Capability, Not Exposure: A Reallocation Model of AI, Work, and Entrepreneurship

Artificial intelligence (AI) is reshaping tasks, hiring, productivity and the economics of venture creation, yet future-of-work and entrepreneurial-ecosystem research are usually analyzed separately. This paper develops a reallocation model linking task-level AI exposure (E) to human-AI complementarity (C), reskilling capacity (S) and entrepreneurial absorption capacity (A). It yields a closed-form transition condition, C+S > ۲/(۲+λA), where λ discounts venture creation for lag and failure. Analyzed over the whole capability space rather than selected scenarios, and under independent uniform capability, the constructive share has the exact closed form ۱-۱/(۲+λ): at λ = ۱ entrepreneurial absorption rescues precisely one third of the configurations that would fail without it, and can never rescue the rest. The paper then stress-tests its own identity. Under randomized misspecification of every pathway equation the condition still classifies ۹۵.۲% of units correctly, though only ۵۹.۶% of those within ۰.۰۵ of the threshold. One proposition fails: when pathways scale differently in exposure, the direction of transition becomes exposure-dependent for up to ۱۴.۵% of configurations. Complementarity and reskilling outweigh entrepreneurial absorption as levers by two to four times.

کلیدواژه ها:

artificial intelligence ، future of work ، job recomposition ، entrepreneurial ecosystems ، generative AI ، SMEs ، reskilling ، scenario analysis