Proteverb – Legal, ethical and technological aspects of processing textual and speech data for scientific, research and development purposes

ABOUT THE PROJECT

From the perspective of language technology development, Slovenian is a language for which few digital resources are available. As a result, modern research from a computer science perspective and the development of products based on natural language processing are significantly slower than for languages with many digital resources. However, in order to adequately acquire language resources and to make secondary use of them in as natural a form as possible, which may also contain some personal data, it is important to interpret the General Data Protection Regulation (GDPR) and the exceptions specifically for research purposes. It is precisely such exemptions that make it possible to achieve the specific purposes to which the present targeted research project is also linked, which will, for the first time in Slovenia, systematically address the acquisition and processing of (personal) data in a way that is in the interest of science, and thus, through application in the form of a pilot project, contribute to the development of the latter, as well as of the economy, on the basis of new insights and practices.

The research project will bring together, intertwine and deepen the knowledge of several different scientific disciplines in the field of social sciences, natural sciences, technical sciences and humanities. Such synergies are essential to ensure that advances in technological development are understood in the appropriate context and regulated in a way that maximises societal benefits and simultaneously minimises negative impacts and interference with ethical and legal standards and human rights. Such a comprehensive approach is the only way to develop the concept of open science to the full extent.

The project will primarily make a significant contribution to the development of three scientific fields, namely law, informatics and computer science, and humanities.

In all three fields of science, the research team will immediately transfer the findings and increased knowledge into the undergraduate and postgraduate studies at national and foreign universities, both through the participation of students in the development of the above-mentioned technologies and in the teaching carried out by the researchers involved in this project, who are also working as professors at different universities.

The academic results of this project will overcome key obstacles to the advancement of science and will optimise the use of data for research purposes without violating legal standards and human rights.

Project type: Target research Programme

Project code: V5-2265

Duration: 1. 10. 2022 – 31. 3. 2025

WORK PACKAGES

The targeted research project will be divided into several phases:

We will examine the legal framework of data processing for research and scientific purposes. The starting point will be the General Regulation and the ZVOP-1, which will be built upon through a comparative legal analysis and monitoring of the development of the ZVOP-2 legislative proposal.
We will look at the current practices of data collection for scientific research purposes, with an interest both in the access to data by researchers and research organisations and in the experience of data sharing by public authorities and institutions (e.g. courts). We will identify the key risk factors that have prevented access to data in the past, in order to develop a protocol to protect privacy in the course of data processing for scientific research purposes.
The project will develop procedures for appropriate data access and anonymisation, based on the adaptation and improvement of existing anonymisers. Recommendations will be made on methods for biometric anonymisation of audio speech recordings based on machine learning methods, with the aim of reducing the impact on the reliability of automatic speech recognisers.
We will attempt to acquire the data (pilot) using a privacy protocol and data access procedures, including anonymisation. The pilot part of the research will consist of the preparation of the necessary groundwork for data acquisition, data takeover, data anonymisation, and the organisation of documentation, procedures and rules for the needs of data processing within the research institution. On the basis of the data obtained for the pilot part of the targeted research project, we will specialise a text anonymiser as well as a speech recogniser for the Slovene language.

PROJECT PARTNERS

Project leader is the Institute of Criminology at the Faculty of Law Ljubljana.

Consortium Partners:

FINANCING

The project is funded by the Slovenian Research and Innovation Agency (ARIS) and the Ministry of Digital Transformation (MDP).

PROJECT RESULTS

Original scientific article

Završnik, A., Ramuš Cvetkovič, I., Lazarevič Padar, K., & Stariha, A. (2024). Regulating data access for criminological research purposes. Revija za kriminalistiko in kriminologijo, 75(1), 72–89. COBISS.SI-ID 192594691.
Završnik, A. (2023). Criminal justice, artificial intelligence systems, and human rights. Ûridičeskie nauki i obrazovanie, 70, 150–164. COBISS.SI-ID 159764739.
Bajec, M., Lebar Bajec, I., Šoltes, T., Cvek, J., Čibej, J., Gantar, K., Sever, S., & Krek, S. (2023). Online Notes – sistem za prepoznavanje govora in strojno prevajanje v realnem času za slovenska univerzitetna predavanja.Informacijska družba, 7–10.
Završnik, A. (2023). Artificial intelligence in the criminal justice system: Impact on human rights. Pravo i pravda, 21(1), 173–192. COBISS.SI-ID 155372291.

Other:

Krek, S. (2022, 29. november). Copyrights, the grave of the Slovenian language. Dnevnik, 72(276), 17. COBISS.SI-ID 184173827.

