Abstract

Three educators discuss teaching history with artificial intelligence (AI) across the secondary and higher-ed classroom. Nate Sleeter argues that history teaching should slow thinking down and value student interpretation over a chatbot's. Kristin Dutcher Mann describes assignments that use and critique AI, including a techno-ethical audit and a comparison of large language model (LLM) answers in Spanish and English with Wikipedia and course readings. Meghan McGlinn Manfra presents a Library of Congress funded learning trajectory for historical thinking that could support personalized pathways, and the session closes with audience discussion in small groups and on Discord.

Chapters

Transcript

AI-generated transcript. It was produced from the recording by speech recognition and cleaned up with AI. It has not been edited by the speaker and may contain errors. The captions on YouTube are AI-generated in the same way. Each timestamp links to that moment in the video.

[0:25] I should we do introductions or just go through yeah all of a sudden have too few hands okay yeah my name is Nate Sleeter I am the associate director at the Roy Rosenzweig Center for History and New Media which is at George Mason University I have a former classroom teacher middle school and high school. I now work on basically creating a lot of digital resources for teaching and learning history, mostly in the kind of 6 to 12, but also into higher education as well. I also teach courses on teaching in the digital age and higher education in the digital age, and I'm really excited to talk with everyone about their ideas for incorporating AI in the classroom and turn it over to my colleagues. Hi, my name is Kristin Dutcher Mann. I'm at the University of Arkansas at Little Rock, which is the urban commuter campus, which has a very large online presence

[1:27] for the University of Arkansas system. I teach classes in Latin American history and I'm the social studies education coordinator. So I supervise student teachers and advise students in that program. I'm here today because I've done some research about student perceptions of the ethics of using AI on assignments with other colleagues at my university across the campus, and I've also written about how to change assignments in the age of AI, how to rethink the kinds of assignments we're giving. Hey there, I'm Meghan Manfra. I am the faculty director for AI for the Data Science and AI Academy at NC State in Raleigh. I'm also in the College of Education. And my forthcoming book is called Fight the Fake, helping students navigate social media and AI for healthy digital futures. So our approach today is to try to make this a conversation as much as possible.

[2:30] I know every session says that, but we are kind of talking about this amongst ourselves and preparing for the session. realize that we're at a time period where we really need good questions as much as we need answers on this. We know just going through the Discord and seeing all the different experiences that you all have come from a lot of really interesting places, really interesting backgrounds and experiences. I think we have a lot that we can help each other out with and inform practice on. So we want to turn this into a discussion. We've got some questions that we're going to put to you and have you a chance to discuss those both on the Discord online and also at your tables. But before that, we're going to get into kind of our perspective and what we see as our main take on AI in the classroom based on what we're working on. So I want to hear from you all. And you can obviously also ask us questions as part of the discussion as well. But just kind of going over, we broke down the issues into kind of three

[3:33] categories. use policies and things like that, but also AI critique and literacy kind of assignments, and then use in instruction, including assessment. So we've got these in main categories, and you can choose which kind of subcategory question you want to talk about. So AI use is something that we've seen a lot of studies come out on recently in the K-12 space. limits by school systems to completely get rid of AI access for students in early grade levels, or even all the way through middle school or even high school. Our research on student use of AI at UA Little Rock is a little bit older. It came from 2023 and 2024 early on, but you can see that there was a gap in every area between students who said they would use AI on assignment and students who thought it was

[4:37] ethical to use AI on an assignment, which is, you know, definitely interesting to investigate further and shows that, you know, students are self-reporting that they might even use it for a task even if they didn't think it was ethical to do so. So on the right, a new research from the Pew Center was disaggregating teenagers' AI use by a lot of different categories. I think the data there is really interesting to look at how students who are of color or in poor families are reporting that they're more likely to say that chatbots are helpful to them as they are doing schoolwork. And this was teenagers, 13 to 19, so it involved high school and college and some middle school students. And so that's just really something for us to consider as educators that we need to keep in mind that student use of AI differs

