The goal of this Research Chair is to propose a legal framework rendering AI systems truly accountable. Important legal issues—e.g. consumer protection, liability, and insurance—need to be resolved before AI gains full social trust and becomes widespread. French administrative law also includes a provision that administrative decisions must be explicable. However, this notion needs to be specified...The goal of this Research Chair is to propose a legal framework rendering AI systems truly accountable. Important legal issues—e.g. consumer protection, liability, and insurance—need to be resolved before AI gains full social trust and becomes widespread. French administrative law also includes a provision that administrative decisions must be explicable. However, this notion needs to be specified and adapted to the emerging use of algorithm-assisted decision-making in the public sector.