A TECHNOPROGRESSIST READING OF THE DECLARATION OF MONTREAL ON A.I.: WHAT ARE THE STAKES FOR THE ETHICS OF TOMORROW

Abstract

On November 28, 2018, at the initiative of the University of Montreal, one of the very first normative texts of a national level has been published, following the Asilomar international conference (2017) on AI. Entitled Montreal Declaration for a Responsible Development of AI, this normative text is positioned in the global geopolitics of AI as an approach aiming to make AI an instrument of progress and inclusion through humanly thought out and ethically implemented algorithms. Developed from an original methodology called “citizen co-construction”, the Montreal Declaration is based on 10 normative ethical principles, the aim of which is to frame the development and deployment of AI. However, by analyzing not only the geopolitical stakes of AI, as well as its ethical structuring around concepts such as “digital humanism” and “human agentivity”, two major observations arise: An absence of real consensus on a fundamental approach, which should guide the development and deployment of AI and above all, a techno-progressive perspective of the Montreal Declaration articulating a techno-optimistic vocation of AI. these observations demonstrate, on the one hand, the complexity of positioning the Montreal Declaration for an ethically responsible AI in the challenge of international regulation; and on the other hand, the challenge of a rigorous ethical compromise between the development of AI and the permanence of human agentivity. Because the techno-optimism of the Montreal Declaration in the face of AI, brings itself the increased risk of the empowerment of artificial intelligence systems, by finally putting forward the question of a transhumanist perspective of the development of AI.