Charting the Course: Incorporating AI Into Assignments to Foster Self-Regulation
CoT Announcements
2023 Annual Conference: The 2nd Annual Civics of Technology annual conference will be held online from 10-3pm EST on both August 3rd and 4th, 2023! You can learn more, register for the conference, or submit a proposal on the 2023 conference page.
Next “Talking Tech” Monthly Meeting on 06/05: We hold a “Talking Tech” event on the first Tuesday of every month from 8-9pm EST/7-8pm CST/6-7pm MST/5-6pm PST. These Zoom meetings include discussions of current events, new books or articles, and more. Participants can bring topics, articles, or ideas to discuss. Register on our Events page or click here.
Next Book Club on 07/27/23: We are reading Meredith Broussard’s 2023 book, More than a Glitch: Confronting Race, Gender, and Ability Bias in Tech. We will meet to discuss the book at 8pm EDT on this Thursday, July27th, 2023. You can register to join on our Events page.
by Brett D. Jones and David Hicks
The release of ChatGPT has provoked a great deal of attention and rung alarm bells across many spheres of teachers’ professional and personal lives. ChatGPT joins a growing band of Artificial Intelligence (AI) tools that have left teachers and teacher educators wondering how it could be used for good, ill, and/or somewhere in between. A recent survey from the Walton Family Foundation found that 51% of K-12 teachers reported using ChatGPT for tasks such as lesson planning, identifying creative ideas for classes, and building background knowledge for lessons and classes. Interestingly, 73% of teachers surveyed reported that ChatGPT can help their students learn more, and 68% of students ages 12 to 17 replied that it can help them become better students.
Although many teachers and students see the potential benefits of using AI, it is unclear as to what relationship they should have with AI. Perhaps AI is simply a tool or resource that should be used much like a calculator to help students solve problems more quickly and efficiently. In this case, there doesn’t seem to be any harm in allowing students to use it to be more productive. Using a calculator, however, eliminates the need for students to learn basic arithmetic skills, which could be useful in the absence of a calculator, or for other reasons such as obtaining mathematical literacy. For these reasons, most curricula still require students to learn basic computational skills even though students carry calculators around with them every day on their phones.
The nature of the working relationship with AI gets more complicated when we consider what skills may be lost by students who use AI. Instead of using a calculator for computations, consider students who are taught to use an abacus instead. Those students are able to not only solve arithmetic operations more quickly using an abacus than a calculator, but they also have the potential to improve some of their other cognitive abilities. Some studies have shown that students who are taught abacus-based mental calculation can improve their working memory and numerical magnitude processing, and that using an abacus can produce functional and structural changes in the brain. Therefore, although students who use the calculator may obtain the same answer to a mathematical problem as those who use the abacus, they have lost out on the opportunity to gain cognitive skills and develop their brain by not learning how to use an abacus. The same concern may also exist for having students use AI to analyze historical sources rather than learning to use sourcing protocols themselves. Our point is that we must not only consider the speed and efficiency with which students achieve solutions using technologies, but we must also consider what they may be giving up by not engaging in the activities that the technologies have allowed them to circumvent. These are, however, not necessarily new concerns brought on by the emergence of AI; Nicholas Carr raised similar concerns when he asked if “Google is making us stupid”(er) back in 2010 and illuminated the dangers of “putting our knowledge in the hands of machines.”
It will likely be difficult for teachers and curriculum developers to decide which skills can be replaced by AI and which skills still need to be taught because they are fundamental to becoming more intelligent and informed citizen and/or to the nature of the discipline being studied. Regardless of how these decisions are made, we believe that AI has the potential to be used as a interactive tool/resource that can facilitate, support, assist, and augment students’ learning. Used in this manner, we suggest it can serve as a means to provide scaffolding to help students “learn how to learn” in ways consistent with cognitive science because effective learners are good at self-regulating their learning. That is, they plan before they begin a learning activity (the planning stage), they monitor their learning during the activity (the monitoring stage), and they reflect on their learning after the activity (the reflection stage).
