I have tried BARD, ChaptGPT and Bing's AI. With each one I have asked the (yes, narcissistic) question "What books has Birrell Walsh written?" At least I know the correct answer.
All three gave incorrect answers. None of them said "Gee, I don't know." They all just gave wrong answers as if they were right.
I then directed them to my website (btwreviews.com) and asked them the question again. Again each gave wrong answers, but DIFFERENT wrong answers than the first time.
When I corrected them, each was apologetic.
I got a couple of Medium articles out of the experience, and a sense of foreboding for anyone who trusts the outputs of these AI's.
I still have to get used to the idea that ChatGPT is a "running" conversation. At least that how it seems when you describe it. I tend to send disjointed queries. Could that impact accuracy?
Loved this episode of Robot Wars!
I love this episode
I love this episode
I have tried BARD, ChaptGPT and Bing's AI. With each one I have asked the (yes, narcissistic) question "What books has Birrell Walsh written?" At least I know the correct answer.
All three gave incorrect answers. None of them said "Gee, I don't know." They all just gave wrong answers as if they were right.
I then directed them to my website (btwreviews.com) and asked them the question again. Again each gave wrong answers, but DIFFERENT wrong answers than the first time.
When I corrected them, each was apologetic.
I got a couple of Medium articles out of the experience, and a sense of foreboding for anyone who trusts the outputs of these AI's.
I still have to get used to the idea that ChatGPT is a "running" conversation. At least that how it seems when you describe it. I tend to send disjointed queries. Could that impact accuracy?