Be cautious about artificial intelligence

At a CA conference last century, I listened to a far-out guy with a mass of dreds talk about virtual reality. (Then I got to don a heavy headset and do battle with colleagues and pterodactyls). That guy was Jaron Lanier. Fast forward to my first copy of "Communications of the ACM" as a new member. It is a livelier read than most equivalent publications, a pleasant surprise for a fusty old org like the ACM. In fact I think it nearly justifies the fee without counting all the other excellent perks of membership. And there is an article by Jaron, Confusions of the Hive Mind. And it says something excellent about AI:

Be cautious about the artificial intelligence approach to computer science. It is impossible to differentiate the actual achievement of AI from the degree to which people change when confronted with what is purported to be intelligent technology. We humans are vulnerable to bending over backward, sometimes making ourselves significantly stupider, in order to make an algorithm seem smart. A great many people in the U.S., as well as elsewhere, demonstrated this danger when they interacted foolishly with deeply flawed algorithms related to the credit and mortgage industries.

Syndicate content