Will humans ever build a real Artificial Intelligence?
This question has been asked in many different ways over the centuries, and the current way of phrasing the question is to do with Aritificial "General" Intelligence. There are various weasel-words that are added to AI, such as "full" AI, "hard AI", or (as above) "General" AI.
And if you do build a real Artificial Intelligence.... how would you know? (We'll leave aside the slightly spookier question: Would it want you to know?)
The popular press generally latches onto a thing called "The Turing Test" as a way of measuring for "real" AI.
Richard Hamming is a bit smarter than the popular press, and is well ahead of the field when it comes to this question of how we'll determine real AI. He gave a series of lectures on AI, which have also been captured in his books (freely available online). His conclusions are simple:
We'll never succeed at building a 'genuine' AI.
Not because we are incapable of building machines that are many times smarter than ourselves (we've long ago achieved that) but because our egos prevent us from accepting any machine as being generally intelligent.
As soon as we achieve General AI, according to the measurement system of the day, that measurement system will be roundly debunked, and a new measure will be invented, by which the current contender fails the test.
In early 2015 I was reading about Hamming's opinions, which date from the 1980s or earlier. Shortly thereafter a lot of news articles appeared claiming that a computer program had passed The Turing Test.
The Turing Test had long been seen as "the bar" for accepting General AI. But no articles were published with headlines such as "General AI has been achieved." The headlines were, at best, making claims about the Turing Test. They were no longer willing to imply that the test itself had any particular significance.
And within a few days, the mood changed considerably. A month later the entire test was declared "Almost worthless".
There is an uncanny valley into which all intelligent machines must inevitably fall.
Would you put your life in the hands of a machine?
Every time you cross a bridge, you are putting your life in the hands of technology. When you fly in a plane. When you receive a pace-maker. The machines keep you alive, and you readily let them.
- Wikipedia: Artificial General Intelligence
- Hamming, Artificial Intelligence Part I
- Hamming, Artificial Intelligence Part II
- Turing Test
- Saturday Morning Breakfast Cereal: Ha! Robots have achieved sentience!