Artificial Intelligence in Sustainable Development Research
Abstract
Artificial intelligence (AI) offers strong potential to support Sustainable Development Goals (SDGs) through data-driven insights and system optimizations. This article reviews 792 studies exploring AI applications related to SDGs, categorizing them by disciplinary scope—from natural sciences to humanities—and by focus, distinguishing economic from socioecological content. Deep learning and supervised machine learning dominate for forecasting and optimization tasks. Yet, a major gap exists: few studies effectively combine advanced AI techniques with in-depth sustainability expertise. To create meaningful, responsible change, sustainability research must balance context-specific detail with broader applicability, with AI playing a central role. Despite high expectations, AI’s full impact on sustainable development remains unrealized.
The 2030 Agenda and AI’s Potential
The United Nations’ 2030 Agenda outlines 17 SDGs aiming for peace, prosperity, and environmental sustainability. Achieving these goals requires integrated approaches addressing interlinked challenges such as poverty, ecosystem protection, peace, and inclusive growth. Sustainability science focuses on systemic solutions rather than fragmented fixes.
Advances in AI provide new tools to speed progress by enabling systems thinking and data-driven decision-making. AI can analyze complex datasets, predict climate patterns, and tackle critical issues. However, AI development often reflects the priorities and values of the countries developing it, with uneven global access linked to economic capacity. Ethical and regulatory frameworks lag behind technology advances, raising questions about how transformative AI truly is for sustainable development.
Reviewing AI Research in SDG Contexts
This review examines AI’s integration with SDG-related research by analyzing 792 peer-reviewed articles sourced from Scopus. Articles were selected based on their engagement with both AI and SDGs, focusing on the 100 most cited studies per goal where possible. Two main dimensions emerged: a disciplinary axis from natural sciences to humanities, and a focus axis separating economic-oriented from socioecological studies.
The analysis reveals that while AI is widely applied, few studies deeply integrate AI methods with sustainability expertise. Bridging this gap is key to realizing AI’s potential for sustainable development.
Research Trends and Geographic Focus
Publication volume has increased sharply since 2019, exceeding 200 articles annually by 2022 and 2023. Most research originates from Europe and Asia, with China, India, the United States, and Spain contributing 38% of the analyzed work.
- Iran, India, and Spain focus heavily on SDG 6 (clean water and sanitation).
- Italy and the UK mainly address SDG 3 (health and well-being).
- SDG 4 (quality education) studies are most common in the US, Spain, and China.
Empirical articles dominate research output, especially from China, the US, and India. Conceptual and review articles provide additional insights but differ in approach and structure from empirical work.
Patterns in Empirical AI-SDG Studies
Using hierarchical cluster analysis on conceptual vocabulary, eight groups of empirical articles emerged, organized along two axes:
- Disciplinary Axis: Ranges from natural sciences to humanities. For instance, health and education studies cluster together, while research on hydrological systems and vegetation forms a separate group.
- Focus Axis: Differentiates economic-oriented studies (e.g., clean energy, industry) from socioecological ones (e.g., hydrology, healthcare).
Each cluster is characterized by dominant themes identified through keyword analysis, reflecting the diversity of AI applications in sustainability contexts.
The Role of AI Methods
Applying a machine-learning taxonomy, AI methods serve key functions in SDG research:
- Forecasting: Prominent in clean energy and vegetation studies, supporting resource management and environmental monitoring.
- System Optimization: Common in clean energy to improve efficiency and performance.
- Data Mining and Remote Sensing: Extracting insights from unstructured data, especially in healthcare and environmental monitoring.
- Accelerated Experimentation and Simulation: Facilitates rapid analysis in clean energy and healthcare research.
The choice of AI algorithms aligns closely with sector-specific needs and data availability. For example, deep learning models like large language models have become distinct areas of focus within AI applications.
However, AI is often studied as an object rather than applied tool in fields like education and industry, indicating room for more empirical research in these areas.
Discussion
Despite growth in AI and SDG research, few studies truly merge advanced AI with deep sustainability knowledge. Most work either focuses on AI techniques or on individual SDGs without fully integrating both to tackle complex sustainability challenges.
Publication trends show increased interest since 2019, with notable geographic differences in focus areas. For example, China emphasizes climate change, clean energy, and education, moving beyond purely economic goals.
Still, official UN reports highlight that half of SDG targets are off-track, with insufficient data for many goals. AI use is strong in health, education, and environmental modeling but minimal in poverty reduction (SDG 1).
Machine learning, deep learning, and evolutionary algorithms have advanced environmental sustainability research, especially in processing large datasets for vegetation and water management. Yet, AI remains underused in social sustainability domains like policymaking, education for sustainable development, and social equity.
Addressing these gaps requires broader data frameworks including qualitative variables, interdisciplinary collaboration bridging AI and social sciences, and algorithm innovation tailored to social complexities. Some AI users maintain techno-optimistic or ecomodernist views, expecting technology to solve sustainability challenges with minimal socio-political change.
Limitations of This Review
This analysis has some constraints:
- Search terms tied to the SDG framework may have excluded relevant studies addressing AI and sustainability outside this political context.
- Focusing on highly cited articles introduces bias, potentially missing emerging research.
- Important findings published in policy reports or conference proceedings may be delayed due to peer-review cycles.
Despite these limits, the identified patterns offer a reliable picture of current research trends at the AI-SDG interface.
Methodology Summary
The study used content analysis and bibliometric methods to analyze AI applications in sustainable development research. Metadata from Scopus was queried in January 2024 for each of the 17 SDGs, yielding 14,423 articles. The 100 most cited articles per SDG were screened for relevance to AI and sustainability.
Full texts were obtained and analyzed using a coding scheme developed inductively by multiple reviewers to ensure consistency. Articles were classified as empirical or conceptual/review, with empirical articles further analyzed through hierarchical clustering of conceptual vocabulary extracted from texts.
This approach highlighted thematic groups reflecting disciplinary and topical diversity within the literature.
Further Information
Additional research design details and data are available in the Supplementary Information, including the full list of analyzed articles and group classifications. The dataset can be accessed via GitHub.
References
- Transforming Our World: The 2030 Agenda for Sustainable Development (United Nations, 2015).
- Allen, C., Metternicht, G. & Wiedmann, T. Initial progress in implementing the Sustainable Development Goals (SDGs): a review of evidence from countries. Sustain. Sci. 13, 1453–1467 (2018).
- Abson, D. J. et al. Leverage points for sustainability transformation. Ambio 46, 30–39 (2017).
- Fuso Nerini, F. et al. Connecting climate action with other Sustainable Development Goals. Nat. Sustain. 2, 674–680 (2019).
- Vinuesa, R. et al. The role of artificial intelligence in achieving the Sustainable Development Goals. Nat. Commun. 11, 233 (2020).
Your membership also unlocks:
