When Generative Artificial Intelligence Creates and Erodes Marketing Advantage

Abstract

Generative Artificial Intelligence (GAI) has moved from experimental tool to marketing infrastructure, but the emerging evidence is inherently mixed: it can raise content, service, and knowledge-work productivity while also creating risks of homogenization, authenticity loss, factual error, privacy exposure, and weakened organizational learning. This paper asks when GAI adoption becomes a sustainable marketing capability rather than a scalable source of strategic value leakage. Using an integrative conceptual review across marketing, strategy, information systems, service, and organizational research, the paper synthesizes evidence through dynamic-capability and human–machine complementarity lenses. The synthesis yields the Generative AI Marketing Paradox: as automation intensity rises, operational output initially improves faster than strategic marketing value; beyond a context-dependent threshold, additional automation can erode differentiation and trust unless firms possess four complementary capabilities—proprietary market grounding, human–AI orchestration, experimentation and learning, and responsible governance. These elements are formalized as Generative Marketing Capability (GMC). The paper develops six testable propositions and a four-stage maturity path from isolated tool experimentation to a governed, proprietary learning system. The framework reconciles productivity findings with consumer-side authenticity and governance concerns and explains why identical general-purpose tools can produce divergent firm outcomes. For managers, the central implication is that competitive advantage is unlikely to reside in access to GAI itself; it resides in the routines that ground, supervise, learn from, and govern its outputs.