(L-R: JSI Applied Technology Center staff Alex Tumwesigye, Leona Rosenblum, Lisa Kowalski, Joy Kamunyori, Jonathan Metzger)
I was lucky enough to attend my first Information Communications Technology for Development (ICT4D) Conference in Kampala this May as part of the JSI Applied Technology Center team and a member of the Global Digital Health Network. Throughout the week, I was involved in a variety of discussions on how to develop and use technology for good globally, both in and outside the health sphere. I also absorbed a host of thought-provoking contradictions on how the ICT4D community operates: the claim and criticism that every ICT4D conference is the same, despite touting itself as a mechanism for sharing and generating innovation, seemed to resonate online (“Tweet, Recycle, Repeat”), and led to worthwhile reflection during and between sessions (more thoughts on that here and on Twitter).
However, it’s not controversial to say that the main topic throughout the conference was data. As IBM’s Hua Ni put it in the conference’s initial supply chain discussion, “data is the new resource.” But like any valuable resource, data must be managed carefully and thoughtfully. Here is a synopsis of my data-focused learning from the conversations at ICT4D 2019.
- Data policy needs to be strengthened at all levels. Throughout the week, I noted that mid-level data policies around the world are still in progress. While at the individual level, everyone and her mother knows that you need written consent from your data source, and at the highest level the EU’s General Data Protection Regulation is dictating the way companies secure client data, data-use policies that encompass the middle level (e.g., countries, counties, or smaller regions) haven’t quite kicked in. Many individual-country data policies are still in their infancy and remain largely unenforced, and regional bodies (such as the East African Community) are still a way off from shared definitions of and requirements for secure data.
- Data need to be secure. It’s no secret that concern over data privacy has exploded in the past few years. Now that our data are stored as not only text and numeric values but also as photos and through other biometric identifiers (also a popular topic at this year’s ICT4D), we have to find new ways to store them securely in our systems. At the event, there was also widespread recognition that the push from funders for open data adds an additional level of complexity to maintaining data privacy.
- Data need to be usable. If a tree falls in the woods and no one is there to hear it, does it make a sound? This statement applies to data accessibility. Thinking about this in terms of JSI’s work: it’s great that frontline health workers and service providers across many levels are able to collect and store their data at the regional or national level, but if they can’t see or access them, or use the data to see how their patients fit in with the larger whole, does it even matter that they have them? Nor can we forget that the individuals providing data should have the right to access them. We must ensure that our digital systems make data available to clients, providers, and implementers, and are, of course, secure. I look forward to seeing more case studies on how organizations have used advanced analytics and other new processes to keep data-collection systems light, effective, and non-duplicative to make data as accessible as possible to decisionmaker.1
- Data still need human involvement. Panel presenter Michael Moszczynski of ImmerLearn phrased this concern succinctly: “There is a risk of automating everything and removing the human element.” With a major spotlight on machine learning at this year’s ICT4D, I particularly enjoyed the “AI Bias in Machine Learning for Good” panel. While machine learning can make data processing easier and faster, you still need to start with data you can trust (aka training data), and that is determined by humans. As it was so aptly put by presenter Amit Gandhi of MIT, “there is a distinction between fairness and bias,” and off-the-shelf tech solutions cannot account for myriad socioeconomic factors and nuances—you still need a data specialist! (This panel really spoke to my inner anthropologist as well as my burgeoning techie).
- Data need to be championed. No one promoted this ideal more throughout the week than the stewards of the Principles for Digital Development (PDD). As an add-on to the main conference, I attended the PDD team’s all-day workshop, and their best practices for digital activities encourage fostering a data-driven environment across all points! However, the “Be Data-Driven” Principle asks designers to “create a data-use culture by prioritizing capacity-building and data-use efforts across ALL stakeholder groups.” As designers of digital systems, it’s our responsibility to ensure that we’re not just creating tech solutions; we’re also spreading the message that data is essential, and fostering its use at all levels to lead to the best possible decisions.
That’s all for now. I’m already looking forward to ICT4D 2020!
P.S. While not directly related to data, these panels also resonated with me:
- Engaging effectively with telecom organizations (Digital Impact Alliance (DIAL), NetHope and Financial Sector Deepening Uganda) provided a thoughtful and refreshingly frank overview of the dissonance between telecom organizations and implementers, framed by the challenges of supply and demand they need to work through together to scale up SMS, IVR, voice, or mobile money programs to national levels. A summary report will be available from DIAL in June 2019.
- Putting the user at the heart of the digital design process: lessons from the SPRING Accelerator. This panel seemed especially relevant, since the SPRING Accelerator product SafeBoda (‘Uber for boda’, which has a specific mission to provide safe access to mobility) was highly visible every day in Uganda—I witnessed dozens of individuals cruising around Kampala with highlighter-orange clad SafeBoda drivers (as in the photo below). Very cool to see the product of a successful human-centered design process, which now operates across Uganda and Kenya.
A SafeBoda driver and passenger in Kampala. Photo: Lisa Kowalski
1.It’s tempting to add another interpretation of “data access” here: the literal access to mobile data. One of the most interesting panels I attended focused on the effects of internet shutdowns on women in conflict areas of India (report here), and I’ve noticed several other recent articles on the shocking number of internet and data cutoffs in multiple countries. A number of humanitarian programs are tackling this, but I’m curious to see if/how it will be addressed in donor-funded international development programs, and to see what the role of implementers like us will be, if any.↩