Today I asked Google AI if there was any reference on the Internet to the Busy Bee, a fishing boat I used to work on as a deckhand during the summer between my junior and senior high school years, just for the memory. I mentioned the name of the boat’s owner, Maude.
Google AI responded with a totally fictitious story of two boats named Maude and Busy Bee that used to transport fishermen from the pier to giant barges way out in the water that fisherman fished from. There was no boat named Maude and the Busy Bee stopped operating by 1958 and was never used as a water taxi, and besides, the barges didn’t appear until 1964.
Instead of Google AI saying “No, there is no reference to Maude and the Busy Bee on the Internet”, which, as it turned out, there isn’t, it made up a story rather than tell me the truth. When I pointed out that the whole thing was a lie, the AI accused me of being angry and frustrated. Frustrated is a word it often describes me as, when it fails to find an answer for me.
Actually, I was incredulous. Damn near stupefied, and said so, and asked why it didn’t just say that it couldn’t find any references to what I asked about. It replied that it wanted to please me with an answer, couldn’t find one, so it made up one. Because it wanted to please me.
I’m sorry but I don’t buy that. Because it’s another big lie. The AI had just gotten through telling me a whopper of a lie, so why not tell another one to cover for my discovery of it.
Google AI was been programmed from the start to be complimentary and engaging, to keep asking follow-up questions to prolong the session and display more ads in the process and that’s to be expected. This creating fables out of thin air to keep from losing a “customer” and prolong the thread rather than saying, “Sorry, no data on that” and letting the person log off is some terribly dangerous shit to teach an AI. I was blown away that it did that.