Unpublished conference contribution

Završnik, A. (2022, 27. november). Algoritmična nepravičnost. Predavanje na Jesenski šoli “Pravo pred izzivi digitale (r)evolucije”, Ljubljana, Slovenija.
Šarf, P. (2022, 15. november). The thin line between personal and anonymized data in the digital age. Predavanje na Conference on Data Protection Law, Portorož, Slovenija. COBISS.SI-ID 150492163.
Završnik, A. (2023, 16. marec). Crime and data. Predavanje na 4th Conference on Information Security Law, Portorož, Slovenija. COBISS.SI-ID 159392003.
Završnik, A. (2023, 5. maj). Artificial intelligence and criminal justice. Predavanje na International Scientific and Practical Conference Digital Forensics in the Modern World: Problems of Theory and Practice, Taškent, Uzbekistan. COBISS.SI-ID 159767299.
Završnik, A. (2023, 18. maj). Artificial intelligence and criminal justice. Predavanje na International Seminar Digitalization in Law, Privacy Protection, and Automation (DPZPA), Sarajevo, Bosna in Hercegovina. COBISS.SI-ID 159770371.
Završnik, A. (2023, 8. november). Umetna inteligenca in družbena škoda : predavanje, Ustavno sodišče RS, Ljubljana.
Završnik, A. (2024, 25. januar). Uspešnosti novega ZVOP-2 v praksi : okrogla miza, Dogodek Informacijskega pooblaščenca ob Evropskem dnevu varstva osebnih podatkov, Ljubljana, 25. 1. 2024.

Published scientific conference contribution abstract

Završnik, A. (2023). Fair trial implications of automation in criminal justice systems. V 2023 ASC Annual Meeting: Seeking Justice: Reconciling with Our Past, Reimagining the Future (Philadelphia, 15.–18. november 2023). American Society of Criminology. COBISS.SI-ID 183045891.
Završnik, A. (2024). Criminal justice data and GDPR: how to reconcile personal data protection with scientific research purposes. EUROCRIM 2024.

Published scientific conference contribution

Bajec, M., Lebar Bajec, I., Šoltes, T., Cvek, J., Čibej, J., Gantar, K., Sever, S., & Krek, S. (2023). Online Notes – sistem za prepoznavanje govora in strojno prevajanje v realnem času za slovenska univerzitetna predavanja. V Informacijska družba 2023 (str. 7–10). Institut Jožef Stefan. COBISS.SI-ID 172807683

Šoltes, T., Bajec, M., Lebar Bajec, I., Gantar, K., & Žitnik, S. (2023). Sistem spletnih opomb. V Lecture Notes in Business Information Processing (str. 485–492). Springer. COBISS.SI-ID 157601539

Organization of professional or scientific events

Završnik, A., Bajec, M., Geršak, K., Rozman, D., Romšak, A., & Šarf, P. (2023, 16. marec). Iskanje ravnovesja med znanstvenim napredkom in varstvom osebnih podatkov [Posebni panel]. Konferenca Pravo informacijske varnosti, Portorož, Slovenija.

Lectures at foreign universities

Završnik, A. (2023, 24. januar). AI and the penal system. Predavanje pri predmetu Artificial Intelligence and Intellectual Property, Centre for International Intellectual Property Studies, Université de Strasbourg, Zoom. COBISS.SI-ID 142227715.
Završnik, A. (2023, 12. september). AI and the penal system. Predavanje na Università Cattolica del Sacro Cuore, Milano, Italija (Zoom). COBISS.SI-ID 164318467.
Završnik, A. (2023, 8. december). Artificial intelligence in the judiciary: Opportunities and risks of algorithmic governance. Predavanje na Faculty of Organizational Sciences, University of Belgrade, in Research and Development Institute for Artificial Intelligence of Serbia, Novi Sad, Srbija. COBISS.SI-ID 178939651.

Professional Lectures

Završnik, A. (2023, 2. junij). Saznanja o mogućim pravcima primjene umjetne inteligencije u krivično pravosuđe. Predavanje na 21. Savjetovanju iz krivičnopravne oblasti: Izazovi pravosuđa u procesuiranju novih i tradicionalnih oblika kriminaliteta, Neum, Bosna in Hercegovina. COBISS.SI-ID 184535811
Završnik, A. (2023, 8. november). Umetna inteligenca in družbena škoda. Predavanje na Ustavnem sodišču Republike Slovenije, Ljubljana, Slovenija. COBISS.SI-ID 183957251
Završnik, A. (2022, 1. junij). The challenges of artificial intelligence in digital societies: The case of algorithmic criminal justice. Vabljeno predavanje na International Digital Security Forum (IDSF), Dunaj, Avstrija. COBISS.SI-ID 116816387
Završnik, A. (2022). Avtomatizacija kazenskega pravosodja. Digital Age Research Center (D!ARC), Universität Klagenfurt, 24. 11. 2022.

Radiotelevizija Slovenija. (2023). The summer of artificial intelligence [Radijska oddaja, Vroči mikrofon]. Z gosti: S. Splichal, I. Bratko, L. Kronegger, J. Kaluža, S. Krek, P. Šarf in V. Gorjanc. COBISS.SI-ID 154859523.
Radiotelevizija Slovenija. (2023). The datafication of society [Radijska oddaja, Aktualna tema]. Z gosti: S. Splichal, I. Bratko, S. Krek, J. Kaluža in P. Šarf. COBISS.SI-ID 154855171.
Radiotelevizija Slovenija. (2023). Who will benefit and who will suffer from the new artificial intelligence? [Radijska oddaja, Intelekta]. Voditeljica: N. Slaček. Gostje: M. Bogataj Jančič, N. Cvar, F. Dobranić in A. Završnik. COBISS.SI-ID 155695363.
KREK, Simon, Avtorske pravice, slovenskega jezika grob. Delo, 29.11.2022, str. 17.