Maybe its just the simplest image to make and each model follows that path because they too find it the simplest.

Perhaps the prompt should include what the scene should contain.