In this workshop, Anca Radu and Yeliz Figen Döker will present their papers entitled ‘Using AI systems to predict the behaviour of parties in criminal proceedings’ and ‘Learning through examples: teaching ethics and law to AI’. Law PhD researchers, Natalia Menéndez González and Marco Lasmar Almada will take part in the presentation as discussants.
'Using AI systems to predict the behaviour of parties in criminal proceedings'
Emerging technologies that fall within the umbrella term of Artificial Intelligence – statistical systems included – have an incredible transformative potential, which raised primary concerns, notably when designing, developing, and deploying these applications in the judicial systems. Within the array of AI systems used in the judiciary, one stood out due to its contribution to impactful decisions based on predicting the probability of something an individual has not done yet. Indeed, we are talking about recidivism risk and needs assessments.
This paper provides a descriptive examination of existing systems of recidivism prediction and maps out the critiques brought forth by academia. More precisely, it pins critical issues throughout such systems' lifecycle and the course of criminal proceedings. However, the human rights dimension of these issues is of greater interest. Here are two angles to capture it: can these applications be adopted across justice systems in a human rights compliant manner, and how does such adoption impact the judiciary's role, of which the parties' enjoyment of their rights is essential.
'Learning through examples: teaching ethics to law to AI'
The necessity for regulation of AI systems has become more crucial in recent years due to systems' elusive outcomes and unpredictable capacities. AI systems should not be equated with other technological tools due to their intricate nature and preposterous potential — at least for the AI systems that are trying to be achieved. For this reason, this research project explores an alternative regulation model where Experiential Learning is implemented. Learning from experiences is not an unprecedented concept both in technology and law. By referring to 'experience' in law and technology, the research seeks whether integrating the Experiential Learning model into the existing learning cycle of AI to learn law and ethics can create an alternative regulation model.
The second-year paper begins with a short history of artificial life, artificial intelligence, and the term AI. The following section discusses the grounds for regulation and the existing effects of AI systems. Then, it establishes a solid basis for the alternative regulation model by discussing the ethics of AI. In order to assist the reader in comprehending the Experiential Learning approach taken in the research, the author highlights legal theorists who put a high value on the importance of experience.
Speaker Bios
Anca Radu is a legal scholar at the European University Institute, where she works under the supervision of Professor Giovanni Sartor. She examines legal and ethical questions around the design, development, and deployment of AI systems in the judiciary. Her research focuses on the impact of these applications on human rights, democracy, and the rule of law. Besides her academic activities, Anca regularly collaborates with international organizations on AI policy and regulation.
Yeliz Figen Döker is a researcher at the Law Department of the European University Institute, holding an LL.M from Bournemouth University (BU) and a Law Degree from Bahçesehir University (BAU). Yeliz is currently working on the matters of Artificial Intelligence Systems and Ethics under the supervision of Professor Nicolas Petit and Giovanni Sartor. Her research interests also cover Artificial Moral Agents, AI Ethics, Machine Ethics, Ethics by Design and Experiential Learning. She is also co-founder of The Digital Constitutionalist (DigiCon) and a member of the Istanbul Bar Association.