Exploring the impact of Generative AI (Gen-AI) on occupational structures can provide quantitative evidence to guide adjustments to college majors. This paper simplifies the analysis by excluding secondary complex factors such as macro-level consumption and regional industries. It integrates LLM text matching, the Logistic diffusion model, and the Malthusian growth equation to construct a differential equation dynamical model. Using long-term data from 1974 to 2024, the study provides quantitative forecasts of long-term employment scale and underlying mechanisms for three occupational categories: software engineers, electricians, and graphic designers. The study indicates that the impact of Gen-AI on employment exhibits significant nonlinearity and occupational differentiation. Due to the combined effects of task attributes and reliance on hands-on skills, the forecast for 2025–2035 shows that software engineers, driven by AI-assisted and AI-enhanced effects, will experience rapid nonlinear employment growth, with the workforce expanding from 2.1 million to 3.19 million (+51.9%), with AI contributing an additional 31.2% of this growth; Electricians, who rely heavily on in-person hands-on work, face minimal displacement and will maintain slow, steady linear growth, with the workforce expanding from 800,000 to 910,000 (+13.8%), representing a deviation of less than 1% from the no-AI baseline; Graphic designers, whose standardized graphic and text tasks are being replaced, exhibit a nonlinear contraction trend that has bottomed out, with the workforce shrinking from 220,000 to 182,000 (-17.3%), representing a net decrease of 24.5% compared to the no-AI scenario. Parameter perturbation tests (error < 6.1%) confirm the model’s strong robustness. The quantitative results provided by this study can serve as a reference for universities in dynamically adjusting enrollment quotas and curriculum systems.

Xingze Zhang, Peiyu Tang, Haoze Liu, Guorui Zhao. An Empirical Study on the Impact of Generative AI Based on a Quantitative Model of Differential Equations on Employment. The Frontiers of Society, Science and Technology (2026), Vol. 8, Issue 3: 74-81. https://doi.org/10.25236/FSST.2026.080311.

[1] Hinton M L, Tippins N T. A database for a changing economy: Review of the Occupational Information Network (O*NET)[M]. Washington, DC: American Psychological Association, 2010.

[2] Frey C B, Osborne M A. The future of employment: How susceptible are jobs to computerisation?[J]. Technological Forecasting and Social Change, 2017, 114: 254-280.

[3] Eloundou T, Manning S, Mishkin P, et al. GPTs are GPTs: An early look at the labor market impact potential of large language models[EB/OL]. 2023.

[4] Autor D H, Levy F, Murnane R J. The skill content of recent technological change: An empirical exploration[J]. Quarterly Journal of Economics, 2003, 118(4): 1279-1333.

[5] Rogers E M. Diffusion of Innovations[M]. 5th ed. New York: Free Press, 2003.

[6] Maslej N, Fattorini L, Perrault R, et al. Artificial Intelligence Index Report 2025[EB/OL]. Stanford Institute for Human-Centered Artificial Intelligence, 2025.

[7] Acemoglu D, Restrepo P. Robots and jobs: Evidence from US labor markets[J]. Journal of Political Economy, 2020, 128(6): 2188-2244.

[8] Box G E P. Robustness in the strategy of scientific model building[M]. New York: Academic Press, 1979.

[9] Saltelli A, Ratto M, Andres T, et al. Global Sensitivity Analysis: The Primer[M]. Chichester: John Wiley & Sons, 2008.