"80 experimental scenarios help us understand how humans judge AIs as opposed to other humans in the same situation"--