New publication: Artificial Intelligence, Humans and the Law
The purpose of LEXplain is to investigate how AI can be safely implemented in public administration to support legal decision-making in a way that does not compromise the explainability requirement. In a new book co-edited by Henrik Palmer Olsen entitled Artificial Intelligence, Humans and the Law several important questions about AI’s role and impact on the public administration of the law is analyzed.
Publisert:
About the book
In the introductory chapter, written by Henrik Palmer Olsen and two colleagues from the university of Copenhagen tackles the question of whether large language models can meaningfully participate in the generation of administrative legal decisions. Distinguishing between existential and functional understanding and inspired by the famous Turing test, the authors argue that large language models can function as a generator of text, which humans can perceive as legally meaningful and that it is therefore possible to include large language models in a legal decision work flow, but we also emphasize that a third mode of understanding, that we call situational is indispensable to legal judgment and that large language models, at least at present, do not exhibit such mode of understanding. A similar analysis conducted from a more general perspective in language philosophy, authored by Anders Søgaard and Thor Grünbaum enables further insight into large language models and their ability to model linguistic understanding. Among other LEXplain relevant contributions in the book is a chapter analyzing human engagement with non-minds (i.e. machines) and hybrid “others”, a chapter on the use of chatbots in public administration and a chapter by LEXplain co-PI Thomas Hildebrandt on Rule-based AI as Transparent, Accountable and Adaptable Computational Interpretations of Law. This latter chapter sets out to explain how rule-based AI can ensure stable and robust modelling of legal rules, which ensures explainability through the entire workflow of the administrative decision process.