Are AI developers deliberately trying to mimic the worst human traits, such as our general reluctance to admit our own mistakes?
My expectations with ChatGPT and Bing have, so far, generally been full of the same sort of frustrations that one might face when talking to a call-centre operative, or when engaging with a know-it-all on a web forum.
If this petulant behaviour hasn't been deliberately programmed into these bots, then it's interesting to see them evolving to exhibit these traits.
And if they have been programmed to be as stubborn as they are then.... Why?!!
Have a look at the transcript snippets below... I'm sure many of us can relate to asking a question from somebody who, while they're trying to be helpful, doesn't really know what they're talking about. And the responses are exactly the sort of thing said person would come up with when called out.
My expectations with ChatGPT and Bing have, so far, generally been full of the same sort of frustrations that one might face when talking to a call-centre operative, or when engaging with a know-it-all on a web forum.
If this petulant behaviour hasn't been deliberately programmed into these bots, then it's interesting to see them evolving to exhibit these traits.
And if they have been programmed to be as stubborn as they are then.... Why?!!
Have a look at the transcript snippets below... I'm sure many of us can relate to asking a question from somebody who, while they're trying to be helpful, doesn't really know what they're talking about. And the responses are exactly the sort of thing said person would come up with when called out.