Ni animales, ni máquinas: conocimiento humano y responsabilidad epistémica

Abstract

The 19th century debate about the overlap between the animal and machine kingdoms has been transformed in the 21st century into a controversy about the overlap between human intelligence and artificial intelligence. If we differentiate between the three domains (animal, human and machine), and leaving aside the question of their possible overlapping, we could ask ourselves: how do humans know, do animals know, do machines know? The question that currently arouses most interest is the latter, which usually takes the following form: can machines know as humans do? I will argue that machines do not share with humans any of the four characteristics mentioned above (consciousness, freedom, normativity, responsibility), although they can simulate them. Animals, on the other hand, as far as we know, only partially share the first of the above-mentioned aspects. I will mainly rely on ideas of William Kingdom Clifford (1876-77) and Ernest Sosa (1991, 2007) to defend this view.