Artificial intelligence (AI) has emerged as a transformative technology with the potential to reshape how societies understand, manage, and respond to complex environmental, economic, and social challenges. The accelerating impacts of climate change, resource scarcity, environmental degradation, rapid urbanization, and increasing demands for resilient infrastructure require approaches capable of integrating heterogeneous information, identifying complex patterns, improving prediction, and supporting timely decision-making. In this context, AI offers powerful capabilities for extracting knowledge from increasingly large and diverse datasets and translating that knowledge into actionable solutions for sustainable development. Recent research has demonstrated that AI is already being applied across a broad spectrum of sustainability-related challenges, including environmental monitoring, resource optimization, energy management, climate action, and sustainable urban development [1,2,3].
The relationship between AI and sustainability, however, is inherently multi-dimensional. AI can improve resource efficiency, environmental monitoring, climate adaptation, agricultural productivity, industrial processes, and urban management, but its development and deployment also introduce challenges related to computational energy consumption, water use, data requirements, algorithmic bias, transparency, governance, and equitable access. Recent assessments have consequently emphasized the dual nature of AI: it can serve as a powerful instrument for advancing sustainable development while simultaneously generating environmental and societal pressures that must be addressed [4,5,6]. For example, recent analyses have highlighted the substantial energy, water, and carbon implications associated with the rapid expansion of AI infrastructure, while other studies have shown that the environmental effects of AI adoption may vary across economic and institutional contexts [4,7]. Thus, advancing sustainable development through AI requires not only demonstrating technical performance but also considering the broader environmental, economic, and social implications of AI-enabled systems.
Against this background, the Special Issue entitled “Advancing Sustainable Development Through Artificial Intelligence (AI)” brings together eight contributions addressing diverse aspects of AI-enabled sustainability. The published studies encompass large language models (LLMs), machine learning, deep learning, computer vision, optimization, remote sensing, unmanned aerial vehicles (UAVs), virtual assistants, and AI-supported qualitative research. Collectively, they demonstrate how AI can support sustainable development across agriculture, land resource management, environmental and climate research, energy systems, industrial operations, and urban development, while also highlighting important challenges associated with responsible and effective deployment.
A particularly important contribution to the Special Issue is the systematic scoping review by Su et al., which examines how LLMs can contribute to environmental sustainability. By synthesizing 20 studies identified across six major academic databases, the authors organize existing applications into six core domains: green transition, carbon-emission management, air-quality assessment, smart-city operations, map analysis, and human cognition and behavioural observation. Their findings illustrate the rapidly expanding role of LLMs in environmental applications while also identifying a fundamental sustainability paradox. Although LLMs can support environmental problem solving and low-carbon transitions, their training and operation require substantial computational resources. This observation is consistent with recent evidence concerning the energy, water, and carbon implications of AI infrastructure and reinforces the need to develop AI systems whose own environmental impacts are considered alongside the benefits they provide [4,5].
The human dimension of climate action is further explored by Sprague et al., who investigate the potential role of AI in qualitative climate change research. By comparing human-only, human-then-AI, AI-then-human, and AI-only approaches to the analysis of interviews with New York City residents, the study demonstrates the potential of AI to complement human interpretation of complex social information. The AI-then-human workflow produced the most comprehensive and contextually accurate thematic results, identifying perceptions related to personal responsibility, community support, governmental and corporate responsibility, future generations, climate impacts, local climate literacy, and competing interests. This contribution emphasizes that sustainable development is not solely a technological challenge; it also depends on understanding human perceptions, experiences, values, and institutional contexts. The findings further support the emerging view that human–AI collaboration can be preferable to fully automated approaches when sustainability decisions involve complex social and contextual considerations.
Several contributions focus on agriculture, where AI can directly support resource efficiency, productivity, and resilience. Koohikeradeh et al. developed a Long Short-Term Memory (LSTM) model integrated with ERA5 data and Normalized Difference Moisture Index (NDMI) information to predict field-scale soil moisture across six commercial potato production sites in Quebec over a five-year period. The model demonstrated very high agreement between observed and predicted soil moisture dynamics, highlighting the potential of deep learning for real-time soil moisture monitoring and irrigation management. Such approaches can contribute to more efficient agricultural water use while helping producers respond to spatial and temporal variability in soil moisture.
