A new artificial intelligence application launched this week aims to help millions of Bangladeshis identify arsenic-contaminated drinking water without requiring laboratory testing, using local knowledge combined with scientific data.
The digital platform, named iArsenic, utilises a decade of groundwater research alongside artificial intelligence to evaluate whether water from hand-pump tubewells poses contamination risks. The system provides immediate, location-based risk evaluations by analysing user-submitted data.
Dr Mo Hoque from the University of Portsmouth’s School of the Environment and Life Sciences emphasised that “iArsenic serves as an initial screening mechanism rather than replacing chemical analysis, offering a defence tool accessible via mobile phones or computers.”
The initiative represents collaboration between the University of Portsmouth, Curtin University, University of Dhaka, and Imperial College London, with backing from Bangladesh’s Department of Public Health Engineering (DPHE). It targets one of the nation’s most enduring public health challenges.
Addressing decades-old contamination crisis
Groundwater arsenic contamination gained widespread recognition during the 1990s. The issue developed following government and NGO initiatives promoting tubewell installation to combat cholera and diarrhoeal diseases from surface water sources. Though microbial threats decreased, numerous wells accessed arsenic-laden aquifers, creating chronic exposure risks for rural communities.
Extended exposure consequences prove devastating, with links to cancers, heart disease, skin conditions, and childhood developmental issues, even below Bangladesh’s 50 µg/L national threshold. The World Health Organization recommends 10 µg/L maximum levels, yet approximately 20 million Bangladeshis continue consuming water from contaminated sources.
More than 10 million tubewells underwent chemical testing through the BAMWSP survey (2000–2005) and the Arsenic Risk Reduction Project (2021–2023). However, tubewells typically last only 10 years, with constant new installations often lacking subsequent testing, while households frequently lose previous results.
Simple technology, powerful results
The application requires users to input three observable elements: well depth, geographical position, and concrete platform staining patterns. Machine learning algorithms trained on millions of historical arsenic measurements predict contamination likelihood for specific wells. Colour-coded outcomes—green indicating probable safety, red suggesting potential contamination—appear immediately.
Kane Swartz, Technical Lead from the University of Portsmouth, explained: “Local residents confidently identify these colours and typically know their well depths. We discovered this basic information, combined with location data, provided remarkable predictive capabilities.”
The technology relies on visual indicators around tubewells. Wells displaying black manganese stains generally proved safe, whilst red or yellow iron staining frequently indicated arsenic contamination, particularly in shallow installations.
Researchers report 84% accuracy in assessing tubewell arsenic risks using the three-input system.
The platform processes user data through machine learning models built on millions of test results and geochemical information nationwide. Users receive colour-coded risk assessments whilst contributing to a real-time interactive map tracking application usage and regional risk patterns.
Professor Adrian Butler from Imperial College London noted: “This grassroots-level risk mapping approach is unprecedented, transforming collective scientific knowledge into public infrastructure.”
Beyond individual household benefits, the application supports national planning initiatives. Research teams developed three-tier classifications identifying thousands of “Tier 1” villages and urban areas where arsenic risks remain high whilst safe water alternatives stay limited.
- Portsmouth students launch 900km space mission
- Hampshire urges residents to ditch single-use plastics
Dr Ashraf Dewan from Curtin University explained: “These represent priority locations requiring urgent coordinated investment. Digital tools like iArsenic cannot independently resolve the problem but can focus efforts, support local initiatives, and extend government screening programme reach.”
Projections suggest widespread adoption could assist millions of Bangladeshis in switching to safer water sources over coming years, protecting health, supporting livelihoods, and reducing long-term healthcare system burdens.
Dr Hoque concluded: “Clean water represents a fundamental right, not a luxury. This constitutes progress toward making that right universal.”
The application currently operates as a web-based platform accessible through browsers including Chrome, Firefox, and Safari, with Android and iOS mobile applications planned for later this year.
