This paper examines task-biased automation affects gender inequality in labor market. We develop a framework in which the gender-differentiated effects of task-biased technological change arise from the endogenous targeting of automation toward tasks that differ systematically in gender intensity. Women are disproportionately employed in mid-level routine cognitive tasks, which are more exposed to automation given their position in the task–skill distribution. We construct task-level measures of automation exposure and study their relationship with changes in gender employment and wage gaps in a stacked first-difference design. We find that tasks with higher female intensity experience greater exposure to automation and larger relative declines in female employment and wages. The results suggest that task-level automation can generate gender-differentiated labor market outcomes even in the absence of gender-specific labor demand.

Why does technological change lead to successful upgrading in some economies, but persistent stagnation in others? I develop a dynamic framework in which technological change shapes the evolution of comparative advantage not only by amplifying existing productivity differences, but—more importantly—by reallocating production toward activities with heterogeneous learning potential. When technological change disproportionately expands low-learning tasks, economies may experience short-run productivity gains while failing to accumulate the capabilities necessary for future comparative advantage. The model generates endogenous divergence in production structures and explains why some countries upgrade and acquire new comparative advantages under technological change, whereas others remain trapped in activities with limited learning opportunities.