Artificial Intelligence (AI) is a transformational technology with the potential to arrive at solutions to the most complex problems. The use of AI in sectors such as genetic modelling and drug discovery have been reported to be successful. Not surprisingly, AI for Good or AI for Development are currently trending buzzwords.
At the same time, there is an aspect about technology that is worrying. Many significant technologies such as Big Data, Internet of Things (IoT), and Blockchain have failed to create the desired or expected impact due to their inability to provide a suitable use-case to address problems of the needy in society. Any effective solution to deep-rooted problems in sectors like healthcare, agriculture, and education calls for the involvement of all stakeholders, especially the end users.
AI too is facing a similar challenge. What can potentially go wrong in the magnificent world of algorithms, sensors, and data? The answer is right in front of our eyes, like the proverbial elephant in the room. Yet, we fail to see it. We get so enamoured with the technology, that the real users at the grass-roots level are all but forgotten. Little wonder that the world is yet to see an AI-led transformative social solution that creates an impact beyond the white papers, labs, and narrow consumer products.
At the Center for Study of Science, Technology and Policy (CSTEP), we took the challenge head-on. We worked with the Government of Karnataka on a pilot programme for designing and deploying a malnutrition management solution in the state. Called SNEHA, our concept solution explores a transformative approach towards malnutrition management.
Child and woman malnutrition is a complex problem in developing countries, including India. Child malnutrition is a predominant cause for early mortality and lifelong disabilities. The stunting of children under the age of 5 stands at a staggering 39 percent in India. A healthy child requires a healthy mother during pregnancy and after delivery.
Maternal health is a cause for concern too, with the Maternal Mortality Rate (MMR) in India currently being 130 per 100,000 births. Health services to women and children are administered primarily by the Ministry of Women and Child Development through its network of anganwadi centres and the Ministry of Health and Family Welfare through the network of ASHA workers and ANM (auxiliary nurse and midwife).
On the face of it, finding a solution is a technologist’s delight. One can easily visualise the ready utility of mobile-based apps in the hands of field workers—AI-enabled tools for detection of health problems and the identification of beneficiaries. Contrary to popular belief, the designing of solutions should start from the grassroots, with the workers on the ground, and not with technocrats in IT labs. A common approach in such complex solutions is Systems Thinking, which focuses on the big picture.
With SNEHA, our researchers took a holistic approach, spending time in the field, observing, understanding and discussing the issues faced by health workers, field functionaries and the beneficiaries—in this case, the anganwadi workers and new mothers. After months of research, we identified four issues that needed intervention.
We realized that the mother’s health management had to take centre stage, with easy access to health facilities. Second, we discovered that the data across multiple departments were inconsistent. Third, the processes of health management across multiple departments, in multiple ministries, were found wanting. Lastly, we realized the need of field workers to go beyond daily record keeping at anganwadis. For instance, they needed a facility to escalate issues and raise complaints for common problems such as electricity, water, and non-functional toilets.
Most problem-solving methods focus on the problem, rather than the entire ecosystem where the difficulty exists. Systems Thinking introduces a wide perspective, and encourages looking at the whole picture—seeing interrelationships between elements— than just a static snapshot.
Given that malnutrition is a grave socio-economic issue, we spent months working alongside government officials to understand the problems on the ground and the likely impacts of potential solutions. As a result, data was cleaned up; integrated processes were established, and the process to provide health access to women was improved. We conceptualized a technology solution only after laying a solid foundation of processes and data, which took into account the ground reality.
Our solution has three distinct components. Firstly, a mobile app in the hands of anganwadi workers based on the principles of inclusive design. Secondly, multi-sectoral process and data integration cutting across many departments. Lastly, the use of advanced technology like image recognition for accurate and evidence-based data entry.
The ‘inclusive design’ ensures that the solution meets the needs of the grassroots functionaries such as anganwadi workers, ASHA workers and the ANMs. For effective adoption of any solution, the benefit has to be felt by those who provide the services—in this case, the field workers.
The ‘multi-sectoral’ solution cuts across the Department of Women and Child Development and the Department of Health and Family Welfare. The solution focusses on end-to-end malnutrition management and not just on routine automation or rudimentary data collection. ‘Advanced technology’ ensures that important data points, such as a child’s height, are captured accurately through the click of a camera, while authentication of critical services like immunisation are done using biometric evidence.
Going ahead, we plan to incorporate advanced analytics for actionable insights, for officials and policymakers. Research is currently underway for using computer-vision techniques for detecting low birth-weight and anaemia.
The experience gained while working on this project highlighted the need to adopt design principles in a comprehensive technology-driven social solution. The more complex a social issue, the bigger the need for a comprehensive approach, extending from design to deployment.
Though a relevant component, technology is just a part of the overall solution, which needs to factor in several non-technology scenarios, including the human element, the processes, and integration with other systems.
Any AI solution that fails to integrate the human element, processes and data, will fail to reach out extensively and achieve the desired outcome.
Views expressed above belong to the author(s).