This is exactly it. Also the reason that inputting an excerpt of a classical text (cfr. Dracula, Frankenstein, …) returns “likely AI” results - it’s literally what it was trained on.