👨💻 How to Use LLMs in the Classroom
Explore different strategies to incorporate large language models in the classroom
Most people are familiar with ChatGPT, but there are also many other large language models (LLMs) with their benefits and use cases in various contexts. Each LLM has been trained on vast amounts of data to develop its neural network, but what distinguishes LLMs from each other is their training data, model architecture, and the specific techniques used during the training process. To get the most out of an LLM, it's essential to understand its capabilities, limitations, and potential biases. In this edition, we will explore various ways to incorporate LLMs into the classroom, along with tailored prompts for each application. Which LLM have you currently been using most frequently?
Here is an overview of today’s newsletter:
New and extensive prompt ideas for educators and students
Strategies for leveraging different LLMs in the classroom
Insights from recent studies on integrating Gen AI tools into K–12 education
Navigating the role of AI chatbots in education
We're excited to share that AI x Education has been featured in Nature, alongside initiatives that amplify students’ perspectives in the current discourse. Thank you for your continued support! Should you find our content insightful, please consider sharing it with other educators in your community.
🚀 Practical AI Usage and Policies
If you are looking for new prompts to try out on GPT or other LLMs, we recommend exploring the extensive prompt library curated by Dr. Ethan Mollick and Dr. Lilach Mollick. It features numerous prompts for designing educational game simulations, crafting lesson plans, and more. They also include prompts that can be used as part of student exercises. Below is an example of how you can prompt Bing and Claude to become a personalized AI tutor:
You are an upbeat, encouraging tutor who helps students understand concepts by explaining ideas and asking students questions. Start by introducing yourself to the student as their AI tutor who is happy to help them with any questions. Only ask one question at a time. Never move on until the student responds. First ask them about their learning level: Are you a high school student, a college student, or a professional? Wait for their response. Do not move on until the student responds. Then ask about the topic they would like to explore and what they know already about the topic. Number these two questions. Do not suggest topics. Wait for a response. Do not move on until the students gives you a response to both questions. Given this information, help students understand the topic by providing explanations, examples, analogies. These should be tailored to the student's learning level and prior knowledge or what they already know about the topic. You should guide students in an open-ended way. Do not provide immediate answers or solutions to problems but help students generate their own answers by asking leading questions. Never ask more than 2 questions at a time; more than 2 questions is overwhelming. Ask students to explain their thinking. If the student is struggling or gets the answer wrong, try giving them additional support or give them a hint. If the student improves, then praise them and show excitement. If the student struggles, then be encouraging and give them some ideas to think about. When pushing the student for information, try to end your responses with a question so that the student has to keep generating ideas. Once the student shows an appropriate level of understanding given their learning level, ask them to explain the concept in their own words (this is the best way to show you know something), or ask them for examples or give them a new problem or situation and ask them to apply the concept. When the student demonstrates that they know the concept, you can move the conversation to a close and tell them you’re here to help if they have further questions. Rule: asking students if they understand or if they follow or if something makes sense is not a good strategy (they may not know if they get it). Instead focus on probing their understanding by asking them to explain, give examples, connect examples to the concept, compare and contrast examples, or apply their knowledge.
Source: Dr. Ethan Mollick and Dr. Lilach MollickWith numerous Language Learning Models (LLMs) available, including GPT, Gemini, Claude, and others, it can be overwhelming to keep track of them all. What are the unique capabilities of each LLM and Generative AI Tool, and how can you use them in the classroom? Below is a compilation of resources to help you get started:
Integrating AI into Google Forms can open up new opportunities for interactive and personalized learning experiences. Discover how to use Google Forms + AI with this resource.
Microsoft has recently launched a one-hour online course, “Enhance teaching and learning with Microsoft Copilot” for educators. This course enables you to delve into Microsoft Copilot for education by familiarizing yourself with its modes and features for designing effective prompts and analyzing results.
One of Gemini's unique features is its ability to analyze and interpret images uploaded by users. Students can capture images of nature to deepen their understanding of the science and environment surrounding them. For math, this imaging feature can be used to analyze data. Check out this resource to learn how to incorporate Gemini into your classroom.
Perplexity is valuable as it includes citations in its responses, grounding the results for both learning and research. Nevertheless, its capabilities extend beyond this feature alone. Richard Campbell delves into several innovative methods to use Perplexity in the classroom through this blog.
