home | science

Can computers think?

In his book A Little History of Philosophy, Nigel Warburton has a chapter titled "Can Computers Think." Warburton is describing the position of others, and he writes of an experiment to detect whether individuals can differentiate between answers from a human and answers from a computer.

Warburton mentions the philosopher John Searle, a professor of philosophy, retired somewhat, at UC Berkeley. He describes Searle as claiming that computers "don't really have intelligence and don't really think. All they do is shuffle symbols around following rules that their makers have programmed into them."

I'm with Searle. I don't see a good definition of what thinking is coming from people who claim that computers can think.

A computer is a robot that can be programmed to respond, maybe even to turn itself off and on. Humans have an ability to choose. Choice allows them to create, and they have an ability to imagine beyond the present moment – to plan. This gives them survival potential rather than walk off a cliff or into oncoming automobiles. I'm waiting for news of a computer that makes choices that are not programmed. Computers can play chess or win at Jeopardy, but these are mechanical and programmed games. Show me a computer that has imagination: a computer that can draw from its experiences and write poetry, plan for the future or can draw from its memory bank to decide whether to convert to Buddhism, Catholicism or agnosticism. Those claiming that computers can think are not giving us a good definition of what thought is.

Questioned by the comedian John Oliver, Stephen Hawkins, in a comedic mode, spoke of artificial intelligence making improvements to itself and being a danger to humanity. He spoke of "a story that scientists built an intelligence" and it was asked, "Is there a god?" It replied: "There is now, and a bolt of lightening struck the plug so it couldn't be turned off." Oliver likes to rest his comedy on the factual, but this was, as Hawkins said, a story. It was a joke. It would take an electronic machine imagination it doesn't have to want to improve itself or to describe itself as a god it invented or as Allah or Jehovah. If you have any evidence to the contrary please let me know.

Humans from infancy learn to identify and to pick up and manipulate objects through trial and error. Computers are faster than humans at complex number crunching but slower than humans at everyday tasks like making a bed or unloading a dishwasher, tasks now described as hard for robots. What robot-makers wish they could do is replicate the wonders of human biology – the product of 400 million or so years of trial and error in biological evolution. I can't say it's impossible. But as of now, in the early 21st century, robot-makers are a long long way from it.

Copyright © 2016 by Frank E. Smitha. All rights reserved.