I'd much rather read someone's actual errored, nonnative writing than whatever an LLM would produce from it. Not only because it's annoying reading the same fake style over and over, but also because the less fluent they are the less able they are to tell when the LLM output is changing things in ways that don't reflect what they're thinking.

And if the main complaint is just a few odd words or structures, it's really not that big of a deal to me.