A cross-sectional survey published in BMC Medical Education reports that high-duration use of generative artificial intelligence, or GenAI, is associated with increased odds of anxiety symptoms and sleep-onset difficulty among medical students in Shanxi, China. The study, authored by Dan Guo, Xiaoyun Su, Cuimei Guo, Yanan Zhao, and colleagues, analyzed data from 1,882 valid questionnaires collected from medical students at four medical universities in the region. The research aimed to analyze the correlation between GenAI usage duration, anxiety, depressive symptoms, and sleep-onset difficulty, noting that the widespread popularity of GenAI has changed the learning and lifestyle of medical students while its association with mental health status and sleep problems remained insufficiently understood.
The researchers adopted a cross-sectional survey design, recruiting participants from four medical universities in Shanxi, China. Data were collected via an online questionnaire in which participants self-reported their usage duration of GenAI, anxiety and depressive symptoms, and sleep-onset difficulty. To explore the associations between these variables, the study team performed multiple logistic regression and linear regression analysis, adjusting for all covariates to isolate the specific relationships between GenAI use and the measured health outcomes.
The central finding of the study is that high-duration GenAI use was associated with increased odds of anxiety symptoms. After adjusting for all covariates, the analysis showed an odds ratio of 2.02, with a 95% confidence interval ranging from 1.48 to 2.77. The P value for trend was less than 0.001, indicating a statistically significant association. This result suggests that as the duration of GenAI use increases, the likelihood of experiencing anxiety symptoms also rises among this specific population of medical students.
In addition to anxiety, the study found a significant association between high-duration GenAI use and sleep-onset difficulty. Based on the sleep latency component of the Pittsburgh Sleep Quality Index, or PSQI, high-duration users exhibited higher odds of sleep-onset difficulty compared to low or non-users. The odds ratio for this association was 1.67, with a 95% confidence interval of 1.26 to 2.24. The P value for trend was also less than 0.001, reinforcing the statistical significance of the link between extended GenAI usage and difficulty falling asleep.
The relationship between GenAI use and depressive symptoms presented a more nuanced picture. High-duration GenAI use was associated with elevated depressive scores, with a beta coefficient of 0.66 and a 95% confidence interval of 0.15 to 1.17. The P value for trend was less than 0.01, indicating a significant association with the continuous measure of depressive scores.
The study emphasizes that GenAI is widely used among medical students in Shanxi, China. The cross-sectional nature of the survey means that the data provide a snapshot of the associations at a single point in time, rather than tracking changes over an extended period. While the statistical models adjusted for various covariates, the design does not allow for causal inferences.
Future research should deploy longitudinal designs to clarify the long-term associations between GenAI use duration, mental health, and sleep problems.
The article was published on 31 October 2026 in BMC Medical Education, an open-access journal.
Subject of Research: GenAI use duration, anxiety, depressive symptoms, and sleep-onset difficulty among medical students in Shanxi, China: a cross-sectional survey
Article Title: GenAI use duration, anxiety, depressive symptoms, and sleep-onset difficulty among medical students in Shanxi, China: a cross-sectional survey
Article References: Guo, D., Su, X., Guo, C., & Zhao, Y. (2026). GenAI use duration, anxiety, depressive symptoms, and sleep-onset difficulty among medical students in Shanxi, China: a cross-sectional survey. BMC Medical Education. https://doi.org/10.1186/s12909-026-10747-0
Image Credits: AI Generated
DOI: 10.1186/s12909-026-10747-0
Keywords: GenAI, duration, anxiety, depressive, symptoms, sleep-onset, difficulty, medical, students, Shanxi, China, cross-sectional
News Source: Glenn Wilkins. (November 12, 2026). Study Links Heavy GenAI Use to Student Anxiety. Scienmag.



