[brushfoot]

I disagree that this is a useful springboard. And I say that as an AI optimist.

A report full of factual errors that a careful intern wouldn't make is worse than useless (yes, yes, I've mentored interns).

If the hard part is the language, then do the research yourself, write an outline, and have the LLM turn it into complete sentences. That would at least be faster.

Here's the thing, though: If you do that, you're effectively proving that prose style is the low-value part of the work, and may be unnecessary. Which, as much as it pains to me say as a former English major, is largely true.