Attitudes toward artificial intelligence tools in university students: associations with body appreciation and e-healthy diet literacy - implications for digital health education

This cross-sectional study aimed to determine attitudes toward the use of artificial intelligence tools, body appreciation, and e-healthy diet literacy (e-HDL) levels among university students, and to analytically evaluate the relationships among these variables.

This cross-sectional study was conducted among 440 undergraduate university students aged 18-30 years enrolled at Gaziantep University. Data were collected with web-based questionnaire that included a Descriptive Information Form, the Artificial Intelligence Attitude Scale-Short Form (AIAS-4), the Body Appreciation Scale (BAS), and the e-HDL Scale.

The mean AIAS-4, BAS, and e-HDL total scores were 5.47 ± 2.62, 38.91 ± 9.00, and 37.89 ± 7.04, respectively. AIAS-4 scores were positively correlated with both BAS and e-HDL total scores. Male students had significantly higher AIAS-4 scores than female students, whereas female students had significantly higher BAS scores. Male students had significantly higher AIAS-4 scores than female students, whereas female students had significantly higher BAS scores; no significant sex difference was observed in e-HDL total scores. AIAS-4 and BAS scores differed significantly across BMI categories, while e-HDL total scores did not. In multiple linear regression analysis, higher BAS scores (β = 0.325, p < 0.001), higher e-HDL total scores (β = 0.247, p < 0.001), and higher BMI (β = 0.107, p = 0.021) were independently associated with higher AIAS-4 scores, whereas female sex was independently associated with lower AIAS-4 scores (β = -0.138, p = 0.003). Notably, BMI was not significantly correlated with AIAS-4 scores in the bivariate analysis (r = 0.083, p > 0.05), yet emerged as a significant independent predictor in the multivariable model (β = 0.107, p = 0.021). This discrepancy is consistent with a suppression effect, whereby sex - which is associated with both BMI and AI attitudes - masks the independent contribution of BMI in unadjusted analyses. After adjustment for sex and other covariates in the multivariable model, the association between BMI and AI attitudes became statistically detectable. This finding underscores the importance of multivariable adjustment when examining predictors that are inter-correlated. Age was not significantly associated with AIAS-4 scores. The regression model explained 22.0% of the variance in AIAS-4 scores.

Among university students, more positive attitudes toward artificial intelligence tools were associated with higher body appreciation and e-HDL. In addition, sex and BMI were independently related to AI attitudes, whereas age and year of study were not. Although nearly all participants reported using AI, the regression model explained 22% of the variance in AI attitudes, indicating that the drivers of more positive attitudes remain multifaceted and warrant further investigation.