Modeling and Forecasting the Dynamic Impact of Artificial Intelligence on the US Labor Market: Heterogeneity Analysis Based on Machine Learning and Mixed Effects Models (2018-2024)

This study systematically evaluates the dynamic impact of artificial intelligence (AI) on the US labor market from 2018 to 2024, with a focus on industry-level heterogeneity. Leveraging machine learning and Mixed-effects Modeling, we integrate random forests, Generalized Additive Models (GAM), and Mixed-effects frameworks to quantify nonlinear relationships and sector-specific sensitivities. Key findings reveal that AI market value (variable importance = 0.553 0.028) dominates AI adoption (0.447 0.032) in predicting net unemployment, reflecting the amplified economic valuation effects. GAM analyses identify a threshold effect: AI market value exhibits significant marginal impacts below 0.4 (slope =0.280.05,p<110), plateauing at higher values. Mixed-effects models further uncover stark industry disparities, with tourism demonstrating robust positive sensitivity to AI development (random slope = 109.37, p < 0.001), contrasting with negative responses in manufacturing (-12.99) and services (-25.43). Methodologically, the study highlights the utility of nonlinear modeling (pseudoR=0.384) but underscores limitations from small sample constraints (n = 8), including wide confidence intervals (AICc = 24.62 vs. AIC = 22.24) and residual right-skewness (skewness = 1.02 0.18). By bridging methodological rigor with actionable insights, this work advances the empirical understanding of AIs heterogeneous labor market impacts.