I work on a rewriting tool. This is the question we get asked most and answer least comfortably, and it comes up in dev circles as often as it does in academic ones now that a model drafts the first version of half the prose anyone ships.
The assumption underneath it is that a disclosure requirement and a detector output are the same object. They are not, and the whole thing falls apart once you separate them. Everything below is what our own public pages commit to, including the parts that are bad for us.
The short answer
No. If your institution or your journal requires you to disclose AI involvement in producing a document, running that document through a rewriting tool afterwards does not remove the requirement. Rewording changes how something is expressed; it does not change what happened while the document was being written. Our own ethics page states this: "The tool changes how something is worded; it does not change the fact that AI was involved."
The assumption worth naming
