Abstract: As AI technology becomes increasingly integrated into oil and gas storage and transportation, data governance in AI scenarios faces significant structural contradictions and challenges. In response, a novel data governance methodology has been proposed. Supported by governance organizations and digital intelligence middle platforms, a multimodal data standard system has been developed to innovate AI metadata-driven mechanisms, resulting in a co-evolutionary "double helix" model in which data governance empowers AI applications, and AI technology, in turn, enhances the governance system. Within this model, the "cognitive helix" utilizes business knowledge graphs to construct high-quality datasets, while the "governance helix" leverages metadata to drive proactive governance for continuous evolution. This methodology and its iterative implementation, in AI scenarios based on innovative practices, offer a reusable theoretical framework and implementation path for data governance in the era of intelligent technology.