FAIR-PReSONS Interview by the University of Aegean

On 8 May 2026, an interview was published in a local online newspaper in Mytilene, touching on the topic of the FAIR-PReSONS project. The interview presented how our two-year European research initiative managed to develop fair, explainable, and gender-sensitive artificial intelligence tools for the criminal justice sector.

The main presenters were:

• Dr. Konstantinos Kotis, Associate Professor and Vice-Chair of the Department of Cultural Technology and Communication, University of the Aegean, and Scientific Coordinator of FAIR-PReSONS.
• Professor Giorgos Tsekouras, Dean of the School of Social Sciences, University of the Aegean.

The interview touched upon the project’s key findings, the gap observed in existing recidivism prediction systems, and the overall mission of the project to create an AI model that would address this disparity through the development of fairness-aware machine learning methodologies, with a particular focus on gender equity, model transparency, and explainability.

1. The FAIR-PReSONS System: Architecture and Methodology

Dr. Kotis presented the AI system developed within the project for the prediction of recidivism risk.

2. Gender Impact Analysis

Professor Tsekouras presented the gender-impact analysis conducted across the project’s participating countries. He addressed the broader debate on algorithmic fairness, emphasising that predictive models trained on historically biased institutional data risk entrenching and amplifying pre-existing social inequalities. He stressed that achieving fairness in AI-assisted criminal justice requires not merely technical intervention but sustained interdisciplinary collaboration between computer scientists, social scientists, legal scholars, and policymakers.

3. Open Datasets and European Data Infrastructure

The creation and public release of curated, anonymised datasets from three European countries was also discussed during the interview. These data are made available through European open-data platforms, intended to support further research in the field of AI-driven criminal risk assessment and provide a reproducible empirical foundation for future work on fairness and explainability in this domain.

To dive deeper into what was discussed, you can watch the full interview on YouTube (English subtitles available) or read the article published in the online newspaper StoNisi.