📄 Rethinking Assignments with AI
AI is reshaping education, but will it lead to more meaningful learning or just smarter ways to cheat?
What if homework could be more than just a checklist? For many students, tackling assignments often feels like a tedious task, with some resorting to copying answers directly from the web. Now, with generative AI, educators are increasingly concerned that students will use these AI tools to obtain answers even more easily. On the other hand, generative AI also offers teachers the opportunity to create unique questions. However, this may lead to a scenario where students might then use AI to find solutions to these new questions, potentially resulting in a never-ending cycle of academic whack-a-mole.
Looking ahead, will assignments become an endless parade of AI-generated problems? Or could they evolve into dynamic tools that simulate active learning? AI has the potential to transform these homework assignments into opportunities for genuine engagement, not just a race to get the right answer. In this issue, we’ll explore how to harness AI's full potential to make education less about busy work and more about meaningful learning experiences.
Here is an overview of today’s newsletter:
Harvard professor reveals innovative use of chatbots in course curriculum
Innovative ways students are using AI to boost their learning
High school student shares how his school’s approach to AI has evolved this year
Impact of AI in helping students sharpen their metacognitive skills
🚀 Practical AI Usage and Policies
Professor Michael Brenner teaches Applied Math 201 at Harvard University, a course designed to equip first-year graduate students with mathematical methods for solving complex science problems. This past semester, he introduced an innovative approach for the class's final project: students were tasked with teaching chatbots to solve problems using prompt engineering. The students had to work in groups to create a dataset of prompts to assess effectiveness and generate new math problems and solutions for evaluation. This novel approach not only deepened students’ engagement with the material but also honed their problem-solving skills, opening doors for continued exploration in pedagogy. Watch the video below to see how it went!
Microsoft’s recent article, “Smart Ways Students are Using AI”, explores how students are harnessing AI to enhance their learning experiences. Here are a few ways students across different educational levels are incorporating AI into their coursework:
Using Copilot to condense lengthy reports into key points, helping them quickly assess if the material is relevant to their study topic and worth reading in full
Learning the lingo of unfamiliar subjects and industries
Revising essays to make them more concise and precise
This blog by Google covers 5 ways Gemini can help students study smarter and shares how their new Gems feature, such as the Learning Coach Gems, can act as custom AI experts to help students create a tailored learning plan. Check out the video below to see how the Learning Coach Gems work!
Professor Ethan Mollick from the University of Pennsylvania offers a compelling read in his Substack, "Post-apocalyptic education: What comes after the Homework Apocalypse." He argues that while the Homework Apocalypse has arrived, two illusions persist: teachers believe they can still easily detect AI use, and students may not fully recognize how relying on AI tools can undermine their learning. His analysis provides a thought-provoking look at the evolving landscape of AI in education.
Tim Evans, Learning Technologies & Data Coordinator at the American International School of Guangzhou, China, created and shared this helpful infographic on LinkedIn, simplifying the UNESCO AI Competency Framework for Teachers and Students with visuals created using Perplexity and Canva. The 10 Key Points for Teachers are shown below and the infographic for students can be found here.
Source: Tim Evans // LinkedIn
📣 Student Voices and Use Cases
This week, we interviewed Yash Yardi, a high school student at the Illinois Math and Science Academy, to discuss the changes he’s noticed in how AI is being used at his school from last year to this new school year.
Q: In what ways have you noticed changes in how your teachers or professors are using AI to support instruction, assignments, or assessments this year compared to last year?
At my school, the most common statement for AI use is as follows in every class outside of humanities: “You are encouraged to explore AI and its capabilities, but ultimately you will be held responsible for the material in this class.” In my opinion, the advancement since last year at my school in AI policy is likely a result of my school’s background–IMSA is a STEM magnet high school–and the launch of the AI Center at IMSA, where research and experimenting with AI is showcased and conducted. At this point, the humanities–notably English, Foreign Language, and History–are the only departments where AI is discouraged, as its usage is commonly for summarizing, note-taking, and “essay-writing guidance.” I say that with quotes since there’s a good and bad side with which all educators are familiar with by now; however, the former are solid options that are permitted by a few of my teachers.
Q: With the growing presence of AI in education, have you noticed any concerns among students or teachers about academic integrity or data privacy?
My school often releases data regarding academic violations for the public, and to start with, there is a surge in plagiarism statistics. This is primarily due to AI being placed under this category in disciplinary infractions at our school and the strict policies of the humanities department. Many of the reports have risen from English essays GPT’d and similar activity, and understandably have concerned a lot of our staff. However, this varies per teacher, as I personally know humanities teachers who have done their research and succeeded at combatting AI misuse in their classrooms. Data privacy is a much less discussed topic when it comes to AI use at my school and this may be due to the high school level, where most of the students’ information is linked or under the control of their parents, so they can trust their information will not be breached easily with their mistakes.
