This isn't a study, it doesn't "objectively show" anything. It's an unreviewed discussion paper with questionable methodology.

The conclusion could just as easily be that an AI is better at writing engaging short-stories than the single author they chose.