Abstract The rapid advancement of Artificial Intelligence (AI) has reignited a longstanding economic debate concerning technological unemployment and labor market transformation. While historical technological revolutions have ultimately generated net employment gains, AI differs in its capacity to automate not only routine manual tasks but also cognitive and decision-based functions. This paper examines the net employment effects of AI by analyzing the balance between job displacement and job creation across sectors. It evaluates whether AI primarily substitutes human labor or complements it by enhancing productivity and generating new occupational categories. The study adopts a labor economics framework grounded in skill-biased technological change and structural transformation theory. It argues that AI disproportionately displaces routine and middle-skill occupations while increasing demand for high-skill, technology-intensive roles. Simultaneously, AI fosters indirect employment through new industries, platform ecosystems, and productivity-induced demand expansion. However, the temporal mismatch between displacement and creation produces transitional unemployment and wage polarization. The paper further assesses sectoral heterogeneity, highlighting differential impacts in manufacturing, services, and knowledge industries. It also examines developing economies where informal labor dominance and limited reskilling infrastructure may intensify adverse outcomes. The findings suggest that net employment effects are not technologically predetermined but institutionally mediated. Education systems, labor mobility, industrial policy, and social protection mechanisms play decisive roles in shaping outcomes. The analysis concludes that AI is unlikely to produce permanent mass unemployment at the aggregate level but will significantly restructure labor markets. The critical policy challenge is not preventing automation but managing distributional consequences and accelerating workforce adaptation to ensure inclusive growth.