The premise that AI and humans should work together sounds great in theory, but have you met a human?
What human won’t stop fact- and double-checking output from a computer? I mean, there are shows to watch and games to play! The output seems plausible enough and who cOuLd poSsiBly geT hUrt. Also, nobody is going to find reason to punish you for trusting the output.
Anyway, this is likely a good time to share my post about Stanislav Petrov: Three lessons from a man who averted nuclear war by not trusting a computer.
https://axbom.com/lessons-from-stanislav-petrov/