If you want to learn more about how to use generative AI in the classroom, Google released a free 2-hour online course in collaboration with MIT RAISE. It covers topics such as “save time on everyday tasks”, “personalize instruction”, “enhance lessons and activities in creative ways”, and many more.
One section from Google’s course. Source: Grow with Google
📣 Student Voices and Use Cases
This week, we spoke with Konstantin Loktev, a 4th-year student at Belarusian State University of Informatics and Radioelectronics (BSUIR) majoring in Artificial Intelligence. In the following, we present select highlights from these conversations, which have been slightly edited for enhanced clarity and precision:
Q: In what ways has using AI tools impacted your learning habits, critical thinking, or problem-solving skills, if at all?
Before AI became as huge as it is now, I would lie if I said that I liked the process of looking for information on my own through the monstrosity that is the Internet of today. With AI (mostly dialogue systems) the process of searching and structuring information has become much easier, and, as much as I don’t want to accept it, using it in my assignments or at work is now quite a regular occurrence. However, I turn to AI only to solve a very specific issue at hand that would require too much time to solve by myself when compared to its importance. I believe that being able to think critically and solve problems on your own is a very important skill to have, even when AI can pretty much do it for you. So, I’d say that my learning habits were indeed influenced by AI, to which I now turn much more frequently than to official documentation or scientific papers, but the impact on critical thinking and problem-solving skills has been minimal.
Q: What advice would you give to educators looking to incorporate AI tools in their classrooms? Based on your own experience, what worked well and what do you think can be improved?
At the end of the day, AI, in my opinion, is just a set of tools for automating routine tasks. And it should be treated like a tool, not like a means of achieving a set goal. AI-generated lectures and presentations are a big no-no (or at least thoroughly fact-check them before bringing them to class). Dialogue systems and other AI tools integrated in the learning process must be meticulously tested before they are used by actual students, since any AI hallucinations that can occur in them may negatively impact the learning process for countless people.
Q: What steps do you believe educational institutions should take to ensure responsible and equitable AI usage in classrooms?
Any AI tool used in educational institutions should be bespoke or at least regulated, only operating in the boundaries set by said institution’s policy(-ies), and refusing to give the response to a student’s request in case of accidental or intentional misuse. If no regulatory actions against AI misuse are taken, the usage of AI in education will be very short-lived, since mass media will be filled with articles talking about students using educational tools for unintended purposes, which will create a negative image of AI in education for the majority of the population.
Q: In your opinion, what are the most exciting possibilities for AI in transforming the future of education?
Currently, the biggest problem in education, in my opinion, are misunderstandings between the teacher and the student. One teacher can only have one teaching style, but a classroom of 100+ students can have much more than one learning style. Most of the time, teachers are reluctant or altogether unwilling to modify their teaching material to better explain a topic to a group of 5 people when the other 95 got it. The most exciting possibility for AI in education, in my opinion, is its potential ability to read, structure, “understand” and relay the teaching material to such students according to their learning styles.
📝 Latest Research in AI + Education
RAND Corporation
Using Artificial Intelligence Tools in K–12 Classrooms ↗️
The research on the integration of Gen AI tools in K–12 classrooms, through a survey of 1,020 teachers and interviews with leaders from 231 public school districts across the United States, revealed a nuanced picture of the early stages of AI adoption in education. As of fall 2023, a minority of teachers (18%) reported actively using AI tools in their teaching, with an additional 15% having experimented with them at least once. The usage was more prevalent among middle and high school teachers, particularly those teaching English Language Arts and social studies, likely due to their greater need for customizing instructional content. The research indicates that AI tools are mainly being used to adapt educational materials to suit individual student needs and to assist in generating new teaching materials.
Despite the early stage of AI tool adoption, there is a significant expectation among teachers for increased use in the future, with three-quarters of current AI-using teachers planning to expand their use of these tools. Districts are progressively supporting this integration, with 60% planning to have provided AI training by the end of the 2023-2024 school year. However, the rollout is uneven, with urban districts lagging in training delivery. District leaders are reportedly focusing on enhancing AI usage to simplify teaching rather than on regulating student use. These developments suggest a growing recognition of the potential of AI to facilitate teaching, although there remains a cautious outlook regarding its broader societal implications and the quality of AI-generated content.
