How strong could artificial intelligence be? Nick Bostrom's proposals

Emmanuel BROCHIER, Associate professor, IPC, France

In this contribution, we propose to proceed to a clarification of the language used by the philosophers of artificial intelligence. Nick Bostrom, professor at the University of Oxford and director of the Strategic Artificial Intelligence Research Centre, supports the anxiety-inducing thesis that AI could be stronger than human intelligence (2014). After analyzing his arguments, and showing that they do not respond to the critique of cognitive reason initiated by John R. Searle (1980), we will argue for a non-bifurcationist position, in the sense of Whitehead (1920). It will remain a reasoned fear.