Let’s consider a couple examples of how teachers can use AI as an interactive tool/resource to help students become more effective self-regulating learners. English teacher Cherie Shields had her students use ChatGPT to create outlines for their essays before they began writing their essays. Similarly, Steven Mintz encouraged his students to develop and input a detailed prompt into ChatGPT to produce an essay. Students then revised the essay and found additional research to support the main points. In both cases, ChatGPT was used as an initiating tool to begin the writing process when faced with a blank page. We imagine these assignments could be adapted for other types of courses, such as mathematics or computer programming courses. These activities seem to be a worthwhile use of AI and could help students to develop specific skills and sharpen their thinking within an academic discipline.
With a few modifications, these activities could be expanded to also teach students how to become better self-regulated learners. First, students need to be aware that the assignment was designed to integrate AI in a way that would help them achieve a goal more effectively. Explicitly explaining and then modeling how the AI can be used as a collaborator to reach a goal is part of the planning stage of self-regulation. In this stage, students plan their time, effort, and strategies to achieve their goals. For the writing examples noted previously, the teacher could explicitly communicate that the purpose of the assignment is to write an essay and that they will use ChatGPT as a tool to get started. Teachers can show students that writing good prompts is a skill they need to learn because prompts can significantly affect the output. For example, teachers could have students input prompts with different levels of specificity to compare the types of output or genres of writing produced. Students could then determine which prompts lead to outputs that are most aligned with their goals. It would also be worth sharing with students current advice on how to cite text produced by AI. Another possibility is to ask students to brainstorm ideas about how they could use AI to write a quality essay. Giving students control and choices can be an effective way to empower students to engage in the process and can help them to consider other ways of planning to reach their goal.
The prompt that students enter into ChatGPT can make a difference in the output that students receive. Within the field of social studies, we are currently playing with creating model prompts to serve as scaffolds to remind and reinforce disciplinary ways of thinking when students chat with ChatGPT. For example, we entered the following prompt into ChatGPT: “In what ways can a student use you as a tool or resource to analyze historical sources as part of an inquiry? What types of historical sources do you work best with?” ChatGPT provided the following output to this prompt:
Using AI, like ChatGPT, as part of historical inquiry can be a valuable way to support students' analytical skills. Here's how students can use me to analyze historical sources:
1. **Contextual Understanding:** When students come across a historical source, they can ask me about the broader historical context of that time. Understanding the background can often provide valuable insights into the source material. For example, if they're analyzing a speech by Winston Churchill during World War II, they could ask about the geopolitical climate of that time, the status of the war when the speech was given, etc.
2. **Clarification:** If a source contains complex language or references to specific events or figures that students are unfamiliar with, they can ask me for clarification. This can help them better understand the source and its significance.
3. **Comparing Sources:** Students can use me to find additional information that can serve as a comparison or contrast to their primary source. For example, if they're analyzing a primary source on the public's perception of the Vietnam War in the U.S., they could use me to find other perspectives on the same topic, perhaps from a different country or demographic.
4. **Critiquing Sources:** I can aid in critiquing the credibility and potential bias of sources. Students can ask questions about the author, publication, or general reputation of the source to determine its reliability.
Regarding types of historical sources, as an AI text model, I work best with textual or written sources. These include speeches, letters, newspaper articles, journal entries, books, or any other text-based document. While I can provide information about the context of images, artworks, or non-written artifacts, my ability to analyze these directly is limited since I don't have visual capabilities.
Remember, while I can provide a lot of information and context, students should always critically evaluate and corroborate the information I provide with other scholarly resources. It's also essential for students to develop their own skills in analyzing and interpreting historical sources, with AI serving as a supplementary tool rather than a replacement for their own critical thinking. (OpenAI, 2023)
We contend that this response could provide the student with a reminder of what they were learning in class via a focused checklist of the necessary steps for undertaking the complex process of historical source analysis.
