As artificial intelligence becomes increasingly embedded in development efforts worldwide, it is often powered by vast quantities of data–much of it generated by individuals and communities in low- and middle-income countries. Yet despite contributing this data, these communities rarely have the opportunity to shape how it is used, reused, or governed.

Current data governance frameworks–particularly those built around individual consent—have not kept pace with the complexity of AI systems. Consent models place a disproportionate burden on individuals, overlook the collective nature of many datasets, and offer little opportunity for long-term oversight or adaptation. In practice, they fail to ensure meaningful agency for those most affected by data reuse.

Digital Self-Determination (DSD) calls for a different approach–one where individuals and communities are not just data subjects, but active participants in governing how data about them is collected, shared, and reused. It’s about restoring agency and ensuring that data governance reflects local values, collective interests, and evolving priorities.

A new report from The GovLab and Agence Française de Développement–Reimagining Data Governance for AI–introduces a promising pathway to make DSD real in practice: social licensing.

Screenshot 2025 05 21 at 16.49.2

Social licensing offers a structured, participatory framework that allows communities to define and negotiate the conditions under which data about them is reused, especially in AI systems. Unlike traditional consent, which is often limited to one-off, individual agreements, social licensing is a collective, ongoing process—one that enables communities to shape how their data is used over time.

The report presents a Social Licensing Toolkit, which supports this process through three practical steps:

  1. Establishing community preferences through inclusive engagement.

  2. Documenting those preferences in clear, enforceable agreements.

  3. Exploring ways to uphold and enforce them, through legal tools, oversight models, and emerging technologies.

What Will It Take to Scale Social Licensing?

To ensure social licensing becomes a viable and lasting part of data governance, the report outlines several key steps:

  • Test it in the real world: Launch pilot projects across different sectors to refine the approach and build a solid evidence base around what works–and what doesn’t.

  • Rethink enforcement: Develop creative compliance strategies that blend legal agreements, community-led oversight, certification schemes, and tech-enabled tools to ensure that community-defined terms are respected.

  • Invest in capacity building: Equip communities, policymakers, and practitioners with the knowledge and tools needed to engage meaningfully in participatory processes around data reuse.

  • Create a dedicated hub for learning and support: Establish a Center of Excellence to offer technical assistance, surface best practices, and advance the field through ongoing research.

  • Incentivize responsible data practices: Integrate social licensing into the policies, funding criteria, and procurement guidelines of development programs to encourage adoption at scale.

  • Keep learning: Support continued research to further develop the social licensing model and explore how it can evolve to meet emerging needs and challenges.

At its core, social licensing is about shifting data governance from extraction to collaboration–from decisions made about communities to decisions made with them.  As AI continues to influence and inform development efforts around the globe, social licensing provides a way to operationalize DSD and build data systems that are fairer, more transparent, and more aligned with the needs and priorities of the people they affect.