Ljupcho Grozdanovski gets a position of Research associate from the FNRS at ULiège

Ljupcho Grozdanovski, a researcher in law, specialized in Artificial Intelligence (AI),  becomes a F.R.S.-FNRS Research associate ( Belgian National Fund for Scientific Research), within the research unit Cité (Faculty of Law, Politicial Science and Criminology) of the University of Liège to pursue his researches that brings together his two main fields of interest: AI and the administration of justice and evidence.


Ljupcho Grozdanovski had the idea to explore, on the one hand, the field of AI and, on the other hand, the field of justice and evidence after publishing, in 2021, an article on Amazon's use of a worker recruitment algorithm. By associating the level of professional performance with the workers’ gender, this algorithm ultimately disregarded the Curricula Vitarum of female applicants, thus becoming the 'perpetrator' of a gender-based discrimination. Against this backdrop, his article addresses the following question: how could women-victims of 'algorithmic' discrimination establish, before a court, that they had suffered unequal treatment performed by an intelligent... and non-human system?

As far as the administration of evidence and justice is concerned,” the researcher explains, “there are going to be problems that we do not encounter when dealing with cases where human agency is beyond doubt. Indeed, quite often, the problem with these algorithms is that their decisions are opaque. So, how can one know that one was discriminated against, if one does not - or cannot - know how the algorithm made its decision?” There is a problem with the access to facts and therefore, to evidence, without which it is impossible to plead one's case and ask a court to award protective and/or compensatory measures.

Moreover, the victims of so-called algorithmic harm are not the only ones who experience procedural difficulties. One must also consider the programmers and deployers of AI, who tend to be designated as the by default culprits for the harmful actions of the systems they have created and put in the market. For them, the main question is: “How can they defend themselves against harms for which they were not directly responsible? In the law of evidence, be it national or at the level of the European Union, there is the age-old dogma that harmful action is invariably caused by a human agent, which implies that it is usually possible to identify a human perpetrator in the causal chain by adducing various types of evidence. But now we have non-human systems that display intelligence similar to that of humans. The age-old dogma of human agency (and thus rationality) is no longer tenable, as more and more disputes emerge dealing with harm caused by intelligent systems acting without apparent human intervention.”

Finally, through his study of the evidentiary difficulties raised by AI, Ljupcho Grozdanovski stresses a major risk: the emergence of a double standard of judicial protection which would include a standard of protection for parties to ‘ordinary’ disputes and another (lower) standard for parties to disputes involving intelligent systems. Our current laws of evidence are adapted to cases in which human agency is not questioned, but they are ill-adapted to cases in which human agency is not apparent. The danger of this double standard is that, in so-called algorithmic disputes, the parties are ultimately left without adequate judicial protection. A solution must therefore be found that would ensure a uniform level of judicial effectiveness, regardless of whether harm is caused by human agents or intelligent, non-human entities.

As for his interest for this theme: “It is a field that is brand new and where there is literally everything to build. I have indeed been able to identify a gap in the relevant scholarship; a gap which I hope to fill. It is an area where I can address issues that interest me anyway, namely, the theory of justice and the theory of evidence. Since AI is an emerging phenomenon, it allows for creativity and prospective research that could be of interest not only to academia, but to regulators and stakeholders as well.”

About Ljupcho Grozdanovski

Born in North Macedonia, Ljupcho Grozdanovski studied law at the University of Strasbourg. He then worked as a teaching and research assistant at the University of Geneva. In 2015, he defended his thesis on presumption in European Union law, at the University Aix-Marseille. Ljupcho Grozdanovski then completed a first postdoc in Switzerland (U.Neuchâtel/U.Geneva). In 2019, he arrived at the University of Liège to work on a project focused on AI and its impact on the labor market. After a fellowship at New York University (NYU) and after teaching European law at the University of Nantes, Ljupcho Grozdanovski permanently joined ULiège. In October 2022, he obtained a mandate as a research associate at the FNRS, the National Fund for Scientific Research.

La preseomption en droit de UE

* Grozdanovski, L. (2020, 25 février). La présomption en droit de l’Union européenne. ANTHEMIS.

An article written by Lola Barnabé, intern at the Communication Dpt of the University of Liège

Share this news