Socioeconomic Challenges of Artificial Intelligence-Driven Digital Transformation: The Functional Mechanism Model

Abstract

Artificial intelligence-driven digital transformation has become a major source of contemporary socioeconomic challenges because firms, sectors, and public authorities need to understand not only whether artificial intelligence is adopted, but also which economic functions it performs and under what organizational conditions it may create durable value. Existing theoretical accounts remain fragmented across sector-specific applications, technology-centered classifications, and general adoption frameworks, leaving no unified explanation of how functionally different artificial intelligence investments may produce distinct return profiles across firms and industries. This study develops the functional mechanism model as a theoretical concept for analyzing the economic contribution of artificial intelligence through a function-based rather than sector-based logic. The model classifies artificial intelligence applications into productivity, information, and innovation functions; links them to cost and cycle-time reduction, reduced information asymmetry, and the creation of new products, services, or decision pathways, respectively; and incorporates governance capacity, competitive pressure, and stakeholder trust as conditions shaping realized contribution. The proposed scientific novelty lies in integrating these elements within one framework that treats artificial intelligence adoption as a portfolio of functionally differentiated investments and derives four explicit propositions for subsequent empirical testing. In contrast to broad adoption frameworks, the model distinguishes economic mechanisms within the same technology and provides a formal structure for examining functional composition, temporal variation, organizational governance, and environmental conditions without treating adoption intensity as sufficient to explain economic outcomes. The model predicts a temporal reversal in function-specific returns: productivity-function returns are initially higher but decay through imitation, whereas innovation-function returns accumulate linearly until, beyond a threshold horizon at which the two return coefficients become equal, innovation becomes relatively dominant. It further predicts that governance capacity has a proportionally larger marginal effect on realized returns for innovation-weighted portfolios than for productivity-weighted portfolios, while stakeholder trust operates as a threshold-like condition below which otherwise attainable economic contribution is sharply constrained. The functional mechanism model therefore provides a basis for future panel, survey, and quasi-experimental testing and offers a structured framework for examining productivity disparities, labor-market adjustment, unequal organizational capability, and regional digital divides as socioeconomic challenges associated with artificial intelligence-driven transformation.