This article explains the essentials of AI and data-driven analytics, analyzes the solution space of a number of human development problems, and identifies the most valuable contributions that AI and data analytics can make.
According to the project website, "This article identifies the most important interventions required to achieve the SDGs, and assesses their dependence on data analytics and AI. Our analysis finds that robust data analytics can play a meaningful role in achieving the SDGs. However, About half of the major interventions required to achieve the SDGs do not need sophisticated data analytics; rather they depend on infrastructure, physical tools, innovative business models, and other on-the-ground programs. To fully implement the interventions that do rely on data, as well as for long-term development, transparency and accountability, countries need to invest in a robust data infrastructure. To measure the breadth and depth of a country’s digital environment, we introduce the Data Density Index (DDI), a composite metric that assesses the volume and variety of data generated across platforms built by both governments (for key public services) and private companies (e.g., smartphones and apps). Based on the available evidence, the DDI for most developing countries ranks them as “data deficient”. This suggests that despite the telecom revolution, much needs to be done before decisions and interventions are informed by robust, granular data. In that context, India’s Aadhar and India Stack initiatives represent a powerful model of digital and data inclusion. If such systems are implemented in other parts of the developing world, they can make the hype of data analytics—and perhaps even AI—become reality, over time. In many ways, the timing may be ideal for emerging economies to make robust investments in their data infrastructure. As such, institutions aiming to impact the SDGs have four options: (1) focusing on more direct interventions (e.g., irrigation and seed hybridization for agricultural development); (2) investing in conventional analytics solutions to improve decision-making on direct interventions; (3) building data infrastructures for the long haul; and (4) betting on new-generation AI solutions. Depending on the context, each of the first three options can be valuable and can be implemented in tandem; however, we believe the AI option is the least likely to lead to impact in the timeline for the SDGs." https://d386wwnkwgr87h.cloudfront.net/wp-content/uploads/2019/03/AI-Report-Mar-2019.pdf
The report reminds us of a possible "knowledge gap between the AI community and the development community" and provides us with a different perspective on AI for SDG.