Governing Trustworthy Generative AI in Enterprise Product Ecosystems

Abstract

Generative artificial intelligence is rapidly becoming a foundational capability within enterprise product ecosystems, enabling new forms of automation, personalization, and decision support. However, its probabilistic behavior, opacity, and deep integration across interconnected products introduce significant trust, risk, and interpretability challenges. This study examines how trustworthy generative AI can be effectively governed in enterprise product ecosystems by adopting an ecosystem-level, socio-technical perspective. Using a mixed-method research design that integrates conceptual modeling, qualitative investigation, and quantitative analysis, the study evaluates the influence of key governance dimensions including transparency, data governance, accountability, model lifecycle management, human-in-the-loop oversight, and regulatory alignment, on trustworthiness outcomes. The results demonstrate that transparency and data governance are the strongest determinants of trust, while ecosystem complexity moderates governance effectiveness through propagation effects across interconnected products. The findings highlight that governance mechanisms operate synergistically rather than independently and must scale systematically with ecosystem complexity. The study contributes a structured understanding of generative AI governance as a strategic enterprise capability that supports responsible scaling, stakeholder confidence, and sustainable innovation in complex product ecosystems.