Penrose, Godel & artificial intelligence

A refutation of Penrose’s Godelian case against artificial intelligence

Selmer Bringsjord & Hong Xiao
Dept. of Philosophy & Cognitive Science
Dept. of Computer Science
Rensselaer Polytechnic Institute

February 2000

This paper is sometimes glibly quoted as a complete refutation of the arguments relative to the Godel theorem and the brain, but in reality the opinions of the authors are much more mixed. The authors emphasise the distinction between ‘Strong Artificial Intelligence’ (AI) and ‘Weak Artificial Intelligence.’ The former claims that the sensation of subjective consciousness could be created by some appropriate computation on a manmade computer. The latter only claims that computers can potentially simulate all the functions of the human mind but would not become conscious in the process. Penrose has adopted a third position, proposing that Godel’s theorem means that computers cannot perform some of the functions of the brain, such as those involved in mathematical understanding, and by extension it would also appear that they cannot produce consciousness.

The first named author of this paper, Selmer Bringjord, takes the intermediate position of opposing strong AI, but accepting weak AI. He has published 13 formal arguments against strong AI. However, he accepts the weak AI position, and claims to expose fallacies in Penrose’s reasoning relative to weak AI, as outlined in ‘Shadows of the Mind‘ and subsequent discussions in Psyche. This involves complex and technical logical processes, which can be accessed at www.citeseer.ist.psu.edu/320350. Bringjord, at the same time argues that when proponents of strong AI, such as Laforte, Hayes and Ford (LaForte et al 1998) have attacked the Godelian case against strong AI, they have fallen into exactly the same fallacies as Penrose.

Bringsjord also critical of Penrose’s scholarship relative to strong AI, but does not think that this represents any refutation of his or Bringjord’s position on strong AI. Bringsjord believes that it is possible to derive the denial of strong AI from Godelian facts. At the same time, he thinks that his own paper ‘Argument from Infinitary Reasoning’ (Bringsjord 1997b) captures Penrose’s core intuitions better than any appeal to Godel.

Because Penrose is a mathematician, he has argued outwards from concepts of mathematical understanding and truth, while consciousness as such has followed from this. The emphasis on mathematical understanding requires a refutation of weak AI. However, a non-computational basis for consciousness and the qualia only requires a refutation of strong AI.

 

Posted by

Leave a Reply