Think Discord chats, but with AI characters in addition to humans.
A UC San Diego study found GPT-4.5 was judged human more often than real people in live chats, raising sharper questions ...
Some results have been hidden because they may be inaccessible to you
Show inaccessible results