There are now multiple calls for data — and increasingly AI — governance to become more participatory. Citizens’ assemblies, community panels, deliberative processes, co-design exercises, and other forms of engagement are routinely proposed as ways to build trust, surface public expectations, and establish a “social license” for how data and AI are developed and used.

This is a direction we have ourselves argued for and sought to advance in our work on social license, data stewardship, and participatory approaches to responsible data re-use. It reflects an important shift: from treating people primarily as data subjects whose rights must be protected toward recognizing them as participants who may have a legitimate role in shaping the purposes, conditions, and safeguards governing data use.

Yet precisely because participation is becoming more widely embraced, it is important to ask a more difficult question: when does participation actually change anything? There is a fundamental distinction that is too often overlooked: being heard is not the same as having power. People may be invited to deliberate, consulted about their preferences, or given opportunities to voice concerns, without those contributions materially affecting what happens next. Participation can expand voice while leaving the underlying distribution of decision-making authority untouched.

The challenge, then, is not simply to create more opportunities for participation, but to understand when participation generates leverage: when those affected by data and AI have a meaningful ability to shape purposes, alter decisions, impose conditions, demand accountability, or ultimately say no. Without such leverage, participation risks becoming procedural rather than consequential, a mechanism for demonstrating inclusion without redistributing meaningful agency.

Our new working paper: From Voice to Leverage in Data Governance: Building Enforceable Social License for Data Re-Use, focuses on what happens after participation ends, and what generates leverage. What makes community input actually change the behavior of a company, public authority, or data intermediary?

We argue that meaningful social license requires social leverage: the capacity to make community preferences costly to ignore through credible legal, institutional, contractual, economic or reputational consequences. We define leverage through three properties: it must be counterfactual, enforceable and durable.

The paper develops this argument in three further directions.

First, leverage depends on legibility: people cannot contest systems they cannot see, understand, or monitor.

Second, we identify six sources of leverage: legal and regulatory, institutional, contractual, economic, data-access, and political or reputational. We then examine how intermediaries such as data trusts, cooperatives, and coalitions can aggregate these sources of leverage.

Third, we propose a practical framework for testing whether participation really redistributes power, based on obligation, enforceability, consequence, monitoring capacity, revocability and institutional durability.

Examples ranging from Digital Trust for Places and Routines (DTPR) and equitable data licensing to Indigenous Data Sovereignty show what different combinations of legibility and leverage can look like in practice.

The underlying proposition is simple: a social license worth the name should function more like a license: conditional, monitorable, and capable of being withdrawn.

The paper is an invitation to move the debate from: how do we give communities a voice? To a harder question: what gives that voice consequences?

The working paper is available here (ArXiv) and here (SSRN).

Back to the Blog