The 17th annual NNN Conference took place in Kigali, Rwanda, from 29 September to 1 October 2026. Organised by the Neglected Tropical Disease NGO Network, this year’s conference placed people, systems and long-term resilience at the heart of the NTD elimination agenda.
Carla Rodríguez Cuesta, Co-Scientific Leader and Study Director of SkincAIr at Sherwood Science, presented the project in the workshop “Digital Innovation in Leprosy and NTDs” and in the Leprosy Disease Specific Group session. You can read about SkincAIr’s participation in the conference here.
What we brought to Kigali
SkincAIr presented what our fieldwork research with frontline health workers, dermatologists and policy makers in Ethiopia, Kenya and Senegal had taught the project:
- A diagnosis needs more than an image.
- An AI tool must give reasons, not verdicts, and speak the health worker’s language. They know scabies, leprosy and leishmaniasis, not the term “skin NTD”.
- A diagnosis tool must build on what clinics already do, and it must reach the cases that never come to a clinic.
In addition to these findings, SkincAIr also showed how these lessons shaped the dataset: ethics by design, with consent the patient understands, pseudonymised data and trained collectors. In the prototype, no consent means no camera.
What we took home
The conference confirmed SkincAIr’s appraoch, and it added lessons of its own.
- Persons affected belong at the design stage, through their own organisations. This was the message of the whole conference, from the opening panel to the Leprosy Disease Specific Group. IDEA International, the global network of organisations of persons affected by leprosy and other NTDs, put it plainly: “Include persons affected at the design stage, not only as individuals invited to speak, but through their own organisations, at the point where decisions are made.” SkincAIr already designs with dermatologists and frontline health workers; the voice still missing is that of persons affected, on stigma, mental health and dignity. How to bring their organisations into user testing and design is the first question the project has taken back to its consortium.
- A digital tool lives or dies by adoption, not accuracy. The session’s own framing highlighted this point: “Today is not only about what AI can do, it is about how we make digital innovation work for people, programmes and health systems.” Several countries presented national digital platforms for NTD case management and asked partners to align with one national plan and one monitoring framework, with no parallel data systems. For an AI tool this means fitting into what a ministry already runs, not arriving beside it.
- Data access and ownership must outlive the funded project. SkincAIr’s own recommendation to the field — engage ministries from the start, and make sure the data remain accessible after the grant ends — was echoed across the sessions on digital innovation. Datasets that disappear when a project closes are a loss for everyone who comes after.
- Skin diseases are now a global priority. World Health Assembly resolution WHA 78.15 made skin diseases a global priority, and the Global Action Plan that will govern the field to 2035 is due for adoption in 2027. Its pillars — data for action, workforce training and task-sharing, data systems and digital health — are the frame in which tools like SkincAIr will be judged.
- The cases are not where the cameras are. Programmes for podoconiosis, Buruli ulcer, yaws and other skin NTDs showed again that many patients are found in schools, communities and campaigns rather than in hospital dermatology clinics, and that conditions such as podoconiosis are still routinely mistaken for others. Reaching those cases, and getting the differentials right, is what a diagnostic-support tool is for.
Since May 2026, SkincAIr’s dermatology teams in Kenya, Ethiopia, Nigeria, Senegal and the Democratic Republic of the Congo have enrolled 560 patients and collected 3,358 clinical images across the twelve target skin NTDs and their differentials (figures from the 23 September 2026 export), with consent the patient understands and data that stays protected. The next step is the evidence: publishing how the dataset was built and how the tool will be validated.
If you want to always stay updated about our project, follow us on LinkedIn, Facebook, X and Bluesky!