I wonder if it is not laziness but boredom. I find that Chat is often too verbose (even when I ask it not to be), and I skim the answers. Reading the answers it provides feels like a chore because it is so darn boring. Even when I'm interested in the information it's pulled up for me. None of the agents I have used have (yet) learned how to provide consistently interesting writing.

I imagine if the researchers had placed a restriction on the subjects using Chat such that they had to do their best to make it look like Chat hadn't written the essay, they would have performed significantly better. People like puzzles, not factory jobs.