The Future of Feedback: How Can AI Help Transform Feedback to Be More Engaging, Effective, and Scalable?
This meeting report synthesizes the perspectives of 50 interdisciplinary scholars to explore the promises, risks, and future research directions of using generative AI to transform feedback in digital learning environments into a more engaging, effective, and scalable practice.
Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer
Imagine a massive kitchen where 50 of the world's top chefs (educational researchers, computer scientists, and psychologists) gathered to solve a very specific problem: How do we feed students the right kind of "feedback" so they actually learn, without burning out the cooks (teachers) or giving the diners (students) indigestion?
This paper is a "meeting report" from that gathering. It explores how Generative AI (the same tech behind tools like ChatGPT) can revolutionize how we give feedback in schools.
Here is the breakdown of their conversation, translated into everyday language with some tasty analogies.
1. The Big Promise: The "Thermomix" of Learning
For years, teachers have been like overworked chefs trying to hand-make a unique, perfect meal for every single student. It's impossible to do this for 30 students, let alone 30,000.
The group sees AI as a super-powered kitchen appliance (they call it a "Thermomix"). Just as a Thermomix can chop, stir, bake, and cook all at once, AI can:
- Grading: Instantly grade essays and math problems.
- Personalization: Serve a "spicy" explanation to a student who needs a challenge, and a "mild" one to a student who is confused.
- Scale: Feed millions of students at once without the teacher losing their voice.
The Catch: Just because you can use the machine doesn't mean you should let it cook the whole meal. The teachers (chefs) still need to taste the food and decide the menu. If we let the AI do everything, students might stop learning how to cook for themselves (a concept called "cognitive offloading").
2. The Problem of "Ignoring the Plate"
The researchers realized that even the best meal is useless if the student throws it in the trash. This is the issue of Feedback Engagement.
- The Reality: Students often ignore feedback. They might see a red "X" on a paper, feel bad, and move on without reading the note on how to fix it.
- The AI Opportunity: AI can act like a smart waiter. Instead of just dropping a plate on the table, the waiter (AI) can watch how the student reacts.
- Is the student staring at the screen? The AI can pause and ask, "Do you need a hint?"
- Is the student frustrated? The AI can change its tone to be more encouraging.
- Is the student bored? The AI can make the feedback more fun or interactive.
The Risk: If the waiter is a robot, students might not trust it. They might think, "Why should I listen to a computer? It doesn't care about me." The paper suggests we need to teach students how to trust and use AI feedback, just like we teach them how to use a calculator.
3. The "One-Size-Fits-All" Trap
In the past, researchers tested feedback in small, controlled labs (like a test kitchen). But real classrooms are chaotic. A student in a quiet library reacts differently than a student in a noisy gym class. A 10-year-old reacts differently than a 20-year-old.
The group argued that we need to stop testing in the "test kitchen" and start testing in the real restaurant.
- The Goal: We need massive, international studies to see what works where.
- The AI Role: AI is the only tool fast enough to run these massive experiments. It can test thousands of different feedback styles across different schools simultaneously to figure out the "secret sauce" for different types of learners.
4. The Human Element: Don't Fire the Chef
A major theme of the meeting was that AI should not replace teachers; it should empower them.
- The Analogy: Think of AI as a sous-chef. The sous-chef does the chopping, the peeling, and the prep work (grading, data analysis). This frees up the Head Chef (the teacher) to focus on the most important part: connecting with the students, inspiring them, and having deep conversations about their learning.
- The Warning: If we just let the AI talk to the students, we lose the human connection. Teachers need to be involved in designing the AI, not just using it. They need to know when to step in and say, "Hey, the robot is wrong, let's talk about this."
5. The Future Recipe
The paper concludes with a call to action. To make this work, we need:
- Open Source Kitchens: Instead of secret recipes owned by big tech companies, we need open platforms where researchers and teachers can share what works.
- AI Literacy: We need to teach students (and teachers) how to read the "ingredients list" of AI. They need to know when the AI is hallucinating (making things up) and how to verify the feedback.
- Collaboration: Psychologists, computer scientists, and teachers need to stop working in silos and start cooking together.
The Bottom Line
AI has the potential to turn feedback from a static grade on a paper (which often gets ignored) into a dynamic, ongoing conversation that happens in real-time. But, just like any powerful tool, it requires a skilled human to wield it. If we use it right, it won't just make grading faster; it will help every student feel seen, understood, and capable of improving.
Drowning in papers in your field?
Get daily digests of the most novel papers matching your research keywords — with technical summaries, in your language.