I do think something that the discourse seems to miss here is that scanlators have been using semi automated OCR tools which hook into translation and LLM api’s for years to streamline thier efforts.

Heck, I’ve done this with open source manga OCR tools for fun as a personal project. It works really well. With billions of dollars, these companies will develop and deploy superior enterprise tools with which an individual could probably translate entire volumes in one workday.

I’m not saying they won’t be lazy and sometimes miss important literary meaning, but they will succeed in making mass translation possible.