Probably not. I assess grant applications for government funding for commercial R&D (not in the USA). We do see some applications which we believe are written by AI, along with a lot written by grant-writers. Now to be clear: we explicitly don't care if an applicant has external assistance. The problem comes if the writer (human or AI) does not have the technical domain knowledge or the market knowledge and they don't get their customer to collaborate on putting this together. Sometimes you get chunks of text which make little sense in context; more commonly it is fluffy generalities where there should be specifics.
We have a very structured way of marking proposals. There are a set of questions, each with a maximum number of words, and a few PDF appendixes with a low page limit. For each question, we have say 10 marks, and for each band of 2 marks, there are criteria. Of course it can't be entirely objective, but the point is that any information not contributing to these criteria is discarded. Worse than that: some applicants end up burying the real information in words, so that it is difficult to tell what they want to convey.
When we get a good application, it is very clear that the writer knows the domain and knows the target market. The risk register is another give-away: if you can't guess the rough area of the project from the risk register alone, there is something seriously wrong. If you can get an AI to write that well, then fine, but you will probably need to involve it in managing the project as well. That day is not yet here.
