It often makes some small mistakes, but little nudge here and there corrects them, whereas with human you have to spend a lot of time explaining why something is wrong and why this and that way would rectify it.
The difference is probably because GPT has access to a vast array of knowledge a human can't reasonably compete with.