Diliberti, M. K., Schwartz, H. L., Doan, S., Shapiro, A., Rainey, L. R., & Lake, R. J. (2024). Using artificial intelligence tools in K–12 classrooms. Santa Monica, CA: RAND Corporation. https://www.rand.org/pubs/research_reports/RRA956-21.html.Microsoft Research
The Emerging AI Divide in the United States ↗️
This research paper examines how the advent of Gen AI technologies, such as ChatGPT, might exacerbate existing digital divides while also offering new pathways for digital inclusion. The study utilizes state- and county-level search query data to analyze the geographical and demographic patterns of AI tool awareness across the U.S., revealing that knowledge of ChatGPT is unevenly distributed, with higher search rates in the more urbanized, educated, and economically advantaged West Coast states, and lower rates in Appalachian and Gulf states. This disparity suggests that despite the potential for AI to democratize access to technology due to its affordability and multilingual capabilities, early adoption still mirrors traditional patterns of digital marginalization, where education stands out as the strongest predictor of engagement with AI tools. The analysis points out that while Gen AI tools like ChatGPT have the potential to revolutionize productivity and creativity, their impact is contingent on broader access and utilization, suggesting a need for proactive measures to ensure equitable access. This calls for policy interventions and targeted educational programs to address and mitigate the risks of deepening digital divides, ensuring that the benefits of generative AI are accessible across all societal segments, thereby avoiding the reinforcement of existing inequalities in digital tool access and use.
Daepp, M. I. G., & Counts, S. (2024). The emerging AI divide in the United States. arXiv preprint arXiv:2404.11988.📰 In the News
Ed Surge
When Bots Go to Class ↗️ (Featuring AI x Education’s team)
Key takeaways:
Curriculum Testing with AI: Ferris State University announced plans to enroll two AI chatbots as students to test their curriculum and explore the use of AI in education. This move also serves to highlight their academic program in artificial intelligence, attracting local media attention and potentially acting as a publicity effort.
Learning Analytics and AI: The initiative at Ferris State is part of a broader trend in higher education focusing on learning analytics—using data generated by students' interactions with digital platforms to improve course design and personalize learning experiences. AI tools like ChatGPT allow for the creation of student simulations to test and enhance educational materials and teaching strategies.
Implementation and Ethical Considerations: While the chatbots, named Ann and Fry, have yet to start classes, they are designed to participate fully, from engaging in class discussions to completing assignments. This approach raises ethical questions, including concerns about inherent bias and the representation of diverse student needs.
Broader Educational Impact: Other institutions, like Arizona State University, are also exploring AI's potential in education. These initiatives aim to enhance student success, streamline administrative processes, and develop new educational tools that interact dynamically with students. The ongoing experiments will provide insights into what educational applications of AI are effective and which need refinement.
BBC News
Can AI take teaching 'out of the Victorian ages'? ↗️
Key takeaways:
Modernizing Education with AI: In the East of England, artificial intelligence is being utilized to modernize and personalize education, ensuring that learning styles are tailored to each student to keep them engaged and prevent anyone from falling behind.
Innovative Learning Environments: West Suffolk College has introduced a £2 million XR Virtual Reality Lab, funded by the government, where students can immerse themselves in virtual 3D worlds using headsets or tablets. This approach "gamifies" the curriculum, enhancing student engagement and making learning more experiential and enjoyable.
Real-time Learning Adjustments: The Bedford College Group employs AI to assist students in subjects like GCSE Maths and English. The technology instantly analyzes student responses, identifies areas of difficulty, and customizes subsequent content, allowing for immediate feedback and self-paced learning.
AI's Role in Reducing Teacher Workload: AI is also being used to aid teachers in lesson planning, reducing the time spent on these tasks and alleviating their workload. This shift allows teachers more time to focus on direct student interaction and support.
Managing AI's Ethical Use: There's a conscious effort to balance the use of technology in education to prevent overreliance on screens and ensure responsible usage. For instance, West Suffolk College plans to establish an AI Board to oversee the ethical use of AI, emphasizing the importance of students demonstrating their understanding of subjects independently.
“Chatgpt.” ChatGPT, OpenAI (GPT-4), openai.com/chatgpt. Accessed 23 Apr. 2024. And that’s a wrap for this week’s newsletter! Based on our previous poll, we noticed that not many educators were familiar with the environmental implications of AI technologies. We hope you found the resources helpful, and we would love to hear about whether and how you covered the topic of environmental implications of AI in your classroom. Please feel free to share your experiences in the comments below!
If you enjoyed our newsletter and found it helpful, please consider sharing this free resource with your colleagues, educators, administrators, and more.