In 2023, Artificial Intelligence (AI) made its big splash. Its emergence has yielded systemic changes across sectors, and the international development industry is no exception. In particular, the water, sanitation, and hygiene (WASH) sector is witnessing how AI could transform traditional approaches to everything from quality testing to monitoring usage rates.
Future access to clean, reliable water sources remains a precarious and concerning prospect for myriad communities. As climate change introduces erratic weather patterns and unbalances water cycles, many peoples are struggling to predict and subsequently mitigate droughts and floods across their geographical zones.
But recent innovations in WASH can help to mitigate or solve these challenges. From portable filtration systems to water waste monitoring methods to systems designed to track the spread of diseases, WASH is expanding to meet a wide range of needs. Given the far-reaching impacts of WASH on education, health, and overall well-being, further innovation could spell significant improvements in quality of life across populations despite the challenges climate change impacts are bringing.
Enter AI. Utilizing predictive analytics and harnessing satellite imagery, artificial intelligence can monitor natural water sources to better conserve WASH resources. AI can identify leaks in municipal water infrastructure and direct teams to repair them. It can also evaluate the best times of day to move water between reservoirs and consumers to avoid evaporation or freezing in high and low temperature zones.
Moreover, AI offers meaningful ways to enhance water quality and sanitation. Operators can monitor water quality in real time and be rapidly notified of significant level changes. By feeding AI historical data, it can also predict and plan for potentially damaging contamination events. In a field that prioritizes reducing its ecological impact, such a robust tool has myriad beneficial applications for true progress.
AI-powered solutions can help identify gaps in WASH services by analyzing demographic, geographic, and socioeconomic data. This information is invaluable for designing tailored approaches to reach vulnerable populations, ultimately leading to more inclusive and equitable development outcomes. Of course, such vulnerable populations are often harder to track and collect data on. Only when such data is available can AI analysis enable evidence-based decision-making and ensure resources are utilized efficiently and targeted interventions are implemented where they are most needed.
Despite the promising potential AI poses to WASH and international development in general, there are real challenges, ethical considerations, and concerns with automating our approach to this work.
One of the primary challenges lies in technology access and affordability. Many regions where improved WASH services are most needed often lack access to advanced technologies and the necessary infrastructure. Without the resources and technology infrastructure to harness most of AI’s benefits, some of the most impoverished countries risk falling even further behind. To bridge this digital divide, the development community must continually explore opportunities to introduce existing solutions or invest in locally originated AI solutions in less technologically rich nations.
Already, such initiatives have shown promise. South African start-up Aerobotics uses AI-informed drone footage to monitor crop yields across Southeast Africa while Planet Labs uses similar satellite imagery to track reforestation efforts in Sri Lanka. Such long range-approaches can harness AI’s advantages in contexts where technological infrastructure is scarce to absent.
But even when its affordable and accessible, AI poses equally important ethical considerations when integrating within the WASH sector. Mainly, this comes down to who controls these AI systems. Predictive algorithms are only as accurate as the statistical models they’re based on. Every instance of sampling bias, unrepresentative bias, biased data, or measurement error will culminate and reflect itself through flawed AI.
While we eagerly embrace algorithmic decision-making for routine processes, such decisions become much more dire when AI is charged with choosing which community is to receive critical rainwater access, for example. As such, transparency and accountability in AI decision-making processes are vital to maintain trust and prevent potential biases or discriminatory outcomes. And moreover, AI systems should be managed and maintained by local communities rather than outside actors.
The integration of AI in the WASH sector has the potential to revolutionize international development efforts. Like with all industries in this moment, international development must remain both open to AI’s potential and sober to its shortcoming when incorporating it into overarching frameworks. Through mindful, collaborative approaches, AI and the WASH sector can flow into an enriching next chapter.
