Photo by Işıl via Pexels. |
Naturally, nobody is more into AI than historians! Just kidding, they hate it. At least, that's what I've observed in an entirely unscientific study in which I read an article about this issue and observe what people with "History" in their job titles tend to say. Obviously this is not true across the board, but it does feel as though the novelty of the past several months has been recieved as a harbinger of doom rather than anything at all helpful.
Undoubtedly skepticism is justified, even beyond the issues of plagiarism and equity and scary future dystopiae (ChatGPT told me the correct word here is "dystopias," which I agree with and have also rejected for reasons of vibes). For example, I gave ChatGPT-3 Jourdan Anderson's letter a while back, and it's not great at understanding satire-- it read Anderson as deeply sincere, which is, as you may know, not the best reading of this source. Read it yourself if it's new to you and see what you think the author intends! Moreover, it also makes stuff up, which as you might have guessed, can be a problem. It's great at assembling a bibliography, for example-- I tried it for sources about one of my main subjects of my disserations, Joe F. Sullivan, and it came up with a handy list. However, as Aaron Welborn notes on the Duke University Libraries blog, it's also wildly capable of making up sources out of thin air! So, great at making a bibliography, not so great at making a bibliography of sources guaranteed to actually exist. It also can't always be trusted even just with making a list-- for example, of presidents. That's just how it works. So it's only really a starting point you then need to guess and check. And let's not forget "Historical Figures Chat", which attempted to make AI chatbots of notable historical figures based on ChatGPT, thus incorporating some ahistorical and incredibly odd "PR nonsense" into responses from multiple architects of historical atrocities. (This tool would have been an interesting concept to ponder in my iteration of the Fiction and the Historical Imagination class, in that many of the less valuable approaches to historical fiction that we looked at dealt with a cultural penchant for believing we can know a person in the past, and mythologizing them based on the image we've created of them, and also in that this thing is effectively creating very specific little historical fictions).
But! As you probably know, my wheelhouse is less locking down and preventing people from using stuff and writing dramatic take-downs of concepts and more about poking into and exploring what things can be good for, even if I, you, and/or the profession in general isn't psyched about them immediately. So, throughout the rest of the year, I'll be releasing a grab bag of AI-related options-- things teachers can do with the current AI tools without either totally forbidding them or making them central to your course creation. Some will help you plan to teach or create teaching materials, while others will be ways you can assign AI usage as part of the learning process. Some I've come up with, while others I'll feature have been suggested by others.
I'll note here that due to my own biases as a trained humanist, I've avoided options that mostly boil down to "have the thing create a thing"-- I tend to find that the writing that text-focused AI tools produce is a little uninspired; many of them literally work by stringing together words in the manner and order that they are frequently strung together, so that's not terribly surprising. It's also the opposite of what I'm usually trying to accomplish in my classes. I've always tried to teach and to get students to find new connections, words that haven't been put together that often, ideas not commonly associated. As a result, the more the tool is creating something toward some sort of final product, the less the options for creating material are interesting to me, though there are undoubtedly uses in that line you can find recommended elsewhere.
Finally, if you have an AI usage for teaching you'd like to share, I'd love to hear from you! This series will be percolated on and added to periodically for as long as I keep thinking of or coming across uses that seem useful and exciting to me.
No comments:
Post a Comment