“Tech for good” is a term that has captured the attention of the innovation community around the world. For the UNICEF Innovation Fund, “tech for good” means to provide early-stage funding and support to frontier technology solutions that benefit children.
In a recent interview, Oasis sat down with Benjamin Grubb, manager of Technology for Development at UNICEF East Asia and Pacific, and Stephanie Sy, founder and CEO at Thinking Machines, a tech consultancy that provides AI and big data solutions. The discussion covered the journey of their partnership as well as the ways that technology has produced a new understanding of poverty in the Philippines. Thinking Machines is also a portfolio company of the UNICEF Innovation Fund.
The following interview has been edited and consolidated for brevity and clarity.
Oasis (OS): How did the collaboration between the UNICEF Innovation Fund and Thinking Machines come about?
Stephanie Sy (SS): The Fund was recommended to us by eight friends. At the time, we were only two or three years old, just getting our feet under us. When we started talking to UNICEF, we found kindred spirits, people who care about how data and technology are transforming. UNICEF shares our core values, such as open source and open data.
OS: Ben, how did you start to work with Stephanie?
Benjamin Grubb (BG): We got to know Thinking Machines near the end of the initial investment phase, when we were looking for more concrete applications in big data. We worked with the Philippine government and Thinking Machines to identify where and how such applications might be possible.
Two years later, we’ve come to the second phase of that partnership. UNICEF has been working closely with Steph to define the steps in this next tranche of work. For me, that means working with the Philippine government to identify what they need, and then sitting down with Steph and her team to work out what’s possible.
OS: What’s the second phase of work?
SS: In places like the Philippines, development happens at a quick pace. However, the national census is always a couple of years late. It is prohibitively expensive to gather the data, make calculations, and use the results to determine the stage of development of different regions. We worked with a team at Stanford to apply artificial intelligence on satellite imagery, so it became possible to not just capture the geographical features, but infer wealth or poverty from the images. That was the first tranche of original work that UNICEF funded.
The second phase of our partnership focuses on the post-processing: building the UI and analytics engines so different stakeholders can come up with actionable conclusions. Next, we’re looking to move beyond the Philippines, so this work can be applied and scaled across different countries in the region.
OS: Why is it necessary for UNICEF to collaborate with companies like Thinking Machines?
BG: UNICEF has a presence in almost every country. But what we don’t have is the technical expertise that groups like Thinking Machines can provide. The vast majority of the UNICEF workforce aren’t tech specialists, they’re epidemiologists or educators. Their expertise is in nutrition and children’s well-being. When we work with governments, it’s about finding organizations like Thinking Machines that can help us address core challenges and problems that are raised from the UNICEF teams in the field. Their deep expertise helps to bridge gaps that large organizations like ours aren’t as nimble with, or in situations where we can’t necessarily develop solutions based on our in-house capabilities.
OS: A lot of partnerships between investors and startups have tension in the beginning. What was it like for you two to work together?
BG: We try to avoid any kind of tension. At the same time, we always welcome disagreements and different perspectives. This is why we value working with companies like Thinking Machines. Reflecting on the collaboration process, there were definitely times when we had to compromise, which is usually a good thing. At the same time, we have these priorities coming from the government and partners, and we have to be cognizant to respond to those needs. Of course, big data is tricky, we do get some impossible requests sometimes, so knowing the boundaries is critical. That’s where Thinking Machines has lots of insights.
SS: Some disagreements arose because of the different natures of our organizations. I know that we’ve given them a few gray hairs. In this process, we provide deep technical expertise. But in other areas, it’s more a matter of policy. It’s about what kinds of hoops we have to jump through in order to make a piece of technology usable from a policy perspective, which is where the UNICEF team truly shines. What’s the point of a great machine learning model if it won’t be adopted because of policy factors?
OS: What’s the biggest developmental problem that you feel tech or data has to address in the coming years?
BG: A lot of misinformation is emerging in the online space. I like to give the example of the COVID-19 pandemic. We’ve also seen online environments propagating lots of hateful messages. From the child care perspective, we’re really cognizant about bullying.
SS: Over the last couple years, our thinking as a group has shifted to embrace issues of sustainability and climate. There is a tension in the Philippines—we want to become more environmentally friendly. At the same time, the country is in need of huge energy, agricultural, and food supply for economic development. Plus, a nation of 7,000 islands can be sent into chaos when typhoons hit.
I want to see if there are better uses of technology to drive decision-making. If you drive purely towards economic development, you’re going to end up with a lot of unintended consequences on the climate side. If you drive purely towards climate outcomes, then you will not see sustainable changes.
BG: I love that answer from Steph. I think it reflects our current work. Thinking Machines and UNICEF are working to understand poverty from both granular and communal levels. Our next mission is to bring new layers to the data. For example, combining environmental data with poverty to see how the environmental factors contribute to poverty. That means we can interpret the data in new ways, and inform policymakers on budgeting, so they can plan more equitably to address emerging climate challenges.
OS: What changes were brought about by the COVID-19 pandemic?
BG: It’s been a huge fast forward. Dependencies on technology have grown in the last 18 months. They were probably fast-tracked by around ten years in some places in the Southeast Asia region. The use of data also gets me super excited. One of the more concrete examples is through data collaborations. We’ve been able to track mobility at a regional level, and we can see movements of populations in real time.
This article was produced in collaboration with the UNICEF Innovation Fund, which provides early-stage funding and support to frontier technology solutions that benefit children and everyone else in the world.