This community innovation has been accepted at the 2026 DHIS2 Annual Conference and will be included a session.
Beyond the Hype: Where Generative AI Can (and Cannot) Add Value to DHIS2
Over the last years, GenAI has generated significant interest within digital health, including the DHIS2 ecosystem. Following earlier proofs of concept that SolidLines developed and tested with several organizations (eg. PAHO, UiO, MSF, Girl Effect) for multimodal data ingestion using image and voice processing, one of the most important challenges identified was the ability to assess the potential for GenAI to realistically improve DHIS2 workflows — and where it should not be applied — as part of a far more complex system of actors, users, business processes and socio-technical factors.
In this presentation, SolidLines shares its perspective as DHIS2 experts and GenAI practitioners, focusing on long-standing bottlenecks in health information systems experienced by key DHIS2 user archetypes:
- Data entry user: We will discuss practical GenAI use cases particularly around data entry and self-reporting, where configuration-driven chatbots approaches can reduce user burden while preserving structured data capture.
- Data analysis user: We will also briefly explore adjacent areas such as assisted analytics using free text and automated reporting.
- System admin: We will explore some Interoperability scenarios (eg. between FHIR and DHIS2, between different DHIS2 instances, etc.)
- Implementer: We will discuss exploratory approaches where GenAI supports the automatic generation of resources such as program indicators and program rules based on natural language descriptions.
At the same time, the adoption of Generative AI in DHIS2 raises critical questions that go beyond technical feasibility like costs, vendor dependency, data protection, etc. The presentation will explicitly talk about these risks, discussing how design choices around architecture, deployment, and governance can either reinforce or undermine local ownership, sustainability, and trust in health information systems.
The session will also present some live demos.
Primary Author: Carlos Tejo-Alonso
Keywords:
Generative AI, DHIS2, Data Entry, Data Use, local ownership