Thu October 27, 2011
Intelligent Design: McCarthy, Myself And A.I.
Adam Frank is an astrophysicist at the University of Rochester. He is a regular contributor to the NPR blog, 13.7: Cosmos and Culture.
What is going to happen when our machines wake up? What will happen when all these computers which run our lives suddenly become intelligent and self-aware. It's a question which makes sense to today as the world marks the recent passage of John McCarthy.
McCarthy was an emeritus professor at Stanford who was known for pushing the boundaries of what was possible with computers. He developed the venerable computer programming language called Lisp, using it develop one of the first computer chess programs. He also helped spur the growth of the internet through his early time-sharing technologies (these were like an early version of cloud computing). But it is for AI that he will be remembered.
Artificial Intelligence has become such standard part of culture that I can refer to it by its initials, AI, and you know what I'm talking about. But the possible consequences of developing an intelligent self-aware machine are just as familiar. If I say "robot overlords" you will still know what I mean.
There are basically two possibilities we naturally imagine when we think about AI. In the first the development of machine intelligence brings us a new age of wonders. In collaboration with these AIs we cure cancer, solve global warming and design ships that carry us to the stars. It's a machine-human paradise. Think of the efficient feminine voice of the computer on the Starship Enterprise if you need an example.
The other possibility — the robot overlord version — tends to get more play. We build machines. The machines wake up. The machines try to wipe us out. From the errant HAL in 2001: a Space Odyssey to the machine masters in The Matrix or Battlestar Galactica we have given fictional flesh (or perhaps steel and plastic forms) to our worst fears.
Many AI researchers can see both possibilities hovering somewhere in the future. While some scientists admit its difficult to imagine why a conscious hyper-intelligent machine would want to continue doing our bidding others think we will understand how to avoid that fate in the very process of building a conscious machine.
Some ambitious scientists in the field speak of the race to develop "good" AI rather than the "bad" kind. For these researchers the dangers of getting it wrong make the global warming look like a game of checkers.
The fact is we don't know what will happen. It's even possible that the intelligence from machines will be so different from our own that entirely new forms of cooperation or antagonism will come to pass.
One thing is certain though. For a culture that is ever more dependent on its machines, humanity will have to learn to share the world with our silicon children.