A significant amount of historical work is re-analyzing existing, known material rather than seeking out novel sources. I do know some people working in class

A significant amount of historical work is re-analyzing existing, known material rather than seeking out novel sources. I do know some people working in classical literature that have been testing LLMs against untranslated sources and finding them perform reasonably well. It's completely within the scope of possibility to imagine them becoming more useful for academic work over time.