This feels like a flawed test. The paragraphs are taken out of context, so the main idea of the work disappears. And the authors are unclear — there are plenty of human writers who write worse than AI. It might be more productive to accept that AI probably won’t create, in the human sense, and move forward with that assumption.