[5:42] by a lot of different factors. And also student access to AI tools differs by a lot of different factors. So the students who might find it most useful and helpful also may be the ones that don't have the access to it as reliably. um this one was a research study done by ProQuest and it was about how students were using ai in college academic work again likely to use it for things like summarizing and brainstorming and then on the right you're seeing where they were required or encouraged to use AI by their instructors for assignments. And this wasn't specific to history. Neither was the earlier slide. So I'm going to give my perspective on sort of education, history, education, and AI, and then my colleagues will go after. I'm kind of heading up the, what's interesting in talking about our experiences and

[6:44] our backgrounds is that we all kind of have overlapping experience with higher education and with K-12 education. So we can all speak to different parts of this. I've kind of been tasked with the secondary version of this, the secondary school version of this. And so I want to go back to basics. And I'm actually going to start even earlier with, if you'll indulge me, my favorite example of historical thinking that I've ever heard. This is because I think, you know, and we're going to hear a lot about different AI approaches to teaching. And my mind personally is open to all of them. I want to see what it can do. I work at a digital history center with some brilliant colleagues who are using AI for lots of different things. And I'm optimistic that it has uses in education. At the same time, I also think that never before has it become so important to emphasize what historians do as active thinking about the past, as active interpretation about the past, and communicating that to students. So this is a real opportunity, I think, to do that.

[7:45] And so this is my favorite historical thinking example. Comes to us via a second grader in Virginia. Can anybody tell me who the individuals behind me are? One of them's obvious. George Washington. And the other one is? Helen Keller. Very good. So these two are connected. You might not know, but they are. Second grade standards in Virginia study kind of famous Americans. One through the Revolution. Washington. In Reformers, you have to study Helen Keller. Okay, that's what the connection between these two are. So the example of historical thinking I have from a Virginia second grader relayed to me by their teacher is, so why do we study George Washington as an adult, but we study Helen Keller as a child? Who thinks that's a good question? And who's a liar who question. No, I'm just kidding. This is a great question because I think this gets to the core of what historical thinking really is.

[8:47] It reminds me of kind of the original research on historical thinking that Sam Wineburg did in the 90s, that he had historians look at primary sources and high school students look at primary sources, and he found that historians were better at it. Amazing breakthrough. And no, it's actually really interesting because how they were good at it, even Japanese medievalists were able to examine these primary sources and do think alouds, right? And I remember specifically one about, you know, wondering, looking at diary entries and saying, oh, these diary entries, we know what those are like. And looking at the date and saying, oh, the Charles River must have been cold that day. That's very similar, I think, to that second grader who's like, wait a minute, like, what are we not seeing here? What's the bigger story here? this is the kind of questions that historians tend to ask and it's something that i think as we move into the the six to twelve um age rants there's six to twelve sixth grade to twelfth grade um we kind of gets lost in schools sometimes uh because 12th graders 11th graders high schoolers college

[9:48] students don't necessarily ask these kinds of questions right they've been trained to read quickly get to the main idea of the paragraph answer the question move on right those kind of meandering like, oh, I wonder what the deal is with this. This is strange that you then have to train graduate students to do after they get through all of their undergraduate and high school training. It kind of gets lost. And if you look at even these tools here for analyzing primary sources that are used at the K-12 level, you'll see there's variation among them, but they're all basically doing the same thing, which I call slowing down your thinking. It's not meant to be fast. It's meant to be slow. It's meant to notice details, wonder about connections, and try to see what the relationships are. What do I already know? What's my background knowledge? And just wonder what you're missing. What part of the story am I not getting? And where can I go look for it? So to wrap up, this is what I see as sort of the opportunity here is that we need to communicate to students that history is something you actively do.

[10:51] We need to make the case for historians as creators of knowledge that history is actively being created and interpreted. It's not just mining facts from the past and bringing them to the future. It's actual active interpretation and thinking. We need to further value student thinking and center student thinking, communicate to them that we care what they think, not what the AI chatbot thinks about what they're looking at. It's incredibly important. And I would say, you know, the great question this morning in the great presentation by Marnie Hughes-Warrington about these AI companies and possibly taking advantage of the experience and insights that people who study history have, this is our best chance to reach the widest audience of people who are going to become the future AI tech gurus are sitting in 6th to 12th grade classrooms right now. We can make that active case to them that this kind of thinking is important. So that would be my sort of base pitch, but I'll turn it over to Kristin.

