What can AI do? This is a question that many companies, organisations and even individuals are trying to answer. If you ask Google Deepmind, the company’s AI division, or NGO Armman, their answer will be: AI can save the lives of women in India. For months now, Deepmind has been working with Armman using its quirky-named AI called Restless Bandit to ensure that expectant mothers stay engaged with prenatal care, thereby reducing dropout rates and improving health outcomes.
“We want to help benefit those who have not benefited from AI,” says Dr Milind Tambe, principal scientist and director of AI for Social Good at Google Deepmind.
While AI advancements are often associated with tools like ChatGPT and Gemini, which write code and novels, Google Deepmind has been working on an AI model with an exceptionally cool name – the Restless Bandit AI – to reduce maternal mortality in India, starting with Mumbai.
According to a report by the Sample Registration System (SRS) for 2018–2020, India’s maternal mortality ratio (MMR) is 97 per 100,000 live births. Excessive bleeding during childbirth or after giving birth is a leading cause of maternal mortality in India.
In 2019, Tambe met Dr Aparna Hegde, the founder of an NGO called Armman, at a Starbucks in Mumbai. Hegde’s NGO has long been using technology to improve health outcomes for underprivileged pregnant women and children, but there was a gap they needed help filling. The organisation has a call centre with health workers calling a large number of expectant mothers.
The NGO was educating over 350,000 underprivileged mothers in Mumbai about pregnancy, nutrition, and prenatal care, sending 2-minute automated messages from the time a mother conceives until she delivers the baby. However, Armman faced a significant dropout rate of 30 to 40 percent. To help fill that gap, Google’s Restless Multi-arm Bandit was born. It was a design to “increase the efficiency of the (NGO’s) calling process”.
Fun fact: The name of the AI model is inspired by imagining a gambler at a row of slot machines, often called ‘one-armed bandits’. The gambler has to figure out which machines to play, how many times to play them, in what order, and whether to stick with the current machine or switch to a different one. Hence, its name: Restless Multi-Armed Bandit.
Tambe says that the AI model identifies who might be at risk of dropping out and how to best allocate the calling resources to get the most effective results. Throughout the pregnancy and childbirth journey, the Armman workers send around 140 automated messages to expectant mothers. The Restless Bandit’s task is to determine which mothers the health workers should call for personal follow-ups, who needs the most support, who is more vulnerable, the frequency of contact, in order to make sure most of these mothers stay engaged for as long as possible, during their pregnancy.
“The model predicts the evolving behaviour of mothers over time,” says Tambe. “It identifies mothers who will benefit from the model.”
The AI for Social Good director explains that from a large list of mothers who need to be reached, the Restless Bandit AI can predict, based on demographic information, which mother is most likely to respond to and benefit from these calls. This prediction is based on their past behaviour, previous pregnancies, etc. In fact, Tambe mentions that Armman workers noticed that the AI model could “somehow identify mothers who may be experiencing a problem or difficulty” in their pregnancy. With its prediction accuracy, the model helped reduce the dropout rate by 30 per cent. The company states that only the listenership data and demographic data of mothers were used for these predictions.
With Restless Bandit AI, Google was able to help Armman workers reach 30 per cent more mothers, get them to involve their families in their health routines, take better supplements, thereby educating them better about maternal health so that they can avoid complications at the time of birth.
With Armman, Google is also working on a national programme called Kilkari, where the company will be working with 3 million mothers across the country. The Restless Bandit AI model, says Tambe, can also be used in Tuberculosis prevention. TB patients often need to adhere to a six-month medication regimen, which makes it easy for them to drop out of the program. By using the model, health workers will be able to identify the patients who would benefit most from personal interventions or visits from health workers to remind them to take their medication.
The key to collaboration between tech companies, academic institutions, and non-profits in tackling public health issues, says Tambe, is to “start with the problem first”. He says, with its collaboration with Armman, Google Deepmind started by analysing the problem and then developing AI models and solutions to address it. The approach should be problem-focused instead of trying to force-fit models or solutions into issues, and that’s also the advice Tambe has for other organisations looking to leverage AI for social good, particularly in the health sector.
Source link
[ad_3]
