Structural Modelling of Generative AI Cognitive Integration, Algorithmic Literacy, and Academic Dependency in Peruvian Higher Education

In the rapidly shifting landscape of modern education, a new kind of learning tool has arrived with the force of a tidal wave: generative artificial intelligence. These are computer programs capable of writing essays, solving complex problems, and generating ideas that once required a human mind. For students in Peru, as in many places, these tools are now available at the click of a button. Yet, having a powerful tool is not the same as knowing how to wield it. The central question facing educators and researchers today is not whether students can use these machines, but how using them changes the way their brains work. Does understanding how the machine thinks help a student learn better, or does simply relying on the machine make them forget how to think for themselves? This tension lies at the heart of a new study conducted across three major regions of Peru, which sought to map the invisible pathways between knowing how technology works, how students integrate it into their thinking, and whether they become dependent on it.

The researchers began by gathering a large group of 642 university students from Lima, Arequipa, and Cusco, representing a mix of public and private institutions and various fields of study, from engineering to health sciences. They asked these students to reflect on their own habits and skills, specifically focusing on four key areas. First, they measured "algorithmic literacy," which is simply the ability to understand how an artificial intelligence generates its answers, spot its mistakes, and recognize when it might be biased. Second, they looked at "cognitive integration," or the extent to which a student uses the AI to deepen their own understanding, connect complex ideas, and improve their arguments. Third, they examined "academic dependence," a state where a student feels they cannot complete a task without the machine, or feels anxious when they cannot access it. Finally, they assessed "perceived academic performance," which is how students feel about their own grades and productivity. The study also considered "digital ethics," the personal moral compass a student carries regarding the responsible use of technology.

The investigation revealed a clear and powerful story about how these factors interact. The data showed that students who possessed a strong understanding of how artificial intelligence works were significantly more likely to use it as a tool to boost their own thinking. Instead of letting the machine do the work for them, these students used it to refine their arguments and connect difficult concepts. In other words, knowing how the technology functions acted as a key that unlocked a more sophisticated way of learning. Conversely, this same understanding acted as a shield against dependency. Students who were literate in the mechanics of AI were far less likely to feel that they needed the machine to survive their coursework or to feel anxious without it. The study found that when students understood the "black box" of the algorithm, they were less prone to the trap of outsourcing their mental effort entirely.

However, the relationship between using AI and doing well in school was not a straight line; it depended entirely on how the tool was used. The research confirmed that when students successfully integrated AI into their own thinking process, they reported feeling more productive and saw their grades improve. But when students fell into the pattern of dependence, relying on the machine to do the heavy lifting without critical review, their perceived academic performance suffered. The machine, in this case, became a crutch that weakened their own metacognitive skills. A crucial finding emerged regarding the role of ethics: students who held strong ethical principles about how to use technology responsibly were better protected. For these students, understanding the technology led to an even stronger reduction in dependency, suggesting that moral awareness amplifies the benefits of technical knowledge.

The study also addressed a common worry in educational research: whether the results were simply a reflection of other hidden factors, such as a student's wealth or internet access. By using advanced statistical methods to control for these variables, the researchers confirmed that the relationships they found were robust and direct. The model they built explained more than half of the variation in how students perceived their own academic success, a level of clarity that is rare in social science. The data ruled out the idea that these effects were random or caused by outside circumstances, pointing instead to a specific causal chain: understanding the technology leads to better integration and less dependence, which in turn leads to better academic outcomes.

These findings carry a significant message for the future of education in Peru and beyond. The evidence suggests that the solution to the challenges posed by artificial intelligence is not to ban it or to ignore it, but to teach students how to understand it. The study indicates that policies focused solely on prohibition are insufficient. Instead, universities should focus on developing "critical computational thinking," helping students decipher the logic behind the algorithms they use. By fostering a deep understanding of the technology and pairing it with strong ethical guidelines, educators can help students transform these tools from sources of dependency into instruments of intellectual growth. The research concludes that the path forward lies in empowering students to be the masters of the technology, ensuring that the machine serves the mind rather than replacing it.