Article:

Die Maschine auf der Couch. Oder: Was ist schon ‹künstlich› an Künstlicher Intelligenz?

The alignment of artificial intelligence with the history of brain research, as can be seen in current debates in media studies, is deeply problematic. The connectionist approach (anchored in neuroscience) – i.e., the assumption that intelligence can be reduced to its physiological wiring in the brain – has gained significant popularity with the rediscovery of artificial neural networks. This article follows the traces of such a cybernetic definition of intelligence and contrasts it with two case studies from the field of computer-based psychotherapy (ELIZA and PARRY). The central question is to what extent the mechanistic world view of connectionism leads to a reductively biologistic idea of intelligence, and how this idea might be reversed by a psychoanalytical critique?