Background
The NIH All of Us Research Program is a historic initiative that aims to build one of the largest and most diverse health databases. By engaging over 868,000 participants from diverse backgrounds across the United States, All of Us seeks to accelerate research and improve health outcomes through precision medicine. The program collects genomic, clinical, lifestyle, and participant-contributed data to provide a comprehensive resource for researchers.
As AI/ML capabilities expand, there is a growing opportunity to leverage this resource for precision medicine, public health, and biomedical research. However, realizing its full potential requires addressing challenges in data usability, interoperability, and participant engagement across diverse communities.
Building on ACT-IAC’s prior work on incorporating digital health technologies into federal health research, this project will help All of Us define strategies and explore prototypes that align AI/ML innovation with the All of Us mission to accelerate health research and medical breakthroughs by enabling individualized prevention, diagnosis, and treatment across many diseases.
Overall Project Objective
- Strategy Development: Identify and prioritize use cases across program operations, participant communications, and tools for researchers that promote increased efficiency and effectiveness.
- Prototype Exploration: Assess gaps in current data readiness and identify potential prototype tools (e.g., AI feature stores, knowledge graphs, APIs, chatbots, model marketplaces).
- Partnership Framework: Build a pre-competitive model for collaboration with industry, academia, and advocacy groups.
- Responsible AI: Leverage the broad spectrum of backgrounds of the All of Us participant cohort to advance unbiased, safe, and reproducible AI models.
- Outreach & Education: Lower the barrier to entry for researchers, clinicians, and participants through intuitive interfaces, community engagement, and educational resources.
- Change Management: Consider the people, process, and policy adjustments (in addition to technology) needed to advance AI adoption within the All of Us Research Program.
Project Approach
- Landscape Analysis
- Map existing AI/ML tools, prototypes, and commercial platforms relevant to biomedical data.
- Identify gaps in All of Us datasets for AI/ML readiness.
- Use Case Framework
- Collect and categorize use cases from stakeholders (researchers, clinicians, tech companies, participants) and their application to All of Us's three major audiences: program personnel, participants, and researchers.
- Define functionality pillars (e.g., APIs, user portals, mobile applications, feature/model stores).
- Data Strategy and Gaps
- Assess completeness, quality, and reproducibility of datasets.
- Align with AI data standards (e.g., OMOP, NIST AI Risk Management Framework).
- Prototype and Concept Development
- Explore opportunities for pilot tools (e.g., research assistant chatbots, patient-clinician decision-support tools, polygenic risk score utilities).
- Encourage external innovators to build on the platform in a pre-competitive model.
- Governance & Sustainability
- Define business and partnership models (e.g., marketplace concepts, consortium engagement, participant-mediated data exchange).
- Ensure neutrality and unbiased facilitation by ACT-IAC.
Duration and Timeline: 6 months
Outcomes
- Clear roadmap for leveraging All of Us data with AI/ML.
- Identification of high-priority use cases and potential partners.
- Stronger ecosystem of pre-competitive collaboration across industry, academia, and advocacy groups.
- Foundation for future tools, standards, and participant-facing innovations.
Deliverables
- White Paper: strategic framework, opportunities, and recommendations.
- Stakeholder Engagement Sessions: workshops, iterative reviews
- Prototype or Proof-of-Concept identified or developed.
Membership
The project team will include members with expertise in:
- AI/ML research and application in healthcare.
- Hands-on AI/ML development
- Biomedical data standards and interoperability.
- Precision medicine and public health.
- Federal health program management and policy.
This project will accept volunteers based on the above qualifications.
This call for volunteers is now closed.