17

The article analyses the publicization of artificial intelligence (AI) and its framing as a public problem under the label ‘AI ethics’, from the point of view of the discourses and actors circulating between the media, scientific, and political arenas. The authors investigate this process by considering a subject that has previously received little attention: the relationship between these three arenas, that is, the mechanisms of circulation and distortion of discourses, the conflicts over definitions, and the resulting reconfigurations for the field and for the modes of public existence of AI. Articulating quantitative and qualitative methods, the study is based on three types of data from the French media, political, and scientific arenas: a corpus of press articles pertaining to AI, covering a period of almost 20 years (2000-2019); interviews with participants in the public debate (scientists, journalists, institutions); and public reports (2006-2019). Driven by increased computing power and the availability of large training databases, the ‘AI wave’ in the public sphere is the product of converging technological, economic, media and political factors, without reflecting any real ‘scientific revolution’ promoted by the media and corporations but contested in researchers’ discourse. The article shows how the public life of AI takes shape at the intersection of competing and often incommensurable discursive registers. Moreover, while discourses on AI and its ethics are ubiquitous in the public sphere where they are emerging as a ‘public problem’ that has become a regulatory issue, outside of specialized spheres they do not really draw on a ‘concerned public’ (Dewey, 2012) engaged in collective inquiry and mobilization. Therefore, the latest ‘AI wave’ and the upsurge of ethical issues seem rather to mark an advanced economization of the field, increasingly subject to private interests, in relation to which researchers and politicians are seeking to position themselves.