NALA Foundation Launches AI Pilot with Lemonade to Predict Schistosomiasis in Ethiopia

The NALA-Lemonade AI disease prediction pilot is a collaborative project aimed at forecasting schistosomiasis risks in Ethiopia through the integration of artificial intelligence, satellite remote sensing, epidemiological data, and field validation.

The announcement signals a new era in the application of artificial intelligence within the enterprise sector for addressing global health issues. Organizations like Lemonade Inc. are extending their technological capabilities through their foundation to tackle complex problems such as neglected tropical diseases. This initiative aligns with broader trends where companies are using their AI expertise to support public health initiatives in partnership with specialized nonprofits. The focus on schistosomiasis, a disease caused by parasitic worms that affects millions in tropical regions, highlights the potential for AI to transform how resources are allocated in disease control programs. By predicting risk areas, interventions can be targeted more precisely, potentially saving costs and improving outcomes for affected populations in Ethiopia and beyond. The partnership combines the on-ground experience of NALA with the AI and data analytics strengths from the Lemonade side and academic input from the University of Haifa.

What is the background of NALA Foundation's efforts against neglected tropical diseases?

NALA has spent more than 15 years collaborating with governments and communities to combat neglected tropical diseases. Their work includes detailed mapping of water points where the intermediate host snails are found, collection of snail samples for laboratory analysis, and building a comprehensive understanding of transmission dynamics. This hands-on experience is crucial for the success of the new pilot because it provides the ground truth data needed to train and validate the AI models. Without such foundational work, the integration of satellite remote sensing and epidemiological data would lack the necessary context to produce reliable predictions. The organization's long history in the field ensures that the AI tools developed will be practical and aligned with the realities of disease control in resource-limited settings. Ethiopia was chosen for the pilot due to the presence of schistosomiasis and the existing relationships that NALA has established there over the years.

Schistosomiasis, also known as bilharzia, is a major public health concern in many parts of Africa, including Ethiopia. The disease is transmitted through contact with freshwater contaminated with the parasite's larvae, which are released by specific snail species. Traditional surveillance methods involve monitoring cases after symptoms appear or after transmission has already begun, which often leads to delayed responses and higher treatment costs. The new approach aims to change this by using environmental indicators captured via satellites to identify conditions favorable for snail proliferation and parasite transmission in advance. This proactive stance could allow for preventive measures such as targeted snail control or public education campaigns before outbreaks occur. The partnership brings together diverse expertise to make this vision a reality, with the pilot serving as a proof of concept for wider application.

The experience gained by NALA over its 15 plus years of operation provides invaluable insights into the practical challenges of disease mapping and control. Collecting snail samples requires trained personnel and laboratory facilities, while mapping water points involves community engagement to ensure accurate data. These activities form the bedrock upon which the AI models will be built, ensuring that predictions are grounded in empirical evidence rather than theoretical assumptions alone. The partnership with the University of Haifa brings academic expertise that can help in designing robust study protocols and statistical analyses. Lemonade Foundation's role is to facilitate the technological aspects and provide the necessary funding and project management support to keep the pilot on track within the 12 month timeframe.

What details define the new AI disease prediction pilot in Ethiopia?

The pilot project is set to begin in two specific communities in Ethiopia and will continue for the next 12 months. During this period, the team will develop AI models that incorporate satellite remote sensing data, environmental indicators such as temperature and vegetation indices, epidemiological records, and ongoing field validation through surveys. The goal is to test how well the model predictions match actual transmission events observed on the ground. This iterative process will involve refining the algorithms based on discrepancies between predicted and observed risks. The involvement of the University of Haifa adds scientific rigor to the data analysis and model development phases. Lemonade Foundation provides the support and resources to scale the initial efforts into a functional system. The project is designed to be the starting point for a larger initiative that could benefit other regions facing similar challenges with neglected tropical diseases.

Key to the pilot is the combination of multiple data sources to create a holistic view of disease risk. Satellite data offers broad coverage of environmental conditions that influence snail habitats, while field data from NALA's previous work supplies the detailed local knowledge. Epidemiological data helps to understand historical patterns of infection rates. By fusing these elements using machine learning techniques, the system can generate risk maps that highlight areas most likely to see increased transmission in the coming weeks or months. This allows health authorities to prioritize interventions in high-risk zones rather than spreading resources thinly across all areas. The 12-month duration provides sufficient time to collect data across different seasons, which is important because transmission patterns can vary with weather changes. Success in this pilot could pave the way for similar projects in other countries where schistosomiasis is prevalent.

