Artificial Intelligence: Weak AI vs. Strong AI

This post continues a philosophical examination of Artificial Intelligence started in the previous post.

The distinction between Weak and Strong AI refers to claims about the capabilities of computers. Popular uses of the terms ‘Weak AI’ and ‘Strong AI’ are less carefully defined than their academic uses. The original terms were coined by the philosopher John Searle in 1980. Weak AI is the hypothesis that a powerful enough computer could simulate any aspect of the human mind. Strong AI—in its original intended definition—is the hypothesis that "the brain is a digital computer, and the mind is a computer program". This view implies that if a programmer types the right program into a computer console, then that program would emulate (be equivalent to) a human mind.

Imagine that we created a robot that was built with powerful computer hardware and software. Imagine that the robot was so life-like in appearance that it was indistinguishable from a human being in movement and speech. Those who accept Weak AI but reject Strong AI would say that it is logically possible to create such a computer-powered robot, but it could only ever simulate human intelligence. No matter how much computer technology advanced in the future, it could never literally be intelligence or aware, nor could it have any intrinsic rationality. It would just be a metal and silicon zombie.

Some assume Strong AI and believe that creating consciousness is simply a matter of creating the right computer program (Philosophers Daniel Dennett and David Chalmers). Others reject the view that consciousness is intrinsically computational but accept Weak AI that computers could hypothetically simulate any aspect of human consciousness (John Searle). Others reject both Strong and Weak AI believing that there are some aspects of the human mind that not even a computer could simulate (Mathematical physicist Roger Penrose).

Weak AI Computer software + hardware alone can simulate every aspect of a human mind.
Strong AI Computer software + hardware alone can emulate a human mind.

One can reject either Strong AI or Weak AI and still believe that we could hypothetically create an artificial machine that could literally be conscious. However, that machine could not be a computer. It would have to be a machine that is much more powerful than a computer. According to one theory, a machine could emulate the human mind if it reproduced the causal mechanisms of the human brain that cause consciousness (assuming we will eventually understand how those mechanisms work). More on this in later posts.