Insights from the First Data Governance Forum: Challenges, Opportunities, and the Role of Large AI Models

The forum, organized by the Shanghai Data Science Key Laboratory and DAMA China, gathered experts from Ant Group, Bright Food, Hong Kong University of Science and Technology, Renmin University, and legal professionals to discuss the latest trends and challenges in data governance.

Key observations included the closed-loop relationship between large AI models and data governance: high‑quality data fuels better models, while advanced models improve data analysis and quality management. Speakers stressed that data governance is not merely a technical task but requires strategic, structural, and organizational alignment.

Industry practitioners highlighted practical obstacles such as fragmented data sources, inconsistent quality, and the difficulty of integrating multi‑granular data, noting that effective governance demands over 70% management and communication effort, especially in traditional large enterprises.

Academic voices pointed out that data fusion and cleaning are core to the data lifecycle and that future large‑model applications can significantly empower these processes, though challenges remain in model stability, interpretability, and the need for knowledge graphs and vector databases.

Legal experts warned about compliance hurdles in data circulation, emphasizing the tension between data openness and privacy, the lack of unified regulations for public data, and the need for robust standards to enable safe data markets.

Panel discussions identified four governance layers—strategic, structural, mechanism, and technical—and argued that genuine data governance requires clear business drivers, cost‑benefit visibility, and long‑term commitment, rather than being treated as a low‑value “dirty work.”

Overall, the consensus was that data governance is a systemic engineering challenge that intersects technology, law, economics, and societal responsibility, and that large AI models can both benefit from and accelerate better data governance practices.