I was just reading Ethan Mollick et al's AI Agents and Education paper (linked above) and wonder if you have read it too. Because he is at the Wharton School its more about simulation tutoring, mentoring, and other repeated tasks in education than your history simulations, but this is a big area of research at Harvard as well, where a Physics knowledge expert was trained to be a first responder for students who would normally reach out to their TAs. I don't see a publication on this yet but saw a presenation at Stanford by Kelly Miller and Greg Kestin.
I was looking for an example of thinking beyond chatbots for teaching that takes advantage of the weirdness of AI and this appeared in my inbox.
https://papers.ssrn.com/sol3/papers.cfm?abstract_id=4871171&utm_source=substack&utm_medium=email
I was just reading Ethan Mollick et al's AI Agents and Education paper (linked above) and wonder if you have read it too. Because he is at the Wharton School its more about simulation tutoring, mentoring, and other repeated tasks in education than your history simulations, but this is a big area of research at Harvard as well, where a Physics knowledge expert was trained to be a first responder for students who would normally reach out to their TAs. I don't see a publication on this yet but saw a presenation at Stanford by Kelly Miller and Greg Kestin.
Can confirm, here's a 600 page story written with it starting in 2015. https://archiveofourown.org/works/46518058/chapters/117135472