Data Commons: Building Infrastructure for Peace Technology
Marine Collins Ragnet and Katerina Siira of NYU’s Peace Research and Education Program are building participatory data infrastructures to enable inclusive, AI-ready humanitarian technologies.
When disasters strike multilingual communities or conflicts emerge in remote regions, the inability to communicate across language barriers can cost lives. This challenge sits at the heart of a broader problem: while artificial intelligence could revolutionize humanitarian response and conflict prevention, it requires vast amounts of contextual, representative data that simply doesn't exist for most crisis-affected populations.
Data commons—shared pools of data resources that are responsibly governed and made available for public benefit—are emerging as critical infrastructure for what researchers call "peace technology." These digital tools and platforms designed to prevent conflict and support humanitarian action depend entirely on access to relevant, timely, and representative data.
The Data Gap in Humanitarian AI
The paradox is striking: we live in an age of unprecedented data abundance, yet the specific information needed for humanitarian and peace applications remains largely inaccessible. Commercial AI systems train on massive internet datasets, but peace technology applications struggle to access the localized, sensitive, and contextual data they need.
As researchers have noted, this "data winter" for public interest AI threatens to leave vulnerable populations behind in the technological revolution.¹ Traditional approaches—whether proprietary databases controlled by single organizations or unstructured open data dumps—fail to address the unique requirements of peace and humanitarian applications.
Consider crisis response in Sub-Saharan Africa, where over 2,000 languages are spoken. Current emergency systems typically operate only in official languages, effectively silencing rural and marginalized communities who are often most vulnerable to disasters. This linguistic exclusion represents just one dimension of a broader data infrastructure crisis.
A New Approach: The Commons Model
Data commons offer a structured alternative that balances accessibility with responsibility. Unlike traditional data repositories, commons establish participatory governance structures that give communities agency over their information while enabling its use for broader social good.
This approach has gained significant momentum with initiatives like the New Commons Challenge, launched by the Open Data Policy Lab (a collaboration between Microsoft and The GovLab). The challenge, which awarded $100,000 grants to develop data commons for humanitarian response and localized decision-making, represents a critical investment in peace technology infrastructure.
From Theory to Practice: The Malawi Voice Data Commons
Among the winning initiatives, NYU's Peace Research and Education Program’s Malawi Voice Data Commons exemplifies how data commons can transform humanitarian response. Led by researchers Marine Collins Ragnet and Katerina Siira, the project addresses the critical challenge of multilingual crisis reporting in Sub-Saharan Africa.
The initiative, developed in collaboration with Ushahidi, UNDP, and the Mozilla Foundation, will create a revolutionary approach to emergency communication. Rather than forcing communities to adapt to English-only systems, it will enable rural Malawians to report emergencies in their native languages—including Chichewa, Chitumbuka, and Chiyao—while creating AI-ready datasets that can train future humanitarian technologies.
"When floods approach or violence erupts, language barriers can be deadly," explain Collins Ragnet and Siira. "The Malawi Voice Data Commons allows rural Malawians to report emergencies in their own languages while preserving cultural heritage. By making these voices AI-ready, we are ensuring that humanitarian response is faster and more inclusive."
The $100,000 award will support pilot implementation in Malawi, with plans to scale across Sub-Saharan Africa—demonstrating how targeted investment in data infrastructure can have continental impact.
Technical Architecture for Peace Applications
The Malawi Voice Data Commons design illustrates several key principles that distinguish peace-focused data infrastructure:
Community Participation: Local communities will maintain control over how their voice data is collected, processed, and used. This participatory governance model ensures that data collection serves community needs rather than extractive research agendas.
Interoperability: The system will integrate with existing humanitarian infrastructure used by UN agencies and international NGOs, ensuring that innovations enhance rather than fragment response capabilities.
Scalability: The technical architecture and governance framework are designed for replication across diverse contexts, recognizing that effective peace technology must work across boundaries.
Dual Purpose: Beyond immediate crisis response, the commons will contribute to language preservation and documentation—acknowledging that cultural heritage and humanitarian effectiveness are interconnected.
Challenges and Governance Considerations
Creating effective data commons for peace technology requires addressing several critical challenges:
Trust Building: In conflict-affected areas, communities may be justifiably suspicious of data collection. The Malawi project's planned emphasis on community control and clear benefit-sharing offers an approach to building necessary trust.
Sustainability: Peace-focused commons require funding models that balance long-term infrastructure needs with donor priorities. The New Commons Challenge's provision of both funding and incubator support—including mentorship and technical assistance—offers a model for sustainable development.
Technical Capacity: Many regions requiring peace technology lack basic digital infrastructure. Solutions must function with limited connectivity while maintaining security and governance standards.
Ethical Safeguards: Data commons must implement robust "do no harm" principles, ensuring that shared information cannot be weaponized against vulnerable populations.
Building the Future of Peace Technology
The recognition of projects like the Malawi Voice Data Commons signals growing understanding that effective AI for peace requires purposeful infrastructure development. As Andrew Schroeder of Direct Relief notes, "From disaster relief to public health, the fastest path to impact is making frontline data findable, trustworthy, and reusable."²
The path forward requires coordinated effort across sectors. Technologists must prioritize community control and responsible use. Funders must support long-term infrastructure rather than just short-term projects. Peace practitioners must engage with technical development to ensure tools meet real needs. Most critically, affected communities must be centered in all aspects of commons design and governance.
For academic institutions like NYU, this represents an opportunity to bridge research and practice, developing theoretical frameworks while creating practical solutions. The Peace Research and Education Program's success demonstrates how universities can lead in building the data infrastructure necessary for effective, inclusive peace technology.
Conclusion
As artificial intelligence reshapes global systems, ensuring its benefits reach conflict-affected and crisis-vulnerable populations requires deliberate infrastructure development. Data commons offer a pathway toward peace technology that serves rather than surveils, includes rather than excludes, and empowers rather than extracts.
The Malawi Voice Data Commons and similar initiatives demonstrate that alternative data futures are possible—ones where communities maintain agency over their information while contributing to technologies that could save lives. With continued investment and commitment to participatory governance, data commons can become the foundation for a new generation of peace technology that truly serves those who need it most.
Learn more about the New Commons Challenge.
References
Verhulst, S., Davis, B., & Schroeder, A. (2025). "Data Commons: The Missing Infrastructure for Public Interest Artificial Intelligence." Open Data Policy Lab Opinion Piece.
Open Data Policy Lab. (2025). "Building the Future of Public-Interest AI." Press Relea