Q: Have AI tools helped you discover learning strategies or resources you wouldn't have used otherwise?
AI tools have most certainly opened doors for how I study and maintain productivity. For starters, I have a list of prompts on a document that I regularly use with ChatGPT (my personal choice, however, Claude is also exceptional with the same) when it comes to note-taking, test prep, and has saved me precious time in my senior year. When it comes to learning strategies, my personal favorite is generating flashcards with AI tools, since it is a pain to manually type out a set often, and with AI, I can generate and begin testing myself within minutes, making for a much more streamlined experience. Lastly, I have seen concepts of AI tools I am eager to try in the format of scroll-studying–where video content is generated and taught in short form reel format on new apps for studying.
Q: If you could make one suggestion for improving how AI is currently being used in your classroom, what would it be and why?
One suggestion I have is for AI to be thoroughly reviewed by all the departments at my school with the goal of creating detailed limitations on its use. Due to the significant gray area hovering over the humanities as well as the general encouragement in technical departments, AI isn’t as organized as it could be at my school. With a limitations outline, we can immediately see a drop in disciplinary infractions. I have imagined this outline to be similar to the DEI Statements in every syllabus teachers provide that help promote inclusion and greatly impact the quality of collaboration and social behavior at my school. With an AI outline in the syllabus, students will firmly know the dos and don’ts depending on the subject that should be followed.
📝 Latest Research in AI + Education
University of Michigan, University of California Berkeley
Backwards Planning with Generative AI: Case Study Evidence from US K12 Teachers ↗️
This paper explores how K12 teachers in the U.S. are integrating generative AI, such as ChatGPT, into their backward planning workflows. The study tracked 24 teachers over the 2023-2024 school year and found that most educators fell into three distinct groups: those seeking both input and output from generative AI, those using it only for outputs, and those who avoid it entirely. Teachers who used generative AI for input into their planning reported improved productivity, whereas those who focused solely on generating outputs saw limited productivity gains. The study also emphasizes the potential for generative AI to enhance teacher workflows when used for ideation and planning, while cautioning against relying too heavily on output generation alone, which may disconnect from key learning objectives.
Keppler, Samantha and Sinchaisri, Wichinpong and Snyder, Clare, Backwards Planning with Generative AI: Case Study Evidence from US K12 Teachers (August 13, 2024). Available at SSRN: https://ssrn.com/abstract=University of Toronto
AI Detection of Human Understanding in a Gen-AI Tutor ↗️
This paper explores how AI can detect and analyze physiological signals like facial expressions, blood flow, and vocal tones to assess a learner's mental state during the learning process. By identifying different phases of understanding, such as nascent understanding, confusion, and deep understanding, AI tools like the "Abel Tutor," can provide real-time, adaptive learning support. This system could enhance engagement and personalize education by addressing students' emotional and cognitive needs, promoting deeper learning and understanding.
Woodruff, E. AI Detection of Human Understanding in a Gen-AI Tutor. AI 2024, 5, 898-921. https://doi.org/10.3390/ai5020045📰 In the News
The Wall Street Journal
Colleges Race to Ready Students for the AI Workplace ↗️
Key takeaways:
Schools are revamping courses and adding AI-focused degrees at an unprecedented pace to equip students with skills relevant to the evolving job market
Job descriptions mentioning AI tools like ChatGPT have tripled, with roles in tech, marketing, art, and media increasingly requiring AI proficiency
Many students believe AI skills are crucial to securing future jobs, with universities offering programs like AI minors and business-AI hybrid majors to meet this demand
AI literacy is becoming as essential as typing was in the past, with employers expecting job seekers to effectively use tools like ChatGPT and understand AI’s limitations
Universities like Emory, Cornell, and USC are introducing interdisciplinary AI programs, with some schools requiring all undergraduates to engage with AI, highlighting its importance across various sectors
The Atlantic
The Metacognition Revolution ↗️
Key takeaways:
AI is increasingly being used in education to help students improve their metacognitive abilities, allowing them to reflect on and analyze their thinking processes
The emphasis in education is shifting from content mastery to understanding learning as a process, with AI tools enabling personalized learning and deeper cognitive reflection
By interacting with AI, students are pushed to break down their knowledge, fostering critical thinking and a healthier skepticism of AI outputs
Traditional assessments, like essays, are evolving to focus more on the learning process, where students’ interactions with AI and their revision histories reveal how they develop and refine their ideas
Educators see AI as a means to teach students "how to learn," preparing them for a future where metacognitive skills and adaptability are crucial in both personal and professional development
“Chatgpt.” ChatGPT, OpenAI (GPT-4), openai.com/chatgpt. Accessed 10 Sept. 2024. And that’s a wrap for this week’s newsletter! If you enjoyed our newsletter and found it helpful, please consider sharing this free resource with your colleagues, educators, administrators, and more.