AI Global Health: Policy, Evidence and Financing for Responsible AI (2026-2027)

Background

AI tools are increasingly used in clinics and public health programs across low-and middle-income countries (LMICs) because there are not enough clinicians, laboratory staff or financial resources to meet population needs. Algorithms support triage, guide treatment, predict outbreaks and help health systems allocate scarce resources. Yet this rapid adoption is outpacing the protections that typically ensure safety, accountability and equitable access.

Three urgent gaps stand out. First, policy: while many countries now have national AI strategies, almost none provide ministries of health with concrete guidance on safety, bias, accountability or liability. In an analysis of 50 national AI policies across LMICs in Africa and Asia, only India and Vietnam directly referenced health. Second, evidence: governments and funders are being asked to approve or scale AI tools without clear, timely proof that they improve outcomes, reduce burden on health workers or advance equity. Third, financing: most AI pilots rely on short-term donor funding, with no durable path to sustained use.

This project will help close these gaps by producing three decision tools: a Responsible AI in Health Policy Toolkit, a Real-Time Evidence Framework and dashboard prototype and a Financing and Scale Playbook.

Project Description

This applied strategy lab is embedded within Duke’s AI for Global Health initiative and is guided by an advisory group that includes PATH, Pluto Health, Villgro Africa, Duke-Margolis and other partners. The team will focus on real policy, evidence and financing decisions facing ministries of health and implementing organizations.

The project includes three interconnected workstreams:

Policy and Ethics — Responsible AI in Health Toolkit
The team will compare AI and digital health policies across multiple LMICs (Kenya, Rwanda, Malawi, Bangladesh, Vietnam and India) along with several upper middle income country/high income country comparators (Brazil, Colombia, Singapore, the U.S.). Deep partnerships with Society for Family Health and the Rwanda Ministry of Health, and with Penda Health and the Kenya Ministry of Health, will allow the team to pressure-test model policy clauses related to safety, bias, accountability and workforce implications.

Evidence and Impact — Rapid Evidence Framework and Dashboard
The team will define key impact domains — clinical outcomes, safety, access, workflow, equity and cost — and will analyze implementation data at 6, 12 and 18 months from six AI-enabled health innovators. The work will produce a rapid evidence framework and a decision dashboard prototype that supports “go / no–go / scale” decisions. During Summer 2026, a Data+ project will convert this prototype into a production-grade tool.

Financing and Scale — Playbook for Sustainable Adoption
The team will analyze how AI tools are financed today, drawing from more than 350 financing commitments across donors, ministries and private-sector actors. The resulting playbook will outline viable pay-for-use and sustainability pathways, including procurement, reimbursement and blended finance options.

The project uses capacity-driven scoping — determining which tasks will generate the most value for ministries and funders given available time, expertise and partner needs. Weekly full-team meetings, weekly workstream sessions, monthly partner interactions and quarterly advisory group reviews will guide progress throughout the year.

Anticipated Outputs

• Responsible AI in Health Policy Toolkit
• Rapid Evidence Framework and prototype decision dashboard
• Financing and Scale Playbook
• Partner and advisory briefings with ministries of health, funders and implementers

Student Opportunities

The project seeks 3 graduate or professional students and 3–6 undergraduate students across global health, law/ethics/policy, data science, engineering and business. Students from LMIC backgrounds and those with multilingual skills are especially encouraged to apply.

Students will gain experience in:

• Mapping AI policy environments and drafting governance language that ministries can adopt
• Conducting interviews with innovators, funders and implementing partners
• Analyzing multi-country implementation data to assess outcomes, safety, workflow implications, equity and cost
• Co-designing a rapid evidence framework and dashboard
• Understanding real-world financing flows and building a Financing and Scale Playbook
• Translating research into briefings and actionable decision tools for ministries and funders

Graduate and professional students will co-lead workstreams, manage partner communication and help guide analytical quality and deliverables.

In Fall 2026, this team will meet on Thursdays from 10:30-11:30 a.m.

See the related Data+ project for Summer 2026; there is a separate application process for students who are interested in this optional component. 

Timing

Summer 2026 – Spring 2027

Summer 2026 (optional):

• Clean and structure partner data
• Build initial dashboard prototype
• Confirm impact indicators with ministries and funders

Fall 2026:

• Team onboarding and scoping
• Policy environment mapping
• Rwanda-focused deep dives with Society for Family Health and the Ministry of Health
• Draft rapid evidence framework
• Map financing options
• Submit IRB or data access requests as needed

Spring 2027:

• Deliver Responsible AI in Health Toolkit
• Deliver Financing and Scale Playbook
• Update dashboard and reporting template
• Conduct partner briefings and refine outputs

Crediting

Academic credit available for fall and spring semesters

See the related Data+ summer project, Policy, Evidence and Financing for Responsible AI (2026).