A new foundation for secure AI systems in the public interest
Across healthcare, social services, and community-based support systems, critical decisions are made every day using incomplete and fragmented data or overlook valuable lived expertise that is hard to quantify. The people most affected—those navigating housing instability, chronic illness, economic precarity, or systemic bias—are the least represented in the systems designed to serve them.
Project Resolve begins from a simple premise:
We cannot fix outcomes without fixing the systems that determine whose knowledge counts, whose data are trusted, and who benefits from the future of AI innovation.
The problem we’re solving
Today’s data infrastructures are not built for the realities of community care or the future of secure, equitable AI:
- Fragmented systems split insights across institutions that rarely interoperate
- Invisible labor—from community health workers to advocates—is undercounted or lost entirely
- Context-poor data strips away the social conditions that shape health, wellbeing, and opportunity
- Extractive models take data from communities without returning value, agency, or control
- Privacy tradeoffs force organizations to choose between protecting people and learning from data
These gaps limit our ability to respond to crises, reinforce inequities, and constrain innovation—and they shape the limits of the data that emerging AI systems rely on.
Our approach
Project Resolve is a co-design and co-development initiative building a new model for how sensitive data can be created, governed, and used—one that centers trust, accountability, and collective benefit.
We work alongside community-based organizations, care networks, and researchers to:
- Design community-governed data infrastructures that reflect lived realities—not just institutional needs
- Develop privacy-preserving systems that enable meaningful insight without compromising dignity or safety
- Capture the full spectrum of care—from formal services to informal support, relationships, and local expertise
- Reduce administrative burden so frontline workers can focus on people, not paperwork
- Create shared, interoperable tools that strengthen collaboration across fragmented ecosystems
Rather than treating data as a static resource, we treat it as shared knowledge and a collective asset—one that communities must be equipped to steward, not extracted.
What makes Resolve different
Project Resolve operates at the intersection of social science, computing, and community practice. It brings together interdisciplinary expertise to rethink the foundations of digital infrastructure and AI systems:
- Centering community knowledge and governance, not just institutional authority
- Bridging qualitative insight and large-scale data systems, recognizing the value of what often goes unrecorded
- Building open, adaptable systems that can scale across contexts without erasing local specificity
- Advancing new models of shared data stewardship that align ethics, security, and equity
These foundations are also essential for the future of AI—where the reliability, fairness, and accountability of models depend on how data is generated, governed, and shared. Resolve contributes to this future by ensuring that the data underpinning these systems is more complete, contextual, collectively governed, and accountable.
Our goal is not just better tools—but a shift in how institutions collaborate with communities to produce knowledge and deliver care.
Toward a new data ecosystem
Project Resolve is part of a broader effort to reimagine how data systems can support health equity, social resilience, and democratic participation. By aligning technology design with community priorities, we aim to create infrastructures that are not only more effective – but more just – and better equipped to support the next generation of responsible, trustworthy AI technologies.
Partnerships
Project Resolve works in close collaboration with organizations leading innovation in community care and shared data stewardship, including Health Leads (opens in new tab) and flok (opens in new tab).