The rise of AI tools like ChatGPT and CoPilot has brought both exciting opportunities and significant challenges to education. While these tools can assist students in completing tasks, they also raise an important question: are they helping students develop metacognitive, critical thinking skills, or are they allowing them to avoid the harder, more rewarding processes of learning?
A recent study explored this very issue, investigating whether AI tools enhance learning or encourage students to disengage from critical thinking. In this blog, we’ll dive into the study’s findings, discuss the risks of over-reliance on AI, and explore how these tools can be used in ways that promote, not hinder, metacognitive growth.
What does research say?
The study investigated how different learning tools affected student outcomes. The students were divided into different groups that included, but were not limited to:
- ChatGPT group: Students used ChatGPT for guidance on their tasks, but the tool was restricted from directly generating essays.
- Human Expert group: Students worked with experienced human tutors for one-on-one academic writing support.
- Control group: Students completed the task independently, with no external support.
The study aimed to compare how the groups performed in terms of motivation, self-regulated learning, and overall academic performance. Here’s what the researchers found:
- Motivation: The study did not find any major differences in motivation between the students who used ChatGPT or had help from tutors, but some trends stood out. In contrast, the Control group had the lowest interest and the highest levels of tension, showing that external feedback can help reduce stress and keep students engaged with the learning content.
- Self-Regulated Learning: The ChatGPT group didn’t show as much self-reflection with their work compared to the Human Expert group. This is what the researchers mean when they talk about AI making students ‘metacognitively lazy’. The students who worked with human tutors showed stronger connections between reading, planning, and evaluating their work. This suggests that human interaction encourages deeper thinking, metacognitive processing and self-reflection.
- Performance: Students in the ChatGPT Group showed the most improvement in essay scores, even outperforming the Human Expert Group. This suggests that personalised, immediate feedback from AI can be effective. However, there was no significant difference in how well the ChatGPT group applied what they learnt to new situations, pointing out a limitation in fostering deeper learning. This suggests that they receive short-term gains without any of the long-term benefits.
The Case for Metacognitive Laziness
These findings highlight a key concern: over-reliance on AI tools like ChatGPT may pose issues that educators should be aware of. One major issue is the potential for students to misuse ChatGPT to generate essays or solutions and present them as their own work. Not only does this amount to plagiarism, but it also undermines students’ sense of ownership over their work. When they lean too heavily on AI, they may fail to take responsibility for their learning, including the feedback they receive.
Not only will students not be accountable for their work, but it often does so also at the cost of deeper engagement with the material. The process of learning involves more than just completing tasks; it requires developing problem-solving skills and engaging in meaningful reflection. Using AI to generate answers bypasses the learning process entirely, robbing students of the opportunity to engage in skills like critical thinking and self-evaluation. These are the types of metacognitive soft skills that students need to navigate the real-world challenges outside of the classroom.
Final thoughts
AI tools like ChatGPT are transforming education, sometimes for the better, sometimes for the worst. At its best, AI should serve as a complement to the learning process, not a shortcut. Unfortunately, the way ChatGPT is designed means this isn’t often the case. It gives quick answers and often isn’t being used by students to develop their thinking. Our blog on the 3 levels of AI use discusses how to frame AI as a tool for reflection and collaboration, rather than a shortcut to complete tasks quickly.
While AI can enhance productivity and task performance, the risks of metacognitive laziness and over-reliance cannot be ignored. By understanding the strengths and limitations of AI, we can better navigate this shifting landscape.
To learn how to use AI successfully, check out our blog on our 8 principles for effective AI use, or book your staff an AI in Education CPD workshop.
