A UC San Diego study found GPT-4.5 was judged human more often than real people in live chats, raising sharper questions about AI disclosure, trust, and online identity.
Some results have been hidden because they may be inaccessible to you
Show inaccessible results