What are the technical specifics involved in the prediction models?

The technical approach relies on artificial intelligence to process large volumes of data from various sources. Satellite remote sensing provides information on land cover, water bodies, and climate variables that affect the snail vectors. Machine learning algorithms are trained to identify correlations between these environmental factors and known transmission sites. The University of Haifa likely contributes knowledge in statistical modeling and data integration techniques. Field validation ensures that the models are calibrated correctly by comparing predictions with actual snail samples and infection rates collected during surveys. This feedback loop is essential for improving accuracy over time. The long-term vision includes making the models and resulting risk maps available as an open-source resource, which would allow other organizations to adapt and use them in their own contexts without significant additional investment.

Environmental indicators play a central role in the models because changes in these factors can precede actual increases in disease cases by several weeks. For example, increased vegetation or certain temperature ranges can signal optimal conditions for snail breeding. By monitoring these indicators continuously through satellite imagery, the system can issue early warnings. The integration with epidemiological data adds another layer, allowing the model to account for human behavior and previous intervention outcomes. The pilot will test the robustness of these models in real-world conditions in Ethiopia, where data quality and accessibility may vary. This technical validation is a critical step before any broader deployment. The use of open-source principles in the vision ensures that the technology can be scrutinized and improved by the wider scientific community.

Data fusion techniques are critical in this project, requiring sophisticated algorithms to handle the heterogeneity of satellite, field, and epidemiological datasets. Preprocessing steps such as normalizing satellite imagery and cleaning survey data are necessary before model training. The choice of machine learning methods, possibly including random forests or neural networks, will depend on the complexity of the relationships identified in the data. Validation metrics such as precision, recall, and area under the curve will be used to assess model performance during the field validation phase. The open-source release will include not only the models but also documentation on how to adapt them to new regions, promoting wider use and continuous improvement by the research community.

What market and stakeholder implications arise from this enterprise AI initiative?

For enterprises in the technology and insurance sectors, this pilot demonstrates how AI can be applied beyond commercial products to address societal challenges. Lemonade Inc. through its foundation is showing a model for corporate social responsibility that leverages core competencies in data and AI. Stakeholders in public health, including governments in endemic countries, stand to benefit from more efficient allocation of limited resources. NGOs like NALA can enhance their impact by incorporating predictive tools into their programs. Researchers at institutions like the University of Haifa gain opportunities to apply their expertise in real-world settings. The market for AI in global health is growing, and this project could serve as a case study for other companies looking to enter this space. Implications also include potential cost savings in disease control programs, as prevention is generally less expensive than treatment after widespread transmission.

The open-source aspect of the long-term vision has significant implications for accessibility. Smaller organizations and governments in low-resource settings would be able to use the tools without paying licensing fees, democratizing access to advanced predictive analytics. This could accelerate the adoption of AI-driven approaches in public health across Africa and other regions affected by neglected tropical diseases. Enterprise stakeholders may see opportunities for partnerships or technology transfer. The success of the pilot could attract additional funding from international donors interested in innovative solutions to persistent health problems. Overall, the project positions AI as a tool for equity in health outcomes by enabling proactive rather than reactive strategies.

From an enterprise perspective, this project showcases how AI can be repurposed for social impact without diverting from core business objectives. Lemonade's background in using AI for insurance risk assessment translates well to assessing disease risk, creating synergies between commercial and philanthropic activities. Stakeholders such as international health organizations may look to replicate this model in other disease areas. The involvement of multiple sectors also fosters knowledge exchange that can lead to innovative solutions. For communities in Ethiopia, the benefits could include reduced disease incidence and improved quality of life if the predictions lead to timely interventions. The pilot's success metrics will be closely watched by the AI community as an example of applied machine learning in challenging environments.

Overview of the 12-month pilot phases and activities

Phase	Timeframe	Key Activities	Deliverables
Initiation	Month 1	Site selection and data gathering setup	Project plan and baseline data
Development	Months 2-6	AI model building with satellite and epi data	Trained prediction models
Validation	Months 7-10	Field surveys and comparison with predictions	Performance metrics and adjustments
Dissemination	Months 11-12	Results analysis and open-source preparation	Final report and tool release plan

What steps will the pilot follow in its implementation?

