Laughing with machines

Abstract

This article will analyse the preconditions of sense of humour for artificial intelligence. Can artificial intelligence have a sense of humour? Is there a difference between human and machine laughter? Some machines already fulfil certain conditions which are associated with the human sense of humour: on the most superficial level machines appear to laugh and produce jokes, and they recognize sarcasm and punchlines, and they can evaluate funniness. In short, artificial intelligence is already able to recognize humour, and reacts to it accordingly. Furthermore, people laugh with humorous machines. However, it is still uncertain whether artificial intelligence can have a sense of humour or not, at least in comparison to a human sense of humour. To build bridges between AI research and philosophy of humour, this article proposes that there are (at least) five notable philosophical issues to be addressed if we are to accept that machines can have a (humanlike) sense of humour. These principles are: 1) worldview, 2) self-consciousness, 3) self-reflection, 4) self-criticism, and 5) losing control.