"Applying AI for Social Good": analysis of a McKinsey Global Institute's research on applications of AI for the SDGs

A simple analysis of a McKinsey Global Institute's research (2018) dating from 2020, which I had previously shared with the European AI Alliance, that is relevant today for current and future sustainable applications of AI.

---

Presentation:

Artificial Intelligence (AI) has the strong potential to be used for social good. For instance, by assisting in the achievement of the Sustainable Development Goals (SDGs). This has been one of my subjects of research during my university studies, in which I encountered this interesting research from the McKinsey Global Institue (MGI). I find it relevant enough to share it here with everyone in the Alliance.

This potential for AI regarding social good and the SDGs is illustrated by this said study by the MGI, which by November 2018 had identified as many as 135 cases of AI use worldwide that support the SDGs. Whether concrete, partial or simply potential, these cases demonstrate real opportunities for AI applications.

Attached is the related discussion paper.

The URL included in this post leads to McKinsey's web article on the topic.

Analysis:

This article shows us an inconsistent distribution of AI applications with the affected SDGs. There was at least 1 case per SDG, as in the case of Objective #14 on "Life below water", whereas some Objectives were represented by more than 20. For example, the Objective #3 for "Good health and well-being", the most represented, had encountered 28 cases.

More generally, the area most represented in these samples was "Health and hunger", which includes contexts such as "Prediction and prevention" or "Care and treatment", with 28 cases. The area least represented was "Information verification and validation", including for example uses against misinformation and false news, with only 4 cases of AI usage. 

It would be tempting to confuse such numbers with an actual assessment of the potential applicability of AI to a particular area or objective. More use cases do not necessarily means higher raw potential. Rather, it is, in my opinion, the truncated representation of the evolution and discovery of AI for applications perhaps more obvious in some cases than in others, also biased by a relatively small sample.

It is nonetheless refreshing to witness so many scenarios, not only of participation in the attainment of the SDGs, but this with the involvement of AI, less than 4 years after the definition of these objectives.