Industry Updates
Unlocking the Power of Feedback with Generative AI
Head of Digital Learning and Artificial Intelligence, Kaplan UKFeedback is crucial in education, serving learners, teachers, and educational providers alike. And a particular kind of feedback, formative assessment – the process of understanding students’ progress, helping them to identify key strengths and weaknesses, and providing strategies for improving their outcomes – is at the heart of effective teaching.
In their 2001 paper Inside the Black Box, Paul Black and Dylan Wiliam argue that the true driver of learning is not external testing or curriculum changes, but the dynamic interactions between teachers and students within the classroom. Considerable research highlights that improving formative assessment can substantially raise learning standards, especially benefitting lower-achieving learners and reducing educational gaps. In other words, it is no exaggeration to say that formative assessment is educationally transformative.
However, using this practice regularly and effectively is challenging, especially for teachers with large classes or limited time. Providing the right amount of feedback is crucial: too much feedback can overwhelm students, while too little can leave them adrift. An example of the wrong sort of feedback is that given by a satnav (aka GPS). Satnavs are great. I use mine all the time. Even when traveling on routes I know well, it can give me up-to-date traffic info, let me know about roadworks or closures, and provide a relatively accurate ETA. But it’s rubbish at teaching me new routes. The feedback it provides means that I no longer have to think or learn. I just do what I’m told, safe in the knowledge that if I make an error the satnav will make the corrections for me. I haven’t learned a new route for years. What’s worse, I’ve become so reliant on the feedback from my satnav that I can no longer navigate independently.
In the classroom, time and resource constraints often limit teachers’ ability to provide meaningful, timely and personal feedback. In online or asynchronous environments, this challenge is even more pronounced. Traditional methods, like multiple-choice questions, offer limited feedback along the lines of correct or incorrect, and lack the personalization necessary for effective formative assessment. Multiple choice questions also lack repeatability as students can remember the answers even if they haven’t learned the solutions.
Enter Generative AI (GenAI), a game-changing tool poised to transform formative assessment.
GenAI can generate detailed, meaningful feedback tailored to each student’s responses. With AI, we can write prompts that provide a model question and answer, and specific feedback. GenAI will highlight strengths, pinpoint weaknesses, provide tips for improvements, offer encouragement, and suggest supplemental resources or exercises tailored to the student's specific needs. This feedback can be aligned with the course objectives and learning outcomes, not only addressing the content of the student's answer but also the underlying skills and knowledge required to reach mastery. This approach also fosters critical thinking and self-reflection, prompting deeper engagement with the material so students become more active participants in their own learning.
Integrating AI into online courses can provide continuous, meaningful feedback throughout a student’s learning journey. Beyond simple question and answer formats, we can also provide AI feedback through discussions, mind-mapping activities, or project submissions. The interactive process helps students develop higher-level skills.
The availability of GenAI feedback around the clock benefits both self-directed and classroom learners. Questions can be repeated as often as a student desires, whenever they need feedback, and the AI won’t get tired or become irritable.
Simply having this immediate, personalized feedback can improve educational standards significantly. But imagine if we could store and analyze all students' work and feedback data. GenAI can create personalized learner profiles, identifying progress and areas of challenge. It can analyze trends and patterns in a student's performance. And it can customize future learning content to pinpoint exactly where intervention is needed.
Aggregating this data across all students can also inform broader educational strategies. Educators and course designers can identify curriculum areas where students struggle and determine the most effective teaching methods. This can lead to a more responsive and dynamic curriculum that evolves based on actual student performance and needs.
This integration of GenAI technology creates a loop of continuous improvement. As students interact with the course material and receive feedback, the system learns and adapts, creating a personalized educational experience that is constantly refining itself. This not only enhances student engagement and satisfaction but also drives up learning standards across the board.
What about the limitations of GenAI? It makes things up, right? How do we know that the feedback is good feedback? Is there a risk that we will replace our teachers with bots? These concerns are valid, but they shouldn’t stop the conversation. Our approach has to involve a multi-tiered system of checks and balances. First, GenAI-generated feedback must follow pedagogical best practices and be routinely reviewed by human experts to ensure accuracy and relevance. AI should be used to complement, not replace, our teachers, enhancing their ability to connect with and support every student. By integrating GenAI responsibly, we can mitigate risks while enriching human interaction.
Assuming we manage these risks effectively and maintain the necessary human oversight and intervention, the combination of immediate, personalized feedback with a data-driven approach to learning creates a powerful educational tool. Aligning with the principles of formative assessment, this approach provides timely, actionable feedback that helps students not just learn, but learn how to learn. By leveraging GenAI in this manner, we can create a more effective and equitable educational environment, offering every student the opportunity to reach their full potential.