During the second stage of self-regulation (monitoring learning), teachers can remind students to check their progress in relation to the assignment goal and make adjustments as needed. Instructors can help students in this stage by ensuring that the assignment instructions are clear and that the evaluation criteria (e.g., rubric) is explicit and understandable to students. In this stage, instructors may provide students with some possible strategies, or they may suggest that students use AI to brainstorm ideas if they get stuck during some aspect of their writing or problem solving. The monitoring stage is a good time to remind students that AI is still limited in many ways and that they need to become critical readers of AI output by checking its veracity, depth, accuracy, and humanness. This requires comparing AI output to other evidence, claims, and products. Through this process, students will become more aware of the errors that AI makes and realize the limitations of AI. Teachers could also have students annotate their work to show their thought processes as they edit the essay.
The third stage of self-regulated learning (the reflection stage) involves students reflecting on their learning after the activity. Teachers could add to the essay assignments described previously by incorporating a reflection component that asks students to reflect on what they’ve learned. Some possible questions that students could discuss and answer, as part of a written or video reflection, include:
· How could you have improved your essay to meet the goals of the assignment more completely?
· To what extent was the AI helpful in providing you with support during the assignment?
· How could you have used AI more effectively?
· Where did AI let you down or limit your own ideas, creativity, and processes?
The reflection should be included as part of the assignment grade and listed in the assignment instructions or rubric. Instructors should also explicitly state that part of becoming an effective self-regulated learner is to reflect on their progress. The instructor might assign a follow-up assignment that builds on this one to incorporate some student ideas that were identified during the reflections.
Of course, the ideas suggested here are not unique to assignments that integrate AI, they have also been applied to assignments that do not require the use of AI. In this way, AI is not necessarily obviating the need for these skills or replacing them. However, as instructors reconceptualize some of their assignments to incorporate AI, it appears to be an opportune time to also more explicitly help students think about their learning and how they can use AI as a interactive tool to become a more effective learner. We would encourage instructors to be intentional, explicit, and advertise exactly how they are using ChatGPT as part of the learning and self-regulation process. Doing so will likely help students to transfer their skills to other activities and disciplines, which is an important learning outcome in itself.
Yet, we wonder how incorporating AI into teaching self-regulation processes may take away from some students learning other skills, much as the calculator may take away the need for some arithmetic skills. Is there something to be gained by staring at a blank screen and deciding how to start an outline of a paper instead of having AI write the outline? Certainly, some cognitive skills are not being used or developed when AI is used to replace that task. The fundamental question seems to be whether the benefits of achieving a better final product in the same amount of time are worth the costs of what may have been lost in the process. By using AI as more of an interactive tool to scaffold self-regulation, perhaps AI can foster students’ development instead of simply substituting for skills that the student will not learn because AI can do them more effectively. That is not to say that AI shouldn’t be used in this manner. However, if one of the functions of education is to develop and hone intellectual skills, then AI should not be used simply to replace intellectual skills, it should help to develop them.
As teachers continue their journey with AI and navigate new terrain, their initial map making endeavors into the uses of AI to support learning—like the first map makers’ representations and sketches themselves—may be somewhat simplistic, flat, and/or mistaken. However, these initial ideas can be tested and eventually serve as a basis for more nuanced representations and explicit approaches to use AI to help students become better at learning. Our guarded idealism and willingness to navigate this new terrain, however, does somewhat conjure up Hector Barbossa’s warning to Captain Jack Sparrow in Pirates of the Caribbean, “ You’re off the edge of the map, mate. Here there [may] be monsters.”
Author Bios
Brett D. Jones, Ph.D., is a Professor in the Educational Psychology Program in the School of Education at Virginia Tech. Dr. Jones researches student motivation and identifies strategies that teachers can use to support students’ motivation, such as by using the MUSIC Model of Motivation (www.theMUSICmodel.com).
David Hicks, Ph.D., is a Professor of Curriculum and Instruction in the School of Education at Virginia Tech. Dr. Hicks’ principal areas of research are related to learning history and social science, explicit strategy instruction and scaffolding, mixed reality and inquiry-based learning, digital technologies and multimedia, and disability studies.