(Photo credit: Adobe Stock)
Students tend to turn to ChatGPT, a generative artificial intelligence tool, when faced with increased academic workload and time constraints, according to new research published in the International Journal of Educational Technology in Higher Education. The study also reveals a concerning trend: reliance on ChatGPT is linked to procrastination, memory loss, and a decline in academic performance. These findings shed light on the role of generative AI in education, suggesting both its widespread use and potential drawbacks.
The motivation behind this research stems from the explosive growth of generative AI technologies in educational settings. Despite their potential to assist in learning and research, there’s a growing concern among educators about their misuse, especially in relation to academic integrity. Previous studies have largely focused on theoretical discussions without much empirical data to support the claims.
“My interest in this topic stemmed from the growing prevalence of generative artificial intelligence in academia and its potential impact on students,” explained study author Muhammad Abbas, an associate professor at the FAST School of Management at the National University of Computer and Emerging Sciences in Pakistan. “For the last year, I observed an increasing, uncritical, reliance on generative AI tools among my students for various assignments and projects I assigned. This prompted me to delve deeper into understanding the underlying causes and consequences of its usage among them.”
To understand these dynamics, the study was conducted in two phases. Initially, the researchers developed and validated a scale to measure university students’ use of ChatGPT for academic purposes. They began by generating an initial set of 12 items, which was refined to 10 after expert evaluations for content validity. Further refinement through an exploratory factor analysis and reliability testing led to the final selection of eight items that effectively measured the extent of ChatGPT’s academic use.
The scale included items such as: “I use ChatGPT for my course assignments,” “I am addicted to ChatGPT when it comes to studies,” and “ChatGPT is part of my campus life.”
In the second phase of the study, the researchers sought to validate the findings from the first phase while also testing specific hypotheses related to ChatGPT’s impact. The sample consisted of 494 university students who were surveyed across three timepoints, each separated by a 1-2 week interval.
This time-lagged approach allowed the researchers to first gather data on predictor variables (academic workload, time pressure, sensitivity to rewards, and sensitivity to quality), followed by the measurement of ChatGPT usage, and finally, the assessment of outcomes (procrastination, memory loss, and academic performance).
Abbas and his colleagues found that high levels of academic workload and time pressure were significant predictors of increased ChatGPT usage, suggesting that students under significant academic stress are more likely to turn to generative AI tools for assistance.
Students who were more sensitive to rewards were less inclined to use ChatGPT, indicating a possible concern about the academic integrity and the potential negative consequences of relying on AI for academic tasks.
Moreover, the study uncovered significant adverse effects of ChatGPT usage on students’ personal and academic outcomes. Increased reliance on ChatGPT was associated with higher levels of procrastination and memory loss, and a negative impact on academic performance, as reflected in students’ grade point averages. These findings suggest that while ChatGPT can be a valuable resource under certain circumstances, its excessive use might lead to detrimental effects on learning behaviors and outcomes.
“One surprising finding was the role of sensitivity to rewards,” Abbas told PsyPost. “Contrary to expectations, students who were more sensitive to rewards were less likely to use generative AI. Another surprising finding was the positive relationship of generative AI usage with procrastination and self-reported memory loss – and negative relationship between generative AI usage and academic performance.”
Interestingly, the study did not find a significant relationship between sensitivity to quality and ChatGPT usage, suggesting that concerns over the quality of academic work do not necessarily influence the decision to use AI tools.
The findings highlight the potential dual impact of ChatGPT in academia, serving both as a helpful tool under academic pressure and as a potential risk to academic integrity and student learning outcomes.
“The average person should recognize the dark side of excessive generative AI usage,” Abbas said. “While these tools offer convenience, they can also lead to negative consequences such as procrastination, memory loss, and compromised academic performance. Also, factors like academic workload, sensitivity to rewards, and time pressure play significant roles in influencing students’ decision to use generative AI.”
The study provides important details about ChatGPT usage among university students. But the study, like all research, includes limitations. For example, the time-lagged design, while more robust than cross-sectional designs, does not entirely eliminate the possibility of reciprocal relationships.
The study suggests a one-way impact of ChatGPT usage on students’ academic workload and personal outcomes, such as procrastination and memory loss. However, it’s conceivable that these relationships could be bidirectional. For instance, students who are prone to procrastination might be more inclined to use ChatGPT, rather than ChatGPT usage leading to increased procrastination.
The research opens the door to investigating the broader effects of ChatGPT usage on students’ learning outcomes and health. Future research could delve into how reliance on generative AI tools affects cognitive skills, mental health, and overall learning experiences.
“My long-term goals involve expanding this line of research to further explore through other methods, such as experiments, how excessive use of generative AI affects students’ outcomes,” Abbas said.
The study, “Is it harmful or helpful? Examining the causes and consequences of generative AI usage among university students,” was authored by Muhammad Abbas, Farooq Ahmed Jam, and Tariq Iqbal Khan.