[11:56] So I'm going to talk a little bit about redesigning assignments and teaching with AI in a more critical stance. This question of how can we encourage students not to use AI to cognitively offload that work, but instead to go deeper. And so there are some categories here. The first one would be to just treat AI like other technologies and to have students take history of technology classes to get that into the AI certificate and degree programs to think about AI in terms that we think of things like industrialization, right? like what were the short and long-term impacts of that. An assignment that I've adapted from a group called the Civics of Technology is a techno-ethical audit of AI technologies. And I'm happy to share that with anyone who would like a copy of that.

[12:57] It's basically a jigsaw activity where students become experts on one area of auditing technology, whether it's labor, the environment, or intellectual property, things like that. Then they jigsaw and they teach each other about it. And then they come up with a list of costs and benefits of using AI technologies for specific kinds of tasks and decide whether it would be a good idea or a bad idea to use AI. I just think we need to do more of those. On my campus, there's so much polarization across campus between people who say hell know AI in any higher ed versus people who have jumped on board, you know, and two feet in. And I think like we can't even have those conversations with a lot of people because they don't know what they don't know about it. So this is kind of a base level thing that I wish we would all try to do with people in our communities and our campuses.

[13:59] Yeah. The second kind is to practice with peer written work and editing. So asking Gen AI to complete an assignment, like whatever your assignment is, put it into Gemini or ChatGPT and see what answer comes out. And then that gives you an idea of what that looks like, but also take it to your students and show them what it looks like. Have them edit, comment, and annotate it based on your rubric and see how it would score and tell them they have to be better, right? Because this is what a machine can do. What can you do to add to it, right? And that becomes an assignment itself. Another one would be to analyze generative AI output. One assignment I have, again, I can send a link to anyone who's interested, is to compare generative AI response to the Wikipedia article and then to your course readings on any

[15:01] subject. I do this in a U.S. Latin American relations class, and we do this in groups, and we query the LLM in Spanish and in English. We look at the native Spanish Wikipedia page versus the English Wikipedia page, and then we look at our course readings. So if you're reading something about like the Cuban revolution, right, in those different formats, it's so interesting to see how the story is told. And it's a really great way to turn that into a thinking exercise about how history is written. You can also analyze Gemini Notebook Studio outputs like the podcast, the slides, and the infographics, or Canva, or MidJourney exhibit mock-ups, or images that are generated. For research, I take my students through using things like Elicit or Research Rabbit. Sourcely is another new one to find primary sources.

[16:04] But I think actually Gemini's been getting pretty good at just helping students locate primary sources that they might not find through like WorldCat. Just the archives, what archival repositories might I need to look into? and what search terms would be helpful. And then our library catalog has an AI research assistant built into it. So teaching them how to use those tools that the library catalog has access to. And then, of course, we've heard a lot about some of the advanced work that people at this conference are doing. And so I think that's more like a step two graduate level. Like after you know how to do some of these basic things, you understand how some of these models work, then you can start putting students on some advanced assignments. Then just thinking about as a teacher, how do we use AI to plan? Is that an okay thing to do with my social studies student teachers?

[17:06] If they don't know how to write a pacing guide or a syllabus or a curriculum map because they're really not familiar enough with the content to know how much time they ought to spend on each. Like what's going to be better, getting Gemini to help them with that or Khanmigo or using the textbook table of contents if they just really don't have that knowledge, right? So that's an interesting dilemma for me. I haven't figured it out, but I'm grappling with that. Also teaching, where do we stand on creating teaching materials with generative AI, making quizzes, rubrics, readings, outlines, graphics, Gamma for our slides, and where should our K-12 teachers be with those things. I had a student teacher last year who was teaching Watergate. She used Gamma to generate her slides. She was not supposed to, but she did, and literally her illustrations for Watergate were like AI generated images and the tapes

[18:09] are on YouTube, right? Like you can listen to Nixon talk about what he's going to do, but she had Gamma, right? So, you know, just helping to walk students through those kinds of choices and thinking about them ourselves. Course chatbots, are we building chatbots to help students interact or with our syllabi when we're not around to help with that. And then all the things that we don't like to do as faculty or grant administrators, like program assessment, core assessment, writing the strategic plan, where do we think we are with that in terms of getting AI assistance with that? I think the bottom line for me is, and then of course, student feedback on student work. A lot of debate over whether we should be, anonymizing student work and then feeding it into the machine to get help with feedback if you've got 120 students and you're trying to have them write essays.

