No. It needs to depend on how close the written code (and maybe the UI design if it passes the threshold of originality) is to the original, otherwise you end up protecting ideas and functionalities, which has always been held as beyond the purpose of copyrights.

People are going to find ways using AI to create functionally equivalent clones where the methods, flows and structures are so radically different from the original, that it can't reasonably be called a derivative work in the copyright sense.