Generative AI and digital productivity: Transitions, institutional misalignment, and human role redefinition

Abstract

Generative AI is widely framed as a productivity tool. This framing understates its significance: when a technology lowers the cost of a fundamental input to the point of abundance, it disrupts not only the tools of an activity but the evaluation systems and professional cultures built around prior cost structures. This paper argues that generative AI constitutes the third transition in digital productivity, following the internet's reduction of information access costs and open data movements that restructured information flows, and develops five contributions toward understanding its institutional consequences. First, a three-transition model characterizes each transition by the input it makes abundant; the current transition makes structured written output abundant on demand. Second, a digitalization prerequisite thesis identifies machine-readable data foundations as a readiness condition most organizations and many individuals have not yet met. Third, a collaborative intelligence framework maps human comparative advantage shifting toward framing, critical evaluation, and contextual judgment as AI absorbs generative work, revealing that institutional credit concentrates at the production stage precisely where AI advantage is now greatest. Fourth, five structural tensions are identified: productivity surplus distribution and automatable job persistence, educational assessment disruption, AI authorship ambiguity, the professional performance threshold, and epistemic risks from fully AI-mediated communication chains. Fifth, a three-part governance framework (disclose, originate, accountable) provides actionable minimum conditions for AI-collaborative knowledge work across institutions and jurisdictions.

The authors' abstract, as published at the source. Social Sciences & Humanities Open, 2026 · DOI ↗