Artificial Intelligence Usage, Cognitive Offloading, and Academic Performance among Rural Undergraduate Students: A Sequential Mediation Model

Imagine a student sitting in a quiet room late at night, facing a difficult assignment. In the past, solving a problem meant wrestling with it, trying different approaches, and slowly building understanding through effort. Today, that same student can type a question into an artificial intelligence tool and receive a complete, polished answer in seconds. This shift is happening rapidly in schools across India, particularly in rural areas where students often lack access to dedicated tutors or extensive libraries. The central question for educators and psychologists is not whether these tools exist, but what happens to the student's mind when they rely on them. Two key ideas help explain the stakes. First, the brain is not a sealed container; it often uses outside tools to help it think, a concept known as the extended mind. Second, learning requires a certain amount of mental struggle, known as cognitive load, to build strong, lasting memories. If a tool removes all the struggle, it might remove the learning itself.

A team of researchers set out to understand exactly how this dynamic plays out for undergraduate students in rural India. They focused on a specific group of 425 students attending government and private colleges in the countryside. These students were navigating a landscape where powerful AI tools are available on mobile phones, yet there is little formal guidance on how to use them wisely. The researchers wanted to see if using these tools helped students learn better or if it changed the way they approached their studies in a way that hurt their long-term performance. They built a model to trace a specific chain of events: how using AI leads to a habit of offloading mental work, how that habit affects the student's ability to manage their own learning, and finally, how all of this influences how they perceive their academic success.

The study began by asking students to report on their habits. They measured how often students used AI tools for schoolwork, how much they relied on those tools to do the thinking for them, how well they planned and monitored their own study sessions, and how they felt about their grades. The researchers found a clear and strong pattern. Students who used AI tools more frequently were significantly more likely to engage in what is called cognitive offloading. This is the practice of handing over mental tasks, like recalling facts or solving problems, to the computer instead of doing the work themselves. The data showed that heavy AI use was the strongest predictor of this behavior, explaining more than half of the variation in how much students relied on the technology to think for them.

However, the story did not end there. The researchers discovered that this habit of offloading mental work had a direct, negative impact on a student's ability to regulate their own learning. Self-regulated learning is the skill of setting goals, planning study time, checking one's own understanding, and seeking help when needed. The study found that when students offloaded too much thinking to AI, their ability to manage their own learning processes weakened. They became less likely to plan their study sessions or monitor their comprehension. Crucially, the researchers found that this loss of self-management was the real reason academic performance suffered. The AI tool itself did not directly lower grades in a simple, one-step way. Instead, the tool encouraged a habit of mental delegation, which then eroded the student's self-discipline and study strategies, and it was this erosion that led to poorer academic outcomes.

Interestingly, the study also revealed a more complex picture. While heavy offloading was harmful, using AI tools did not automatically destroy a student's ability to learn. In fact, for some students, using AI was linked to better self-regulation, perhaps because the tools helped them clarify difficult concepts or access information they otherwise could not find. The key difference lay in how the student used the tool. Those who used AI as a substitute for thinking saw their learning skills decline. Those who used it to support their own thinking maintained or improved their study habits. The researchers confirmed that the most dangerous path was a specific sequence: using AI led to offloading, offloading reduced self-regulation, and reduced self-regulation lowered academic performance. This chain of events was statistically significant and accounted for a large portion of the impact AI had on student success.

The findings took on a sharper edge when the researchers looked at different groups of students. They compared those who were already struggling academically with those who were performing well. The results showed a stark difference in vulnerability. Students with lower grades were far more likely to fall into the trap of offloading their thinking to AI. For these students, the link between using AI and handing over mental work was nearly twice as strong as it was for their high-performing peers. This suggests that the students who need the most support are the ones most at risk of being harmed by unsupervised AI use. They are more likely to use the technology as a shortcut that bypasses the effort required to learn, while higher-performing students are better able to use the same tools to scaffold their own learning without losing control of the process.

The researchers concluded that the solution is not to ban these tools, which are already deeply embedded in student life, but to change how students are taught to use them. The study suggests that simply giving students access to AI without teaching them how to manage their own learning is dangerous. Instead, colleges need to integrate specific training that helps students distinguish between using AI as a helper and using it as a replacement for their own brain. This includes teaching students to check their own understanding, to plan their study time, and to view AI outputs as something to be critiqued rather than accepted. The goal is to turn AI from a convenience that erodes learning into a tool that supports it. For rural students, who often lack other academic resources, this distinction is vital. Without guidance, the technology risks widening the gap between those who can learn effectively and those who cannot, but with the right support, it could become a powerful equalizer.