The utility of AI in education
Bradley Busch, Director at InnerDrive
Artificial Intelligence( AI) is rapidly transforming many aspects of our lives, and education is no exception. Teachers and students alike are exploring how AI can be harnessed to enhance learning, improve grades, and streamline teaching processes. But with all the excitement around AI’ s potential, it’ s important to take a closer look at its real impact in the classroom; both positive and negative.
In this article, we’ ll take a closer look at the opportunities and traps that come with AI in education. Specifically:
• How AI affects student learning and grades
• Whether current strategies to detect and limit AI use really work
• How AI can be integrated into teaching to improve feedback and support students
AI and academic performance
At first glance, research on AI in education seems overwhelmingly positive. Deng et al.( 2025) published a meta-analysis of 69 studies suggesting that ChatGPT boosts student performance( e. g., Hsu, 2023), increases motivation( e. g., Wahba et al., 2024), and reduces the mental effort needed to complete tasks( e. g., Li et al., 2024). These findings paint an encouraging picture. After all, who wouldn’ t want students to achieve better results with less struggle?
However, a closer look reveals significant limitations that challenge these conclusions. Most of the studies( 84 %) were conducted on university students, a group that is older, more independent, and academically successful. This focus raises concerns about whether the same benefits apply to younger secondary and primary school students, who often have different levels of motivation, cognitive development, and study habits.
An even deeper flaw emerges when we examine how students were allowed to use ChatGPT. Many studies permitted its use in final assessments( e. g., Bašić et al., 2023), and nearly half failed to clarify whether AI use was restricted in exams( e. g., Song & Song, 2023). This is a crucial oversight because real-world education often prohibits AI in high-stakes tests.
Importantly, research suggests that when students use ChatGPT primarily for quick answers rather than deeper understanding, their long-term learning suffers. For instance, Bastani et al.( 2022) found that ChatGPT hindered learning when it was available for study but not for exams, as students rarely used it for meaningful engagement with the material. This contradicts wellestablished learning principles that emphasise effortful thinking( De Bruin et al., 2023). While AI can enhance student output, it may not necessarily improve learning, especially if students rely on it to bypass cognitive effort( Kuldas et al., 2014). The key challenge for educators is ensuring AI is used as a tool to deepen understanding rather than replace the thinking process entirely.
CORRECT ANSWERS
0.45
0.36
0.27
0.18
0.9
Control group AI group
0 STUDYING
FINAL EXAM
Innerdrive. co. uk
4