Malaria in 2026
In 2026, malaria remains a major global public health challenge. Despite substantial progress over the past two decades, gains have been fragile and uneven and progress has stalled. Estimated global malaria cases rose from 241 million in 2020 to 249 million in 2022, with deaths increasing alongside. The WHO African Region bears the highest burden, accounting for nearly 90% of cases.
The reasons for this stagnation are complex: changes in transmission hotspots and high-risk populations; source–sink dynamics driven by human mobility; gaps in spatiotemporal data coverage; changes in vector behaviour following sustained interventions; the emergence of drug-resistant parasites and insecticide-resistant mosquitoes; altered dynamics under a changing climate; and declining global health funding.
Read the full World Malaria Report 2025 (WHO) →
The challenge in Zanzibar
Two decades of intensive malaria control in Zanzibar produced a rapid initial decline in transmission, but progress has stagnated. The annual parasite incidence has remained stubbornly persistent, characterised by seasonally and geographically heterogeneous low-level transmission.
Several overlapping challenges explain this persistence. Transmission is increasingly outdoors, concentrated among mobile high-risk populations, with a high proportion of imported infections through homan mobility between Zanzibar and Tanzania. Surveillance and targeted interventions aimed at outdoor transmission foci, population hotspots and importation from mainland Tanzania can help to overcome some fo these challenges.
What we do
MalarAI develops and applies high-resolution risk mapping, spatiotemporal modelling, and decision-support tools that integrate surveillance, climate, mobility, genomics, and environmental data to support targeted, evidence-based malaria elimination. Our pilot is focused on Zanzibar, in direct partnership with ZAMEP, with a framework designed to be transferable to other pre-elimination settings.
We are building an AutoML pipeline for near real-time data integration and analysis, enabling ZAMEP to move from reactive to proactive intervention strategies, identifying micro-sources and sinks, mapping importation corridors, and targeting resources precisely where transmission actually occurs.
How we work
Open and reproducible methods with transparent assumptions and explicit representation of uncertainty. All code and data products are published openly.
Partnership-driven research, co-designed with ZAMEP, regional partners, and local collaborators. Workshops with African stakeholders shape the research agenda from the outset.
Actionable outputs, designed for real-world planning, targeting, and evaluation. Dashboards, early warning alerts, and intervention guidance, not just academic papers.
Capacity building: training ZAMEP staff to own, maintain, and interpret AI tools beyond the life of the project. The goal is local ownership, not dependency.
Funding
Our pilot project is funded by the Web Science Institute Pilot Project Fund 2025–26 at the University of Southampton, providing pump-priming support for data landscape assessment, AutoML feasibility testing, and stakeholder co-design workshops. We are actively pursuing full programme funding from a range of international funders.