Using AI for text corpus research (informal discussion / brainstorming)

At our meeting yesterday, we briefly discussed the possibilities for using artificial intelligence to research large collections of books, articles, and other texts. We thought it would be worthwhile to hold a special meeting devoted specifically to this topic and to explore it together in an informal conversation.
The central question will be: How can AI be used to investigate or conduct research on a large text-based corpus?
We will consider what tools are currently available, how one might organize and approach a large collection of texts, and what kinds of questions could usefully be asked of such a collection. For example, can AI identify recurring themes, locate relevant passages, compare the views of different authors, trace the development of an idea, or help us discover patterns that would otherwise be difficult to see?
We will also consider some of the practical limitations and problems involved, including the reliability of AI-generated answers and the difficulty of working with collections too large to submit directly to an AI system.
This will be an informal exploratory discussion rather than a presentation or technical workshop. No special knowledge of AI or computer programming is required.