Can we embrace generative AI in education?
The learning process should be valued over measures of achievement, such as grades
Generative AI like ChatGPT is popular among students, but the educators we work with across countries like Germany, the Netherlands and Finland see it as both an opportunity and a risk. They are particularly concerned about its impact on future learning and teaching methods. We are researchers dedicated to understanding learning strategies to enhance learner agency in the AI era. While we acknowledge educators’ concerns, we believe in the potential of AI if we rethink the way we view and assess learning.
An education system that prioritizes the learning process over measures of achievement, such as grades, provides an opportunity for students to realize their agency. With less pressure to achieve high grades, students can focus on and experiment with ways of learning to find which work best for them while engaging with the learning material. This enables students to gain a better understanding of their own learning process, and in time to expand their knowledge of subject matter. Ultimately, this would create more resilient, adaptive students who are better prepared to address complex, real-world problems.
“An education system that prioritizes the learning process over measures of achievement, such as grades, provides an opportunity for students to realize their agency.”
Rethinking learning in the AI era
By tailoring the digital learning environment to students’ preferred learning strategies, digital learning tools and AI can help students learn better. However, giving students the freedom to choose their own method of learning could make it harder for them to regulate their learning, as not all students possess the necessary self-regulation skills and self-awareness to do so. Students may become overwhelmed with the multiple options offered by open-ended, digital settings and choose familiar learning methods instead of more effective approaches. They may also mistakenly take superficial cues, such as producing text fluently, as a sign of an effective learning strategy, as shown in our research. Students need to develop the ability to determine whether a given learning method works for them, and to make changes if it doesn’t.
The challenges students face in regulating their learning may be explained, in part, by the fact that that they tend to focus on measures like exam scores and course grades rather than on how well they have learned the material. In our study of essay writing in Dutch secondary schools, students were given feedback on their learning process. However, they mainly wanted to evaluate their performance with their essay grades, struggling to reflect on their learning process independent of those measures. In some private schools in Germany, students often exceeded the stated word limit, perhaps believing they would receive better grades if they provided more information. In Finland, when asked to describe their learning process, some students focused first on their essay’s grading. All these examples, observed during our recent data collection, show the same grade-oriented mindset.
This focus on grades makes it difficult for students to judge how well they are learning. Consequently, they are unable to identify aspects of their learning that they could change to refine their learning strategies. Changes need to be made so that students are able to understand their learning processes, self-regulate, and become learners who actively monitor their own learning. This is important because learning regulation is dynamic and cyclical, shaping lifelong learning through each learning experience.
Why does the learning process matter?
Many agree that the aim of education, especially in schools, is not only to help pupils achieve excellence, but also to promote effective learning that can be sustained and provide long-term benefits. Education should equip learners with the competencies they will need for the future. It is therefore important to consider the immediate effect of digital tools on educational performance, as well as their longer-term impact on student development.
In some cases, generative AI can take over the learning process instead of simply providing assistance. For example, a student might copy an essay directly from ChatGPT without reviewing it. While the tool might, in that case, help the student achieve a higher grade in the short term, it does not develop the student’s skills or knowledge.
“To have a long-term impact, students must understand how AI works, and what it can and cannot do.”
We believe that to have a long-term impact, students must understand how AI works, and what it can and cannot do. They must also be able to take charge of their learning while using AI tools. It is our hope that if school lessons focus on the learning process and not simply on measures of achievement, students, too, will shift their attention towards the learning process.
As researchers, we need to explore how students can work with generative AI to develop self-regulated learning skills and ultimately also boost their learning outcomes. Generative AI is not something to shy away from, but a tool to be explained and taught to students so that it is helpful instead of harmful.
Footnotes
Our research is part of a pioneering initiative known as the Center for Learning and Living with AI (CELLA). This initiative is a collaboration among education and technology researchers and practitioners that is dedicated to finding strategies for enhancing learner agency in the AI era. CELLA focuses on leveraging digital technologies to support self-regulated learning in K-12 education settings across five different countries. In promoting learner agency, our team seeks to increase students’ awareness of their learning processes and their self-regulation skills. This, in turn, could have a positive effect on students’ performance and grades.
We would like to express our gratitude to the following individuals for their help with and contributions to this article: Susanne de Mooij, Joni Lämsä, Maria Bannert, Sanna Järvelä, and Inge Molenaar.
Authors
