It's an ok study to show people are predisposed to assume AI works are bad.
However the headline and article imply that people are wrong that the quality of the writing is bad. So a study "independent of the actual quality of the writing" goes directly against the conclusion they are pushing here. (the paper itself does not have this issue).
They have to construct the study independent of the actual quality of the writing. And it’s clear here that many people are just biased.