Digital Economy

Can artificial intelligence help us achieve all 17 Sustainable Development Goals (SDGs)?

Not just doomsday scenarios and labour crises, the new McKinsey report is a breath of optimism.

18 August 2024

4' min read

Not only doomsday scenarios, labour crises and laboratory-generated viruses. Artificial intelligence can help mankind achieve all 17 of the UN Sustainable Development Goals (SDGs) and in fact there are already concrete examples, use cases in all these areas.

The AI and the 17 objectives

In the days when AI is being used - among other things - to generate electoral misinformation and alarms are growing over the booming energy consumption in data centres due to this technology, the report just published by McKinsey is a breath of optimism. In 29 pages it delves, with many examples, into the ways in which AI is already contributing to goals of utmost importance to the future of humanity, which the United Nations would like to achieve by 2030: eradicating poverty, hunger, ensuring health and wellbeing, providing quality education, achieving gender equality, ensuring clean water and sanitation; ensuring clean and affordable energy, promoting sustained and sustainable economic growth, building resilient infrastructure and promoting innovation; reduce inequality; make cities and communities sustainable; ensure sustainable patterns of consumption and production; combat climate change; conserve and sustainably use oceans and marine resources; protect terrestrial ecosystems; promote peace, justice and strong institutions; and strengthen partnerships for sustainable development. The United Nations in 2023 recognised that in seven years since the 17 goals were initiated, humanity is still 15 per cent of the way there.

The point is that AI could help us speed up, then, 'but it is not a magic bullet' acknowledges McKinsey, and there are several major challenges to overcome in order to exploit this opportunity.

Positive cases

Experts interviewed by McKinsey agree that AI has a particularly high potential to make a difference for five SDG goals: good health and well-being (SDG 3), quality education (SDG 4), affordable and clean energy (SDG 7), sustainable cities and communities (SDG 11) and climate action (SDG 13). In fact, 60 per cent of AI uses for social good for non-profit purposes were in these areas.

For example, in Kenya, Jacaranda Health uses AI-based solutions to improve the quality of care for women and reduce maternal deaths. Their Prompt service, an exchange of personalised text messages to women, increased the likelihood of mothers attending more than four antenatal visits by 20 per cent. In India, Armman developed mMitra, an automated voice messaging system that provides key information on preventive care. Some 3.6 million women in nine states were reached in this way.

Another notable example is DeepMind's work with AlphaFold to predict the structure of proteins. This tool has been instrumental in medical research to develop treatments for neglected diseases and combat antibiotic resistance.

In the field of education, Livox uses intelligent algorithms to adapt content to students with different disabilities. More than 25,000 people with disabilities have used this service. In India, Educate Girls uses a machine learning model to reduce the operational costs of identifying girls who do not attend school. This enables staff to reach more potential students more quickly.

In terms of climate action, Global Forest Watch (Gfw) uses satellites, computer vision and deep learning to monitor illegal deforestation. Friends of the Earth Nigeria used Gfw data to monitor deforestation and return almost 14,000 hectares of land to communities. Google Flood Hub, on the other hand, provides flood forecasts up to seven days before a hazard, protecting livelihoods in more than 80 countries.

There are cases, also less important, for Sdg 1 (No poverty). GiveDirectly in Togo, with UC Berkeley and telecom operators, exploited AI during covid to accurately estimate poverty and improve the targeting of aid to the population.

AI is also improving processes and productivity in agriculture. Outgrow in India connects farmers with high-value products and services using an AI-based platform. It supports over 200,000 farmers.

In Colombia, Future Seeds uses predictive models to identify habitats that host important plant biodiversity. This preserves key specimens and supports food security.

How funding for AI initiatives supports SDGs

The report also analyses AI funding for SDGs and finds that it is concentrated with the highest potential areas, such as health, education, clean energy, sustainable cities and climate action. Most of the funding comes from private capital, but there is a geographical disparity: most funds go to organisations in high-income countries.

The challenges

This alone is a sign that it is a perfectible system. The challenges are indeed numerous. Starting with the availability, accessibility and quality of data to be used for AI, especially in places that would need this help the most. Ditto for the necessary computational resources and other issues. The distribution of AI talent is uneven, with a high concentration in high-income countries. Another point, adopting AI requires organisations to change, but this is difficult especially for non-profit organisations.

The advice

McKinsey also makes some suggestions. Creating partnerships to accelerate impact is crucial. Collaborating with other organisations to share resources, data and talent can make a difference. So can supporting the development of digital public goods, making models, software and data accessible for sustainable development.

In this regard, it is important to strengthen the quality and usability of data by improving data collection and management, especially in resource-poor environments. Expanding the AI talent pool for SDGs requires investment in training new talent and promoting the borrowing of talent from large companies. Adopting an inclusive and user-centric approach can ensure that AI solutions are developed in collaboration with end-users. This will increase their adoption. Finally, an ad hoc 'business model' is needed to financially support AI initiatives on an ongoing basis.

Brand connect

I prossimi eventi

Tutti gli eventi

Notizie e approfondimenti sugli avvenimenti politici, economici e finanziari.

Technology