1. Select two communities in Ethiopia with known schistosomiasis presence for the initial test.
2. Compile historical and current data on environmental conditions, snail distributions, and infection rates.
3. Train machine learning models using combined satellite imagery and epidemiological datasets.
4. Deploy field teams to conduct surveys validating model outputs in real time.
5. Analyze discrepancies between predictions and observations to refine the algorithms.
6. Prepare documentation and code for open-source release based on pilot outcomes.

What reactions from experts and authorities can be expected?

While specific expert reactions beyond the announcement are not detailed in initial reports, the involvement of established entities like the University of Haifa suggests academic endorsement of the approach. Public health experts in Ethiopia and international organizations focused on neglected tropical diseases may view this as a promising innovation that complements existing control programs. The use of AI for prediction aligns with global trends toward data-driven decision making in health. Shai Wininger's announcement on social media platforms indicates enthusiasm from the Lemonade side, highlighting the potential for technology companies to contribute meaningfully to global health. The NALA Foundation's long track record lends credibility to the project, likely generating positive responses from partner governments and communities. As results emerge over the 12 months, further reactions will provide insights into the practical value of the predictions.

The pilot also serves as an opportunity to build capacity among local teams in Ethiopia through training on data collection and interpretation of AI outputs. This capacity building aspect ensures that the benefits extend beyond the immediate predictions to long-term skill development in the communities and local health systems. Partnerships like this one between foundations, universities, and nonprofits demonstrate the collaborative nature required for successful AI deployments in global health. The focus on schistosomiasis specifically addresses a disease that has been neglected in terms of technological innovation compared to more high-profile conditions. By bringing modern AI tools to this area, the project could revitalize interest and funding in neglected tropical disease control efforts worldwide.

Excited to announce the launch of our disease prediction pilot in Ethiopia with @NALA_Foundation, where AI is used to help contain diseases before they spread.Shai Wininger, Co-founder & co-CEO of Lemonade

What comes next after the initial 12-month pilot period?

Following the completion of the pilot, the partners plan to evaluate the accuracy of the AI predictions and make necessary refinements to the models. The long-term vision is to develop a scalable, open-source resource that can be used by governments, NGOs, and researchers worldwide to identify disease risk earlier and target interventions more effectively. This could involve expanding the geographic scope to additional communities or countries where similar diseases are a concern. Continued collaboration between the entities will be key to sustaining momentum. The data and insights gained from the Ethiopia pilot will inform future iterations, potentially incorporating more advanced AI techniques or additional data streams as technology evolves. Success here could inspire similar initiatives for other neglected tropical diseases or even broader applications in predictive public health.

The next phase would likely include seeking additional funding and partnerships to scale the solution. The open-source release will allow the global community to contribute improvements and adaptations. Monitoring the adoption and impact in other regions will provide valuable feedback. Enterprise AI applications in this domain may grow as more organizations recognize the benefits of predictive modeling for resource optimization. The pilot serves as a foundation for building a robust system that could ultimately contribute to the reduction of schistosomiasis prevalence in Ethiopia and similar settings. By focusing on prevention, the initiative aligns with sustainable development goals related to health and well-being.

The next phase would likely include seeking additional funding and partnerships to scale the solution. The open-source release will allow the global community to contribute improvements and adaptations. Monitoring the adoption and impact in other regions will provide valuable feedback. Enterprise AI applications in this domain may grow as more organizations recognize the benefits of predictive modeling for resource optimization. The pilot serves as a foundation for building a robust system that could ultimately contribute to the reduction of schistosomiasis prevalence in Ethiopia and similar settings. By focusing on prevention, the initiative aligns with sustainable development goals related to health and well-being.

The next phase would likely include seeking additional funding and partnerships to scale the solution. The open-source release will allow the global community to contribute improvements and adaptations. Monitoring the adoption and impact in other regions will provide valuable feedback. Enterprise AI applications in this domain may grow as more organizations recognize the benefits of predictive modeling for resource optimization. The pilot serves as a foundation for building a robust system that could ultimately contribute to the reduction of schistosomiasis prevalence in Ethiopia and similar settings. By focusing on prevention, the initiative aligns with sustainable development goals related to health and well-being.