I don’t think that works the way they think it does, though. Just imagine analog data instead of digital, and take two copyrighted works, break them down into blocks, mix them together with some randomizers, and what do you get? A work derived from multiple copyrighted sources. No one needs to “own” the entirety of the source, if any part of it is derived from copyrighted materials in a way that isn’t exempted via eg fair use, then the entirety of the resulting work is in violation and is fair game for lawsuits, etc.
Simply moving this to the digital domain from analog doesn’t change the logic, does it?