[19:10] Instead, I teach students how to do that themselves by putting my assignment instructions and the rubric and their work into the Gemini and then getting their own feedback and then using that to improve their work before I even see it and give them additional feedback. And then transparency and disclosure. I think this is important for us as teachers too. Like, did we use Diffit to create the reading for the second grade student or the sixth grade student? You know, did we use Gemini to help us put together a reading list for the syllabus for this course? Did we use a large language model to help us give feedback on student work? Students are asking for that kind of transparency, and they get kind of angry at the thought that they might be paying a human to teach the course and then getting a machine doing some of the scoring or the

[20:13] feedback or the creation of the materials. So I think we do need to be clear about that. Wow, these are great questions. Thank you. I'm going to stand up just because I'm right in the middle of my slides here. So I think just tracking what we've heard so far in terms of these issues around transparency and how we may, you know, be responsible in the ways in which we're also preparing our students to use these tools. You probably heard about a story that went viral about a history faculty member at a university who ended up failing 32 of 35 students because they used AI to complete a midterm exam. He had posed the question to them to write an essay about the sort of the difference or the comparison of the industrial revolution to today's digital age. And his point, he said, was to encourage these students to really to read

[21:13] the output and to the output that they're submitting. In some ways, though, I think this is also this kind of gotcha pedagogy that I'm wondering if we may want to get away from and think differently about the way in which we teach. And so often what I think about is that perhaps the gap in AI or what AI is disrupting in education is really just highlighting gaps in our current approach to education. I was in a, if you were in this room previously, there was some mention about the sort of traditional coverage-based approaches to history education. And I think there's a real opportunity here for us to rethink that and finally do it with the tools that are available to us to actually bring this about. So we've been talking a long time in my field of learning sciences about how students learn history and trying to approximate that in the classroom. And I think we finally have the tools that can make that personalized learning available. So one of the projects I'm working on, similar to what we see in other

[22:15] fields like math and science, where they've mapped a learning trajectory around conceptual knowledge, is attempting to do the same in historical thinking. So this is a URL to a website that has the complete project. It was funded by the Library of Congress Teaching Primary Sources program. And I'll just tell you a little bit about what I mean by a learning trajectory. So it's identifying, getting really clear with the learning goal. And I think both of what we heard here today from my colleagues is the importance of first really surfacing what our learning goals are and getting really clear and really transparent with students about what they are. And then thinking about the conjectured pathway or the progression that students need to move through in order to be successful, to gain proficiency. And that, of course, are those instructional activities that come along with it. So what we've done is identified a series of eight learning indicators based on

[23:16] the research literature on historical thinking from the National Research Council. And then what we did is we tested these in the classroom. So we, for each learning indicator, identified proficiency levels from beginning, developing, proficient to exemplary, and then moving, trying to move students through these trajectories. And I think this is applicable in undergraduate and graduate education, just like secondary education. And this is kind of what the matrix looks like. So again, you can see the learning indicators like evidence, and then the idea of what a beginning level could do, and then developing proficient and exceeding. And the aim here is to layer in instructional strategies to move students. But the point of this is that our students come at different proficiency levels. And so what if we begin with a pre-assessment to see where they are and then give them a sort of personalized learning trajectory?

[24:17] It could be differentiated based on the complexity of the primary source or the source they're given, the type of work that they're given, and then move through so we can get them proficient so they can do that really deep historical thinking. And I think what we can do working alongside teachers who are doing this work in the classroom as well as AI tools is to create really personalized learning pathways. So this is where the project is currently. So rather than wait till the end, that project on, you know, what is the similarities or difference between the industrial revolution, for instance, and the digital age, but think about all those things, all those moves that students have to do to get to that final product and really try to support them through active scaffolding, personalized learning, and then also consistent assessment. So this is a sort of invitation to start thinking differently about the types of

