Artificial intelligence and sustainable development goals nexus via four vantage points

Abstract

Artificial Intelligence (AI) should aim at benefiting society, the economy, and the environment, i.e., AI should aim to be socially good. The UN-defined Sustainable Development Goals (SDGs) are the best depiction to measure social good. For AI to be socially good, it must support all 17 UN SDGs. Our work provides a unique insight into AI on all fronts including Curricula, Frameworks, Projects, and Research papers. We then analyze these datasets to extract meaningful information for policymakers and researchers alike - shedding light on how AI is being used and can potentially be employed in the future to achieve the SDGs. To this end, we devised a methodology using keyword-matching and keyword-similarity to compute the relevance of the SDGs for a given document. SDG metadata and AI4SDG Projects (Oxford initiative on AI4SDGs) were used to validate our methodology. We find an imbalance of coverage with SDG 9 (Industry Innovation and Infrastructure) having the highest representation (with 50.3% of our data containing references to it) compared to SDGs 5, 6, 14, and 15, which have the lowest representation (5% of observed data). Findings from this study suggest that the development of AI technology is focused on improving the current economic growth, but it might neglect important societal and environmental issues. 2022 The Authors