@cshlan
People are using it exactly for that purpose - finding information.
Follow
@bibliotecaria @cshlan@toot.cat and that is a huge mistake. Given the way ChatGPT was discussed by media as well as in academic conversations, it’s an understandable mistake. But still a very fundamental one.
@ChrisWilms23 @bibliotecaria @cshlan
I totally agree that there is a pervasive misunderstanding about the capacities of large langue models, but with the state-of-the-art changing so rapidly I can't say I totally blame anyone for taking the bot at its word.
That said, I think #OpenAI could have been a lot more safety-focused and made the disclaimers around factuality a lot more forward.