From Assistance to Dependence: Investigating the Effects of Generative AI on Student Learning, Retention, and Study Effort

Purpose: Generative Artificial Intelligence (GenAI), including tools such as ChatGPT, is increasingly used by university students for explanations, summaries, problem-solving, feedback, and academic support. This paper examines whether such assistance supports learning or may encourage dependence that affects long-term retention and independent thinking. Methods: The study uses secondary research and a descriptive synthesis of published experimental studies, randomized controlled trials, systematic reviews, and meta-analyses. No primary data were collected and no new statistical meta-analysis was conducted. Results: The reviewed studies generally report positive effects on academic performance and engagement. Reported effect sizes include approximately g = 0.712 across 69 experimental studies, Hedges’ g = 0.577 across 37 studies, and an effect size of approximately 0.804 across 57 studies. However, one randomized controlled study reported 45-day retention of 57.5% for a ChatGPT-assisted group compared with 68.5% for a traditional-learning group. Conclusion: GenAI can be a useful learning assistant when students actively use it for explanation, feedback, comparison, and clarification. However, completing tasks efficiently should not be treated as proof of deep or lasting learning. Students should retain opportunities for independent recall, reasoning, analysis, and problem-solving.