Joseph Weizenbaum: Fighting the False Gods

It was with unexpected sadness that I read this morning about the passing of Mr. Joseph Weizenbaum.

As author of a computer program called Eliza that was designed to simulate a psychiatrist, Massachusetts Institute of Technology professor Joseph Weizenbaum created a beguiling artifact of early computing. But after test subjects told him the program really empathized with their problems, Mr. Weizenbaum became a digital Jeremiah, and spent decades preaching the computer apocalypse.

You remember Eliza, at least if you were part of my generation.  That first glimpse into artificial intelligence as a kid, trying to stump a computer in such inane conversation as:

Young woman: Men are all alike.
Eliza: In what way?
YW: They're always bugging us about something or another.
E: Can you think of a specific example?
YW: Well, my boyfriend made me come here.
E: Your boyfriend made you come here?
YW: He says I'm depressed.
E: I'm sorry you're depressed.
YW: It's true, I am unhappy.

Well, ok, I didn't have that specific conversation.  But I bet you played it once or twice.  Mr. Weizenbaum went on to identify some potential dangers of computers.  Occasionally he may have strayed into the deep end, but some of his thoughts were unfortunately prescient.

"How long will it be before what counts as fact is determined by the system, before all other knowledge, all memory, is simply declared illegitimate?" he wrote. 

He later expounded on this dire prediction.

He soon soured on computers and condemned automated decision making as antihuman.  In a lighter moment, he called the, "a solution looking for a problem."  At other times compared them with National Socialism, Karl Marx and Stalin.  Mr. Weizenbaum then gave up working on artificial intelligence.

Like I said, some of it is a little... extreme.  But the fundamental concept provoked some thought.

"He was deeply troubled by the fact that it was easy for people to mistake such simple pattern matching for true understanding," say Bruce Buchanan, a University of Pittsburgh computer science professor who debated artificial intelligence with Mr. Weizenbaum in the 1970s.  "He raised questions that are as relevant today as they were when he first raised them."

How true.  It is very easy to get brainwashed into the idea that a computer creates order and understanding out of complexity, when the complexity may not exist in the first place.  Which is why we continue to rant against the dangers of what we call the false god of the almighty algorithm.  Instead of applying a complex computer solution to a simple problem, perhaps it's easier to just run out to Staples.  Fix the unnecessary complexity and find the underlying simplicity before considering a software band-aid.   

Comments

[revere]

See our own personal reflections on Joe's passing here:

http://scienceblogs.com/effectmeasure/2008/03/rip_joseph_weizenbaum_berlin_j.php