Opinion: AI — unintended consequences

As far back as the early 1960s, computer scientists have toyed with the idea that computers might one day possess artificial intelligence (AI) that would surpass and — perhaps — supplant human intelligence.

When I taught computer science in the early 1980s at Bakersfield College, I introduced my students to Eliza. Eliza was developed by Joseph Weizenbaum at the MIT AI Laboratory between 1964 and 1966. Eliza was a computer psychoanalyst, and I had my students use the program just to get used to interacting with the computer.

The “school” of psychoanalysis to which Eliza subscribed was Rogerian. Carl Rogers developed the “client-centered” approach to the treatment that Freud called “talking therapy.” Rogerian psychoanalysts generally echoed back the comments that their clients made. In like fashion, when my students interacted with Eliza, the computer would “echo back” to them using a method that Weizenbaum called “pattern matching.” So, if a student typed “I’m feeling stress because of my astronomy class,” Eliza would respond, “Why do you think that your astronomy class is causing you stress?”