics and administration. Automated processes for anonymizing these texts are necessary so that large volumes of court decisions can be made available in a time- and resource-efficient manner in compliance with the applicable data protection regulations. The judgments made available in this way could, for example, be used as training data for the development of processes based on artificial intelligence (AI), which will make the work of courts and authorities easier in the future.
The use of legal field-specific software solutions for the automatic anonymization of court judgments is a novelty in the field of e-justice and legal tech. Pioneering work is also being done in the project by systematically combining results from different areas of law for the first time and comprehensively anonymizing and pseudonymizing court judgments. This will make urgently needed data available on a large scale for the digital transformation of processes in state institutions. In addition, not only the private sector legal tech community and existing start-ups will benefit from the large database of anonymized and pseudonymized judgments, but also the digitalization of the state and administration as well as the innovative strength of Germany as a business location as a whole.
Further links and literature:
https://www.linguistik.phil.fau.de/projects/leak-anger/
https://www.forschung-it-sicherheit-kommunikationssysteme.de/projekte/anger
Project managers: Axel Adrian, Stephanie Evert, Michael Kohlhase, Lutz Schröder, Andreas Maier
Employees:
- Aurelius Adrian
- Anil Basaran
- Nathan Dykes
- Michael Gritz
- Merlin Humml
- Max Rapp
- Verena Stürmer
- student assistants
Start: March 2024
End: February 2027
Acronym: DIREGA
Funding: Federal Chamber of Notaries (BNotK), Bavarian State Chamber of Notaries (LNotK), Bavarian State Ministry of Justice (StMJ)
Research objective:
This is a multidisciplinary research project „Digital Register Assistant“ on behalf of the Federal Chamber of Notaries and in cooperation with the Bavarian Chamber of Notaries and the Bavarian State Ministry of Justice.
The project is concerned with basic research in the field of law and computer science and, in particular, questions relating to the extent to which legal decisions can be supported by machines or even made automatically using artificial intelligence (AI). Specific legal issues from commercial, corporate and register law are to be examined in connection with various sub-disciplines of AI research.
With the help of so-called corpus and computational linguistics (NLP), for example, natural language texts in legal documents are transformed into machine-readable language and information is automatically extracted. With so-called symbolic AI, e.g. legal argumentation must be formalized, legal expertise and practical knowledge must be represented automatically and machine reasoning processes must be evaluated. Pattern recognition and sub-symbolic AI methods can be used to generate legally relevant training data, for example. Finally, results generated by machines must always be evaluated for legal consistency. By developing a prototype as part of the research project, fundamental insights are to be gained into how law can be interpreted and applied in the future using technology or hybrid AI systems.
Project managers: Aurelius Adrian, Axel Adrian, Timo Meinhof
Start: April 2019
End: open
Funding: own funds
Research objective:
Tools for legal service processes will be developed, such as contract drafting automaton, money laundering documentation tool, semantic full-text search assistant, etc.