This paper examines how AI at work impacts on-the-job learning, shedding light on workers’ reactions to the groundbreaking AI technology. Based on theoretical analysis, six hypotheses are proposed regarding three aspects of AI’s influence on on-the-job learning. Empirical results demonstrate that AI significantly inhibits people’s on-the-job learning and this conclusion holds true in a series of robustness and endogeneity checks. The impact mechanism is that AI makes workers more pessimistic about the future, leading to burnout and less motivation for on-the-job learning. In addition, AI’s replacement, mismatch, and deskilling effects decrease people’s income while extending working hours, reducing their available financial resources and disposable time for further learning. Moreover, it has been found that AI’s impact on on-the-job learning is more prominent for older, female and less-educated employees, as well as those without labor contracts and with less job autonomy and work experience. In regions with more intense human–AI competition, more labor-management conflicts, and poorer labor protection, the inhibitory effect of AI on further learning is more pronounced. In the context of the fourth technological revolution driving forward the intelligent transformation, findings of this paper have important implications for enterprises to better understand employee behaviors and to promote them to acquire new skills to achieve better human–AI teaming.

Purpose: Existing studies on the association between BMI and depression report conflicting results with some demonstrating a positive relationship, while others a negative link or insignificant correlation. Very limited research on the nonlinear relationship between BMI and depression has yet to clarify the reliability and robustness of the potential nonlinearity and whether a more complex association exists. This paper aims to systematically investigate the nonlinear relationship between the two factors applying rigorous statistical methods, as well as explore the heterogeneity of their association. Materials and Methods: A large-scale nationally representative dataset, Chinese General Social Survey, is used to empirically analyze the nonlinear relationship between BMI and perceived depression. Various statistical tests are employed to check the robustness of the nonlinearity. Results: Results indicate that there is a U-shaped relationship between BMI and perceived depression, with the turning point (25.718) very close to while slightly larger than the upper limit of the range of healthy weight (18.500 ≤ BMI < 25.000) defined by World Health Organization. Both very high and low BMIs are associated with increased risk for depressive disorders. Furthermore, perceived depression is higher at almost all BMI levels among individuals who are older, female, lower educated, unmarried, in rural areas, belonging to ethnic minorities, non-Communist Party of China members, as well as those with lower income and uncovered by social security. In addition, these subgroups have smaller inflection points and their self-rated depression is more sensitive to BMI. Conclusion: This paper confirms a significant U-shaped trend in the association between BMI and depression. Therefore, it is important to account for the variations in this relationship across different BMI categories when using BMI to predict depression risk. Besides, this study clarifies the management goals for achieving an appropriate BMI from a mental health perspective and identifies vulnerable subgroups at higher risk of depression.