[25:18] things we can bring in the classroom, instead of just layering AI onto what we already do, to think about new things and new opportunities that we can leverage. So with that in mind, we're going to move into our discussion, and we have a variety of discussion topics for you all, and then we will come back and share. So I'll turn this back over to you, Nate. so like we said at the beginning we've kind of divided this into three sort of main areas which we think are encapsulate absolutely every single thing that could possibly no impossible but these are kind of general areas in which i think people's interests tend to lie one is what do we do about students and using ai policy which includes do you allow students to use ai policy and do you let them opt out because increasingly students um a non-significant significant minority of students i think are very opposed to any use of ai in the classroom i tell people all the time the most anti-ai person i know in my life is my 19 year

[26:19] old very anti when i look up something on google she yells at me for trying to read the first line of like that's the i suggest it don't even look at that um so there's going to be more and more of that i think among students and then what are we doing in terms of assignments that engage students with AI in a form of critique and form of AI literacy and digital literacy. We have some suggestions up there. What assignments have you built that require students to engage with AI? How does your program build AI awareness and ethical issues into assignments? Important things to think about. And then instructor use is also something that we talked about. switching it up okay can you hear me yeah okay uh and then also instructor use like how do you uh i thought it was me so i'm glad it wasn't it was the technology how how can we use it in planning right and how much should we be transparent with our students it you know we wouldn't tell our students that we

[27:23] googled something but we might may feel more uh pressed to tell them that we used an AI tool to organize something, to summarize something, however you might use it? Does your program have systematic AI approaches? Is it already being incorporated? We talked about the Gemini, right, already being incorporated into Google tools and already being right there for students and instructors to use. And then what should, if any, the role of AI in assessment be? Does it have a role? Should we keep it out of assessment completely? Or are there possibilities there. So pick your topic and we will reconvene in how many minutes do you think? 10 minutes or so. And we'll kind of share conclusions, go around and have people share. And it's your chance to ask us questions as well. And we hope that the folks online, I see there's some good questions going in Discord and we hope that you'll discuss it online as well. Yes, please.

[46:33] ...the point of class, and I'm still holding them responsible and accountable for the content, But because they can get information so quickly and in so many different ways, I'm finding myself becoming less concerned about the content piece and more about the active thinking piece. I think that's great. The educational researcher Brenda Santos says this thing about trying to ask students figuring out questions versus finding out questions. the way I kind of interpret that is try to ask students questions that you wrestle with and not one that you already know the answer to. And you're just waiting because nothing deadens a discussion like, and who did this? And you know, I know, and you're like, do you don't want to say it? So, but if, if it's one that you find difficult too, it's really compelling, which I think also gets the Meghan's point about getting away from coverage and getting into the deep dives. So yeah, I think that's really important.

[47:35] Other thoughts that people raised. So the Discord chat has been really interesting with several different issues. One that's come up is the fact that many institutions do not have AI policies yet, not even the requirement that an instructor have an AI statement on their syllabus, which is an absolute bare minimum, because students are just left clueless otherwise. And then also having an AI statement that's enforceable, right? If you can't enforce it, then you like, it's not really good idea to put it on there, right? And then the issue of adjunct and contingent faculty, and how policies might differ in different institutions, and how that can really make it difficult for adjunct and contingent faculty to design courses that could work at multiple institutions. So that's something to that

[48:36] I think we could continue to talk about. And then the other thing that you all just said, this tension between content and skills is something that always comes up in our profession. And I think for a while, our social studies standards with the C3 framework at the K-12 level shifted towards skills, like skill, skill, skills, really de-emphasizing content. But then we've got this problem because if you don't understand the content, then you don't have the context you need to do the analysis of the documents. So we do still have to keep a balance between the content and the analytical skills. And we can't just do it on a need to know basis. I mean, they do need to be grounded in that historical knowledge. And perhaps that's a place where some of these other kinds of interactive platforms that students can, you know, interact with in a more gamified way to learn some of the content. I don't know if that's an appropriate place for that or not, but I know there are people that are trying that. That's lots of cool stuff in the chat.

[49:40] That's great. Thank you for flagging that. Other thoughts? Well, I think we have our contact information on here. We would be super happy to continue the conversations or share some of the resources that we use for assignments, the projects that we're working on at our institutions and continue the conversation. For sure. And I think we solved all the problems having to do with AI and history instruction. So give yourselves a round of applause. It was tough, but you did it. Good job. But thank you very much for your time. If you have any other questions for us, we're happy to answer those as well.