Studying at university is meant to help you gain skills and confidence — for many, AI usage seems to do the opposite.
15 July 2026
By Emily Reynolds
For better or for worse, generative AI tools are now widely used across academia. Their use is so widespread that some figures suggest over 90% of students use AI in one way or another, be that creating summaries, conducting searches, or completing essays. As these tools become more and more embedded, researchers are asking whether there may be downsides to relying on them too heavily.
In a new study, published in BMC Psychology, Wenlong Wang and colleagues found that students who relied more heavily on AI tended to have less confidence in their ability to succeed — a pattern that was linked to higher levels of burnout and anxiety. Their findings suggest that using such tools to help cope with the stresses of their workload may inadvertently undermine confidence and leave students more vulnerable to distress.
To understand the impact of AI tools, the team surveyed 1,623 undergraduate students from China. Participants completed questionnaires measuring academic stress, anxiety, burnout, and their belief in their ability to cope with challenges and succeed at tasks. They also completed a self-report measure of AI dependence, which focused not on how often they used AI tools, but how strongly they felt they relied on them for thinking, problem-solving, and getting academic work done. The researchers then looked at how these factors were related to one another.
Students who were under greater academic pressure were more likely to depend on AI tools, and those who relied more heavily on AI also reported higher levels of burnout and anxiety. But the researchers found that AI dependence was not just associated with these outcomes, but may actually help explain why they occur.
In fact, students who relied more heavily on AI tended to feel less capable of dealing with academic challenges on their own, a reduced sense of competence that was, in turn, linked to higher levels of burnout and anxiety.
This suggests that, while AI can help students navigate academic challenges in the short term, it might also reduce opportunities to build the confidence and sense of mastery that protect against burnout and anxiety in the long term. The findings also echo a growing body of research suggesting that regularly offloading cognitive work to AI can undermine people's capacity to think through problems independently, as well as their confidence in their own reasoning.
The study's analyses don't show whether dependence on AI actually causes lower self-efficacy, burnout, or anxiety, and it's possible that students who already lacked confidence were more likely to rely on AI in the first place. Future research could track students over time and test whether educational approaches that encourage students to critically engage with AI-generated content help preserve confidence and reduce reliance on AI.
Overall, the study suggests that AI is at its most beneficial when it acts as a scaffold for learning rather than a substitute for it — but that this is easier said than done (especially given its well-documented limitations, including inaccuracies and fabricated information).
SOURCE:
No comments:
Post a Comment