3 ms·
> It’s almost always the headlines and PR pieces that exaggerate it. That is definitely happening here. https://jamanetwork.com/journals/jamainternalmedicine/
by ceejayoz 4y ago
> It’s almost always the headlines and PR pieces that exaggerate it.
That is definitely happening here.
https://jamanetwork.com/journals/jamainternalmedicine/fullarticle/2804309?guestAccessKey=6d6e7fbf-54c1-49fc-8f5e-ae7ad3e02231 https://jamanetwork.com/journals/jamainternalmedicine/fullar...
> evaluators did not assess the chatbot responses for accuracy or fabricated information
Yikes. (I do fault the researchers quite a bit for quietly slipping that little detail into a page-long "limitations" section.)