"Judge weighs punishment for lawyer who didn't bother to verify ChatGPT output."
Also this: "OpenAI isn’t doing enough to make ChatGPT’s limitations clear"
*weary sigh*
I'll just say, once again, that I think it was a bad idea for LLM/AI to be co-opted for use in things like search engines and the like, i.e. things that traditionally rely on... well... being reliable.
But then, as that Verge article points out, it's not as if Wikipedia or Google or whatever have always been perfect sources of inviolably accurate information themselves, even before the proliferation of LLM/AI. I remember when my teachers would typically ban the use of Wikipedia as an original source in school assignments (and I would hope that's still the case today as well, because it's not like Wikipedia has magically become more accurate in the intervening years since I was last in school). The same should be the case for ChatGPT and such, too, what with ChatGPT's tendency for making shit up whole cloth (not unlike some Wikipedia editors).
Also this: "OpenAI isn’t doing enough to make ChatGPT’s limitations clear"
*weary sigh*
I'll just say, once again, that I think it was a bad idea for LLM/AI to be co-opted for use in things like search engines and the like, i.e. things that traditionally rely on... well... being reliable.
But then, as that Verge article points out, it's not as if Wikipedia or Google or whatever have always been perfect sources of inviolably accurate information themselves, even before the proliferation of LLM/AI. I remember when my teachers would typically ban the use of Wikipedia as an original source in school assignments (and I would hope that's still the case today as well, because it's not like Wikipedia has magically become more accurate in the intervening years since I was last in school). The same should be the case for ChatGPT and such, too, what with ChatGPT's tendency for making shit up whole cloth (not unlike some Wikipedia editors).