The Role of AI in Achieving the SDGs

The Role of AI in Achieving the SDGs Francesco Fuso Nerini, As. Professor, ffn@kth. se Ricardo Vinuesa, As. Professor, rvinuesa@mech. kth. se 2021 -02 -27 Footer 1

Content 1. The Interconnectedness of Global Goals 2. Artificial Intelligence and the SDGs 3. Conclusions 2021 -02 -27 Footer 2

The Interconnectedness of Global Goals 2021 -02 -27 Footer 3

Global Goals in practice SDGs as a Powerful tool to: • Collect and categorize research • Understand policy needs • Cross-sectorial coordination 2021 -02 -27 Footer 4

An exciting field of research 2021 -02 -27 Footer 5

2021 -02 -27 Footer 6

Motivation We want to answer the question: “Is there published evidence of AI acting as an enabler or an inhibitor for each of the SDG targets? ” We needed to assemble a multi-disciplinary team spanning the wide range of required areas of knowledge. 2021 -02 -27 Footer 7

The Team H. Azizpour R. Vinuesa Fluid mechanics, AI fundamentals Applied AI S. Domisch Biodiversity 2021 -02 -27 A. Felländer AI ethics I. Leite AI and social interaction S. D. Langhans Freshwater ecology Footer M. Balaam Interaction design M. Tegmark Cosmology, Applied AI V. Dignnum AI ethics F. F. Nerini Energy systems, sustainability 8

AI means BUT: NOT (YET) SOFTWARE TECHNOLOGY WITH EITHER: • Perception (e. g. face recognition) • Decision-making (e. g. medical diagnosis systems) • Prediction (e. g. weather forecast) • Automatic knowledge extraction and pattern recognition from data (e. g. discovery of fake news circles in social media) • Interactive communication (e. g. social robots or chat bots) • logical reasoning (e. g. theory development from premises). 2021 -02 -27 Footer 9

Methods – 169 evaluations 2021 -02 -27 Footer 10

Results 2021 -02 -27 Footer 11

Impact of AI on each of 169 targets - We divided the 17 SDGs into 3 main categories (Stockholm Resilience Center, 2017; United Nations (2019): Society, Economy, and Environment. - Percentage of targets where positive (79%) or negative (35%) impact of AI is documented: Vinuesa et al. , Nature Communications 11, 233 (2020) 2021 -02 -27 Footer 12

Dividing the 17 SDGs in 3 main pillars - We divided the 17 SDGs into 3 main categories (Stockholm Resilience Center, 2017; United Nations, 2019): Society, Economy, and Environment. Vinuesa et al. , Nature Communications 11, 233 (2020) 2021 -02 -27 Footer 13

Impact of AI on each of 169 targets Detailed analysis considering the targets from each of the 17 SDGs: Vinuesa et al. , Nature Communications 11, 233 (2020) 2021 -02 -27 Footer 14

Impact of AI on each of 169 targets Vinuesa et al. , Nature Communications 11, 233 (2020) 2021 -02 -27 Footer 15

Impact of AI on each of the 169 targets Poverty, economy and Industry: AI-enabled technology may help new jobs and productivity. Satellite data to track poverty. Vinuesa et al. , Nature Communications 11, 233 (2020) 2021 -02 -27 Footer 16

Impact of AI on each of 169 targets BUT with inequality: Certain jobs will be lost Uneven opportunities to access AI resources may end up increasing inequalities Vinuesa et al. , Nature Communications 11, 233 (2020) 2021 -02 -27 Footer And algorithms may polarize societies and increase discrimination 17

Impact of AI on each of 169 targets Polarized and unequal societies can in turn have impacts on peace and stability Vinuesa et al. , Nature Communications 11, 233 (2020) 2021 -02 -27 Footer 18

Impact of AI on each of 169 targets AI can potentially enable smart and low-carbon societies. e. g. • it can help to integrate variable renewables by enabling smart grids applications. • It can support the creation of smart and low pollution cities (800, 000 deaths per year due to pollution in EU. Vinuesa et al. , Nature Communications 11, 233 (2020) 2021 -02 -27 Footer 19

Impact of AI on each of 169 targets And some AI applications can support the management of natural resources and resource efficiency Vinuesa et al. , Nature Communications 11, 233 (2020) 2021 -02 -27 Footer 20

Impact of AI on each of 169 targets However, at the same time, AI can have large effects on the global energy demand. the total electricity demand of information and communications technologies (ICT) could require up to 20% of the global electricity demand by 2030, from around 1% today. Vinuesa et al. , Nature Communications 11, 233 (2020) 2021 -02 -27 Footer 21

Interaction of AI with society: Very fast development of technology Vinuesa et al. , Nature Communications 11, 233 (2020) 2021 -02 -27 Footer 22

Conclusions 1. AI is already affecting everyone´s lives • Positively: e. g. smart grids and cities, e-health • Negatively: ´big nudging´, citizens scores 2. And that will only increase, potentially enabling an AI-fueled Sustainable Development but • Technology evolves faster than citizens and governments • Large research gaps to manage the transition • Vulnerability of infrastructures 3. Substantial work is needed for overcoming AI gaps in transparency, safety and ethical standards 2021 -02 -27 Footer 23