AI advances Over the past two years, advances in AI have accelerated dramatically, giving us the ability to design, train, and most importantly, run AI at scale for millions of users. The potential impact of AI spans many sectors and has the potential to bring about positive change in society.
In fact, AI is already being used to advance all 17 of the United Nations' Sustainable Development Goals (SDGs), from eradicating poverty to building sustainable cities and communities to providing quality education for all. Generative AI has opened up even more possibilities. In 2018, it was becoming clear that AI could play a major role globally in driving productivity and economic growth as well as societal benefits. A report we published at the time outlined how AI capabilities, from natural language processing to speech recognition and tracking, could be used in 170 use cases to benefit society, including promoting equality and inclusion, improving crisis response, and protecting the environment. Today, we have found a total of nearly 600 use cases, a more than threefold increase.
By mapping innovation and impact to the SDGs, the 2024 report will once again examine how AI can and already is a key part of solutions that benefit people and the planet (see sidebar, “Methodology”). The SDGs consist of 17 goals and 169 targets aimed at improving lives around the world and protecting the planet. Yet, according to the UN’s 2023 SDG progress update, the world has achieved only 15% of the SDG targets. In practice, that means that today, 2.2 billion people lack access to safe water and sanitation, 3.5 billion lack access to safely managed sanitation, about 3.3 billion people live in environments highly vulnerable to climate change, and 750 million people face hunger.
The potential of AI for the SDGs and how funding it can support progress
While AI impacts all SDGs, experts we surveyed believe that AI has the potential to make a big difference in five goals in particular: Good Health and Well-being (SDG 3), Quality Education (SDG 4), Climate Action (SDG 13), Affordable and Clean Energy (SDG 7), and Sustainable Cities and Communities (SDG 11). In fact, 60% of non-profit AI deployments for social impact were in these areas. Zero Hunger (SDG 2), Life on Land (SDG 15), and Peace, Justice and Institutions (SDG 16) have more use case deployments relative to the perceived potential of AI, whereas Quality Education (SDG 4), Affordable and Clean Energy (SDG 7), and Climate Action (SDG 13) have fewer (Figure 1). We identified partnerships that address the goals of Decent Work and Economic Growth (SDG 8), Building Infrastructure for Industry and Innovation, and SDG 17 from our nonprofit deployment, foundation grants, and private capital analysis, because the scope of AI is broad enough that we can tag most of the projects into these areas.
When we analyze where funding is being directed to leverage AI to achieve the SDGs, we find a clear alignment with expert opinion on the five most promising areas: the combined view of grant and private capital funding focuses on good health and well-being (SDG 3), quality education (SDG 4), affordable and clean energy (SDG 7), sustainable cities and towns (SDG 11), and climate action (SDG 13). Additionally, about 40% of private capital investments in the 20,000 AI companies analyzed directly or indirectly contribute to at least one of the 17 SDG thematic areas.
But exciting opportunities exist: consider the relatively low private capital funding for quality education (SDG 4). Even when it comes to affordable clean energy (SDG 7) and climate action (SDG 13), over 50% of private capital investment is going towards autonomous vehicles to improve energy efficiency and reduce emissions. This suggests further potential for private companies to deploy capital in SDG themes where AI holds great promise.
Geographic disparities in grant allocation remain significant: An analysis of grantee headquarters locations from a database of U.S.-majority foundations revealed that only 10% of grants allocated to AI initiatives addressing one or more SDGs from 2018 to 2023 went to organizations based in low- or middle-income countries. While organizations can have impact outside of their home countries, 60% of experts responding to our survey agreed that current AI efforts are not sufficiently focused on benefits to low-income countries (as opposed to high-income or developed countries), where the needs and SDG impacts are likely to be highest.
Challenges and risks of scaling AI for social good
The challenges of scaling AI for philanthropy are persistent and daunting. 72% of respondents to our expert survey noted that the majority of efforts to deploy AI for philanthropy to date have focused on research and innovation rather than adoption and scale. 55% of grants for AI research and deployment across the SDGs are $250,000 or less, consistent with a focus on targeted research and small-scale deployment rather than large-scale scale. Outside of funding, the biggest barriers to scaling AI continue to be data availability, accessibility, and quality; availability and accessibility of AI talent; organizational acceptance; and change management. For more information on these topics, see the full report.
While overcoming these challenges, organizations should also be mindful of strategies to address a range of risks, including inaccurate outputs, biases embedded in the underlying training data, the potential for misinformation at scale, and malicious impacts on politics and personal well-being. As several recent articles have noted, AI tools and technologies can be misused, even if they were originally designed for social good. Experts cited reduced fairness, malicious use, and privacy and security concerns as the top risks, followed by explainability (Figure 2). Nonprofit respondents expressed relatively strong concerns about misinformation, workforce issues such as job losses, and the impact of AI on economic stability, while for-profit respondents were more concerned about intellectual property infringement.
Accelerating the adoption of AI for social good
Scientific advances have improved AI's effectiveness in pattern recognition, prediction, and creation. This progress has coincided with a surge in successful AI deployments, but as noted above, challenges remain in the expanded use of AI to solve the SDGs. To realize this potential, stakeholders must work more closely together to ensure access to sufficient talent and robust data solutions, as well as access to more open-sourced or scalable AI applications and models across user geographies around the world, to respond to people in their time of need.
By coming together to find ways to use AI at scale for good, mission-driven organizations, governments, foundations, universities, developer ecosystems, and businesses can help solve some of the world's toughest, most intractable problems: stopping human trafficking, ensuring girls and children around the world get the education they deserve, protecting forests from illegal deforestation, supporting the health and safety of pregnant women and newborns, and so much more. If these things aren't worth fighting for, then what are we fighting for?