Although the use of AI may improve students’ performance on average, what matters more is how students use it. If students use AI instrumentally, their grades may even be lower than those of students who do not use AI at all. By contrast, using AI to seek feedback and identify missing perspectives may lead to significantly better performance than other approaches to AI use.

These findings emerged from a recent study conducted at the University of Eastern Finland, which examined how students use AI in their studies and whether different approaches to using AI affect student performance.

The study identified three distinct approaches to using AI:

• Instrumental use of AI involves delegating assignments to AI and using AI-generated responses without substantial critical reflection or revision.
• Ideational use of AI refers to using AI to generate ideas and perspectives for an assignment. Ideational use leads to better learning outcomes than instrumental use, but students often place too much trust in the quality of the ideas generated by AI.
• Reflective use of AI refers to an approach in which students first generate content themselves and then ask AI to provide feedback or identify perspectives that are missing from it. In this approach, students draw on their own expertise and skills, while AI becomes a means of supporting their learning and the development of their competence. This is also directly reflected in students’ grades: students who use AI in this way achieve the highest grades, compared both with students who use AI in other ways and with those who do not use AI at all.

“This is why it is important to teach students how to use AI effectively and purposefully, without using it as a substitute for their own work or thinking, so that it becomes a useful part of their learning and competence development,” says Associate Professor Outi Hakola of the University of Eastern Finland.

The use of AI is neither inherently good nor bad. What matters is whether it is used in a meaningful way.

“Based on our findings, we see reflective use of AI as one possible path towards using AI in a way that is both responsible and supportive of student performance.”

According to Hakola, it is often assumed in public debate that students use AI in only one way. Previous research on the topic has also primarily focused on how well AI can complete assignments.

“This is an overly simplistic view of a complex issue.”

The data consisted of take-home exams completed by students in 2023 and 2024. Students were permitted to use AI, provided that they disclosed its use. The study involved 891 students, 159 of whom reported using AI.

“During the course of this study, the number of AI applications increased and people’s awareness of them grew. The data therefore come from the very period when AI was gradually becoming part of university teaching and learning.”

According to Hakola, it would be interesting to investigate in around five years’ time how AI has become integrated into students’ learning processes.

Research article:
Outi Hakola, Jari Kukkonen, Teemu Valtonen, Self-regulated approaches to students’ generative AI use in higher education, Computers in Human Behavior Reports, Volume 24, 2026, 101340, ISSN 2451-9588, https://doi.org/10.1016/j.chbr.2026.101340