Why, and how, is an autonomous system different from an automated system? How does implementation of autonomous systems affect, and reflect, the Rule of Law? Can we identify the essential requirements that need to be reflected in the design-process of an autonomous system?
In order to find essential requirements which need to be reflected in the design-process it is necessary to adapt to the fundamental rules and regulations. Therefore ‘essential requirements’ refers to the translation of fundamental rules and regulations into the research design of autonomous systems. This is a difficult task since autonomous systems are not easy to define, and ‘essential requirements’ are not easy to find. Therefore I would like to discuss three examples of automated systems/autonomous systems, to try to clarify what is at stake: a candy machine, a smart home thermostat and the algorithmic system used by the Dutch Tax Authorities.
This presentation brings together philosophy, law and technology in an interdisciplinary setting, in order to consider potential answers to applied questions. Concerns regarding legality, liability, responsibility, accountability, proportionality and subsidiarity need to play a role during design-choices in AI and ML models, and we need first to situate the relevant questions at the level of applied cases, in order to investigate the possibility of Legal Protection by Design.
human autonomy, autonomous vehicles, essential requirements, autonomous systems, Rule of Law, Legal Protection by Design.