Crime Research and Responsible AI

UZH's university funding program TRANSFORM provides seed funding to enable innovative ideas to be implemented and new organizational structures to be set up in cutting-edge research fields. The Executive Board of the University has now approved two new projects. It has allocated CHF 1.5 million over the next four years to support the CrimeLab^UZH project, which takes an interdisciplinary and transdisciplinary approach to crime and criminal justice research. It has also granted CHF 1.7 million to the Responsible AI (RAI) project, which conducts interdisciplinary research into the accountable use of AI.

RAI aims to advance responsible AI research, teaching and knowledge sharing at the University of Zurich. It also seeks to promote technology transfer and pool the interdisciplinary expertise that already exists in this area at UZH. "Responsible AI is not a clearly defined concept, but a broad, interdisciplinary field of research," says Reinhard Furrer, one of the project managers and also a professor at the Department of Mathematical Modeling and Machine Learning (DM3L).

The project focuses on transparent, fair and socially responsible AI systems, as well as on the methodological, ethical, legal and social questions that need to be addressed in relation to its use. UZH has deliberately decided not to join the race to train large language models, and is instead conducting basic research and critically reflecting on AI and its societal context and implications, notes Furrer. RAI is intended to become one of the main pillars of UZH.ai - the platform that pools all the initiatives on the topic of artificial intelligence at UZH.

Many fields already focus heavily on the topic of responsible artificial intelligence in both research and teaching - from informatics and economics, to social sciences, ethics and law. RAI aims to better pool and showcase this expertise, which is distributed across different disciplines at UZH. The project also seeks to establish a professorship at the Department of Informatics and one at DM3L to support early-career researchers, bolstering research at both departments. Reinhard Furrer firmly believes that, given the dominance of the big tech companies, universities have a significant role to play as independent bodies for critical and responsible AI research.

CrimeLab^UZH is designed to become a transdisciplinary hub for crime and criminal justice research. Both basic and applied research into issues relating to crime and criminal justice have long been researched at multiple UZH faculties, but also at external institutions, such as the Zurich Forensic Science Institute, the research department of Zurich's law enforcement, and the municipal and cantonal police. "It was often left to chance whether the different research groups were even aware of each other's existence and whether they would pool their resources in interdisciplinary projects," says professor of law Thierry Urwyler, member of the Board of Directors of CrimeLab^UZH.

The idea therefore came about to bundle strands of research and to systematically tap into synergies. "With its external research partners, the University of Zurich has enormous potential to study crime and criminal justice in an interdisciplinary and multi-method way," says Urwyler, "but for these types of connections between research groups and working partners to become reality, a shared platform is needed." This is precisely what the TRANSFORM financing has made possible.

One of the four research priorities of the CrimeLab^UZH looks at new forms of crime. The focus of this includes digitalization and the advancing possibilities offered by AI: "The question is already being asked, for example, if and how criminal law can continue to be applied when (fully) autonomous systems are used - such as in the context of road traffic," says Thierry Urwyler. What's more, artificial intelligence will also have an impact on areas of crime that we already know about, for example, large language models allow people with no relevant coding skills to develop malicious software, the professor adds.

AI also makes criminal activities more accessible and scalable, which is further fueled by the extensive digitalization of society, Urwyler explains. Phishing, ransomware attacks and numerous other offenses can therefore be committed by many people against many people, without the need for specific IT knowledge. "The science community needs to wake up to this emerging - or in some cases already existing - problem, and new strategies need to be developed regarding prevention and criminal law in order to keep pace with the momentum of technological development," says Thierry Urwyler.

One of the sub-goals of the CrimeLab^UZH is to consolidate the research data from various projects. If this is successful, researchers with specific thematic interests will be able to access appropriate data sets more quickly. A growing data pool creates optimal conditions to systematically analyze large data sets using machine learning. To support early-career researchers in this field, CrimeLab^UZH will also offer postgraduate continuing education programs and an interdisciplinary Master's program.

TRANSFORM funding line

Prof. Dr. Reinhard Furrer

Prof. Dr. iur. Thierry Urwyler