Skip to main content

About this free course

Download this course

Share this free course

Machines, minds and computers
Machines, minds and computers

Start this free course now. Just create an account and sign in. Enrol and complete the course for a free statement of participation or digital badge if available.

3.3 Artificial intelligence – the quest

The goal of artificial intelligence is to build intelligent systems. So far, we've considered what intelligence might be and how we might recognise an intelligent system when we see one. But now let's try to unpick the real nature of the quest for artificial intelligence a bit further. When we say we want to build intelligent systems, what are we really trying to achieve?

You might recall a question I posed earlier about the attitude of the automata builders of the 18th century to their creations. I asked whether Vaucanson, for example, might have imagined his duck was to some degree truly a living thing, rather than just a clever simulacrum of a real water fowl. If it could be made into a much, much more accurate simulacrum could it become, in some way, the real living thing?

This is a very difficult question. But it is directly relevant to the quest for artificial intelligence. What are we really trying to achieve when we build intelligent computer systems? There are two distinct possibilities:

  1. We are trying to build practical systems that will do certain clever things. This may give us certain insights into the human mental processes that underlie intelligent behaviour along the way, but no more than that. Such systems are not intended to be accurate imitations of mental processes. Moreover, each simulation might be quite narrow in scope – good at playing chess, say, but useless at checkers, language or medical diagnosis.
  2. We are trying to build systems that faithfully copy mental processes. If that is our aim then the question we started with – Could the imitation ever become the reality? – becomes pertinent. Suppose we could very precisely reproduce mental process on a computer; might we end up with something that is genuinely intelligent, is aware, has a mind? Something like you and me?

In 1980, the philosopher John Searle proposed the terms weak artificial intelligence and strong artificial intelligence to describe exactly these two possibilities. Here is a brief outline of his ideas.