"Chatbots were generally bad at declining to answer questions they couldn’t answer accurately, offering incorrect or speculative answers instead.
Premium chatbots provided more confidently incorrect answers than their free counterparts."
Sounds promising.
https://www.cjr.org/tow_center/we-compared-eight-ai-search-engines-theyre-all-bad-at-citing-news.php
Premium chatbots provided more confidently incorrect answers than their free counterparts."
Sounds promising.
https://www.cjr.org/tow_center/we-compared-eight-ai-search-engines-theyre-all-bad-at-citing-news.php
Comments
There is no other possible result of using these models. To even 'test' these asinine uses is like repeatedly dropping things to 'test' if gravity still works.