Maybe waxing too philosophical for this listserv, but I don't think it's fair to use words like "lazy" for any AI - any words, in fact, that have a moral judgement attached. AIs, being just machines, do what they do. If Saddam Hussein drops a political opponent in the wood chipper, no one thinks the wood chipper was evil for doing what it was designed to do, nor (in my opinion) is the inventor of the wood chipper a villain.
It isn't clear to me how an AI COULD consider whether its information is, for example, truthful. The lawyer who asked the AI for a legal brief got what the AI thought looked like a legal brief; the fact that some or all of the precedents weren't real didn't figure into its goals. And by the way, Rob, I know you didn't mean anything much by "lazy" - probably just the first word that came to you. I'm just springboarding into a topic much on my mind recently. --- Bob Bridges, robhbrid...@gmail.com, cell 336 382-7313 /* D'you call life a bad job? Never! We've had our ups and downs, we've had our struggles, we've always been poor, but it's been worth it, ay, worth it a hundred times I say when I look 'round at my children. -from _Of Human Bondage_ by W Somerset Maugham */ -----Original Message----- From: IBM Mainframe Discussion List <IBM-MAIN@LISTSERV.UA.EDU> On Behalf Of Rob Schramm Sent: Tuesday, January 14, 2025 19:30 It is all about the prompting to prevent the AI from being lazy and making things up or mixing things up. I think it would be less of an issue the more examples that were consumed by the AI during training. ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
