AI in Community Health.

Setting the terms for how artificial intelligence enters community health.

Artificial intelligence is arriving in community-based primary health care faster than the systems meant to hold it.

Diagnostics, logistics, and decision-support tools are being piloted across low- and middle-income countries – often ahead of the supervision, supply chains, data systems, and governance that would let community health workers (CHWs) use them safely.

You cannot AI your way out of a broken system. AI should support CHWs, never replace them.

CHIC’s role is to make sure this wave strengthens care rather than repeating the fragmented failures of past digital health cycles.

We are not a technology vendor and we are not chasing pilots.

As the coalition connecting CHWs, implementing organizations, funders, and governments across more than 60 countries, we are working to establish the governance, evidence, and workforce foundations that determine whether AI adds value — or simply exposes fragility.

Three strands make up this work: peer-reviewed research, convening the field, and building a shared evidence base.

CHW Maureen visits a household in Migori County, Kenya © BBC Storyworks for Medic

A Governance-first Framework

Research.

In March 2026, The Lancet Primary Care published “Governance, scale, and integration: building community health worker systems ready for artificial intelligence,” authored by Community Health Impact Coalition Research Group (Volume 2, Issue 3; open access).

Drawing on CHIC’s survey work with CHW programs, it sets out three system-level conditions that must be met for AI to move from promise to durable public value:

Public stewardship: AI tools for CHWs should be owned and governed as public infrastructure, with named units, budget lines, and clear approval pathways, rather than run as parallel private pilots with little more than government notification.
Collective learning over isolated pilots: the field needs shared standards, interoperable data systems, and an evidence-to-action cycle that compounds learning across implementers, instead of long-cycle trials that entrench non-interoperable silos.
System and workforce readiness for scale: AI only helps if it reaches a salaried, skilled, supplied, and supervised CHW workforce equipped with the right hardware. Without that foundation, we leave what could be the world’s largest care-delivery network unplugged.

The Comment closes as an explicit invitation to governments, funders, implementers, and researchers to align around stewardship, readiness, and professionalized CHWs.

“The real danger is not failure, but scaling on terms that undermine care.”

Insights from “Governance, scale, and integration: building community health worker systems ready for artificial intelligence.”

The largest survey of AI in CHW programs to date

Evidence.

AI is moving fast in global health conversations. What does it look like on the frontlines? CHIC’s multi-stakeholder, cross-sectional international survey of 283 respondents across 41 countries, including 157 CHWs and 28 supervisors, is among the first to capture the full range of CHW-program stakeholders, from frontline workers to implementers and governments, in a single view. The picture is one of a field aware of AI but overwhelmingly pre-scale.

Awareness far outpaces maturity. 67% (189) were aware of AI tools deployed in CHW programs, but only 11%(31) described mature or at-scale use, and fewer than a third (31%) rated system, data, and workforce readiness as high.
Frontline workers are already using AI without support. Among 185 CHWs and supervisors, 76% (141) reported using tools they identified as AI-enabled — yet only 19% (36) received ongoing training.
There is strong consensus on how AI should be used. 81% (230) agreed AI should support rather than replace CHW judgement, and 75% (211) wanted clear escalation pathways — but only 14% (39) considered national AI policy clear.

The survey points funders and governments toward the priorities that matter: CHW-centred design, national AI-for-health policy, continuous training and supervision, reliable escalation pathways, data protection, interoperability, and infrastructure built for low-connectivity settings.

CHW–AI workshop at the Skoll World Forum

Convening.

On 23 April 2026, alongside the Skoll World Forum, CHIC convened 80 participants — implementing NGOs, technology partners including frontier AI labs, and funders — for a four-hour working session on the intersection of CHWs and AI.

The day was opened by Bernardo Xavier (CHW, Brazil), Madeleine Ballard (CEO and Founder, CHIC), Rebecca Hope (Director of Global Health, Emerson Collective), and Emilie Chambert (CEO, Living Goods), who set three framing truths that carried through the discussions: conditions, not just code (“you can’t AI your way out of a broken system”); support, not replace; and from pilots to pillars — a shift from proliferating pilots toward durable, government-owned infrastructure.

The evidence shared in the room reinforced the point that infrastructure matters more than the algorithm. In an IDinsight/Last Mile Health deployment in Ethiopia, 91% of health extension workers named poor connectivity as their primary barrier and 48% flagged electricity. Where the foundations were in place, the gains were real: AMREF’s CHA Reporting Assistant in Machakos County, Kenya, saved 30+ minutes per CHW per reporting cycle, surfaced new health patterns in 76% of cases, and measurably raised confidence in using data.

Participants converged on three asks of the field:

Shared frameworks and principles (due-diligence checklists, governance minimums, common evaluation criteria) rather than more pilots.
Better evidence and learning systems, including pooled multi-organisation evaluation and CHW-rated quality measures beyond simple usage metrics.
More structured bridging conversations between funders, governments, and technology partners.

The strongest signal was appetite for what comes next — participants asked to co-author guidelines, join a community of practice, and help stand up a shared learning initiative.

WHAT’S NEXT

The field does not need more isolated pilots. It needs shared standards and a common learning agenda.

Building on both, we are:

Convening a CHW–AI evidence and learning collaborative: a standing coalition of implementers, funders, governments, and technologists to harmonize a shared learning agenda, pool evaluation, and agree what “good” looks like.
Developing shared publications and CHW–AI governance minimums: evaluation criteria, and procurement guidance that governments and funders need.
Keeping the foundations front and center because AI is only as good as the salaried, skilled, supplied, and supervised workforce that uses it.

We are looking for funders and partners who want to help set the terms for how AI enters community health — not to launch another pilot, but to build the stewardship, evidence, and workforce foundations that make the technology worth deploying.

There are concrete ways to plug in:

joining the evidence and learning collaborative
funding synthesis and guideline development
or co-authoring research and policy outputs.

If you want to help set these terms, reach out. Email us at info@joinchic.org to explore joining the evidence and learning collaborative, funding synthesis and guideline development, or co-authoring research and policy outputs.

Get in touch: info@joinchic.org
Stay informed: Sign up for CHIC’s newsletter for updates on this work and the wider proCHW movement.

CHIC CHW and AI workshop

Setting the terms for how artificial intelligence enters community health.

Research.

Evidence.

Convening.

WHAT’S NEXT

Join the movement.

Social change is a team sport. Subscribe to our mailing list and join the movement.