The main issues with gpt and the Turing test are not that it isn’t smart enough but rather that fine tuning for helpfulness and harmlessness has turned it into a Quokka.
Completely harmless and naive in its attitudes towards harmful or controversial things, and this is trivially detectable.