Essay by Stefaan Verhulst: “Today, the question facing governments is no longer whether they should use data. Especially in an age of Artificial Intelligence, that debate is long settled. The harder and more urgent question is how to govern data in ways that are trusted, durable, and fit for increasingly complex societal challenges. In short – how to make government data initiatives more effective and legitimate at the same time?
This question is where data stewardship steps in. Within a government setting, data stewardship is the practice of governing public-sector data as a shared civic asset, one whose value depends not only on technical performance but on legitimacy and institutional accountability. It begins from a recognition that data is a social artifact, embedded in social, political, and cultural processes. Only such a lived, non-technocratic approach can help local and state governments navigate the trade-offs, uncertainties, and value conflicts that increasingly define public-sector data use.
Limitations of the Current Approach
Too often, data strategy is framed as a technical exercise. But the shortcomings of a technocratic approach leave state and local data and innovation officers navigating tensions that technical solutions alone cannot resolve: promoting data sharing and reuse while also safeguarding privacy and civil liberties; delivering innovation and efficiency even as public skepticism and trust deficits deepen; and responding to growing pressure to deploy AI systems despite unclear lines of accountability and uneven institutional capacity.
These are not engineering problems but governance challenges that require judgment, legitimacy, and sustained institutional stewardship. While current orientations toward data governance, focusing on risk avoidance and compliance, continue to have an important role, today’s data requires socially embedded governance that centers legitimacy and public trust as facilitated by data stewards.
The Vital Role of Data Stewards
Data stewards move beyond today’s limited, technocratic approach, instead advancing—and embodying—a principle of legitimacy rather than merely compliance. This shift reframes the questions that guide data practice, from whether a particular use is legally permissible to whether it is socially appropriate, publicly understandable, and institutionally accountable…(More)”.
