Gödel's theorem and the possibility of thinking machines: “Do androids dream of electric sheep?”
Burton H. Voorhees
Complexity 1995 DOI: 10.1002/cplx.6130010308 (opens in new tab)
Study at a glance
AI-extracted from the abstract| Characteristics | Theoretical or philosophical paper Peer reviewed |
|---|---|
| Key points | Argues that the debate over machine thought and self-consciousness has become a practical question with the digital computer. The authors describe the "strong AI" hypothesis and its proponents' claims about downloading minds and computers as evolutionary successors, and note that the Gödel-based challenge to strong AI fails to establish that the mind possesses the transcomputational capacities it claims, which proponents cite to call the theorems irrelevant. |
Abstract
he question of whether or not machines can think is intimately connected to a number of age-old questions:T What is thinking?What is consciousness?What is selfconsciousness?These questions have been central to philosophy from its very beginnings, but now the invention of the digital computer has brought them down from the mountains of philosophical speculation to the plains of practical application.The claim that computers are capable of self-consciousness is called the strong A1 (Artificial Intelligence) hypothesis.Proponents of this idea have gone as far as to suggest that immortality may lie in downloading minds onto mainframes 111 and that computers are our evolutionary successors [a].Challengers of strong A1 cite Godel's incompleteness theorems.These theorems, they claim, demonstrate that there are capacities ofthe human mind that can never be matched by a computer [3].The weakness of this challenge is that it does not establish that the mind actually possesses the claimed transcomputational capacities l e g , 4,5].For this reason, proponents of strong A1 say the incompleteness theorems are irrelevant to strong A1 and have no place in the debate [6].