Entry Barriers to the Labor Market in the Era of Generative Artificial Intelligence: A Critical Review with a Two-Dimensional Adjustment Framework (2022–2026)

Generative artificial intelligence (AI) is restructuring labor markets in ways that earlier analyses did not anticipate. Adjustment has not fallen uniformly across experience levels; it concentrates on entry-level workers. This critical narrative literature review examines the empirical evidence from 2022 through early 2026 on how generative AI has affected hiring, wages, and the composition of junior positions in four high-exposure sectors: technology, finance, consulting, and administration. Studies from Harvard, Stanford, IESE, Brookings, the World Economic Forum, and the International Labour Organization converge on a common pattern. Firms that adopt generative AI reduce entry-level hiring while retaining incumbent workers, a phenomenon the literature now labels seniority-biased technological change. The dynamic compresses starting wages, narrows access to first jobs, and threatens the continuity of organizational human capital. Evidence from outside the United States is more mixed. Danish administrative data show precise null effects on early-career earnings and hours, indicating that institutional context shapes both the pace and the form of adjustment. A parallel body of evidence shows that, within adopting firms, generative AI complements the productivity of the juniors who remain. Substitution and complementarity therefore operate on different margins at the same time. This review contributes two original elements beyond synthesis. First, a structured visual synthesis of seven reported labor-market effect sizes and four reported within-firm productivity effects exposes the substitution–complementarity paradox along distinct outcome margins and identifies three sources of heterogeneity that explain the dispersion in reported effects. Second, a two-dimensional framework defined by institutional flexibility and AI adoption intensity reorganizes the evidence into four adjustment regimes—acute compression, latent vulnerability, cushioned adjustment, and pre-adjustment—generates testable predictions for understudied labor markets, and identifies the lower-left cell of the framework (most of the global South) as the most important empirical gap for future research.