Artificial Intelligence and Public Administration : A Journey

This book is about public administration’s decision-making (and policy development activity) in the context of an increasingly ICT-driven environment. Already today, AI systems can deliver more precise results than humans in the context of data analysis. Public administrations can certainly use AI-based algorithms successfully to analyse past decisions and acquire a greater understanding of how laws operate in practice, thereby making more informed decisions. But when it comes to inventing the future (public policy development) and imagining the future (single-case decision-making), Artificial Intelligence lacks the imaginative component, which is (at least so far) solely a prerogative of the human brain and a consequence of its capacity to infer from situations and events that apparently have no connection at all. Overall, even technical competence is insufficient to properly decide on individual cases: the decision of an individual case, or single-case decision-making, necessitates adapting a general and abstract rule to a concrete situation. The remaining margin of uncertainty, which is an intrinsic human prerogative, must be reserved for human public officers. Moreover, despite the diversity of ADM technology and the rapid pace of technological development, basic legal requirements for the legality of public decision-making must remain unchanged and fixed in principles of the rule of law and good administration, implying that even if the AI revolution comes, public law does not go away! If left to operate without regulation, AI may indeed pose a tangible and significant threat to human rights, democracy, and the rule of law. The notion that we should allow the private sector and innovative businesses to operate freely, without restrictions (not “clip their wings”), is fundamentally wrong: in doing so, we would emulate the renowned Angelus Novus in Paul Klee’s painting, which I intentionally selected for the cover of this book.