Complementing this data-driven prediction approach, Ronzhin et al. investigated UAV-assisted pesticide application in potato cultivation under waterlogged soil conditions, where conventional ground-based machinery can face severe operational limitations. Their integrated UAV framework incorporated aerial soil analysis, orthophotomap generation, weed segmentation, targeted spraying, and yield assessment. The field experiment demonstrated improved plant condition and a reported 15.2% increase in yield relative to the control plot, while maintaining a comparable proportion of marketable tubers. This study illustrates how AI-enabled remote sensing and UAV technologies can support timely agricultural interventions and strengthen the resilience of production systems under challenging environmental conditions.
At a broader spatial scale, Nguyen and Vu addressed an important challenge in AI-based land resource management: the ability of models to generalize to geographical domains that were not represented during training. Their F-SegFormer framework combines a Variational Information Bottleneck module with the SegFormer architecture to improve feature selection and generalization to unseen domains, while Online Hard Example Mining accelerates model convergence. The study highlights the importance of developing AI systems that remain reliable when deployed beyond their original training environments. Such transferability is particularly important in remote sensing and land management, where environmental conditions can vary substantially across regions and where extensive target-domain training data may not always be available.
The Special Issue also demonstrates the potential of AI to improve the sustainability of energy-intensive industrial systems. Huang et al. developed a data-driven many-objective optimization framework for wet flue gas desulfurization at a 600 MW ultra-supercritical coal-fired power plant. Using operational data, the authors employed an extreme random forest model to predict desulfurization efficiency, unit power consumption, limestone supply, and operating cost. These predictive models were subsequently combined with differential evolution and NSGA-III to identify Pareto-optimal operating conditions. The optimized operation improved desulfurization efficiency while reducing energy, material, and operational costs. This study illustrates the value of combining machine learning with multi-objective optimization when environmental, economic, and resource-efficiency objectives must be considered simultaneously.
In the renewable energy domain, Gavilánez et al. developed an AI-based virtual assistant for solar-radiation prediction and sustainable energy applications. Their hybrid framework combined Random Forest, CatBoost, Gradient Boosting, and a one-dimensional convolutional neural network and achieved strong predictive performance. Importantly, the model was integrated into a chatbot capable of providing real-time personalized information and alerts. This approach illustrates how AI can bridge the gap between predictive modelling and accessible decision-support services, with potential applications in renewable energy planning, environmental awareness, and public health protection. More broadly, recent research suggests that AI can influence environmental sustainability through interactions with energy efficiency, renewable energy adoption, and broader technological and institutional conditions [7,8].
The urban dimension of AI-enabled sustainability is examined by Cina et al., who explore the role of AI-powered predictive modelling in infrastructure optimization, energy management, environmental monitoring, and climate adaptation. Their analysis emphasizes that technological capability alone does not guarantee sustainable outcomes. Data limitations, algorithmic bias, ethical concerns, and governance challenges must be addressed to ensure that AI-supported urban development remains transparent, inclusive, and accountable. This perspective is consistent with recent scholarship emphasizing that sustainable AI must encompass environmental, social, and economic dimensions rather than focusing exclusively on computational performance or environmental efficiency [6,9].
Taken together, the eight contributions reveal several important cross-cutting insights. First, AI applications are increasingly moving from prediction toward actionable decision support and intervention. Soil moisture prediction can inform irrigation, UAV-based analysis can guide agricultural operations, machine learning can optimize industrial processes, and AI-based assistants can translate environmental predictions into accessible information. However, the usefulness of such decision support depends on whether the spatial and temporal scales of AI-derived information are compatible with the processes being assessed and the decisions being made. Scale-aware approaches are therefore essential for translating increasingly sophisticated environmental observations and predictions into effective sustainability actions [10].
Second, heterogeneous data integration is becoming central to AI-enabled sustainability. The contributions in this Special Issue utilize satellite and remote-sensing observations, meteorological data, field measurements, industrial operational data, UAV imagery, and qualitative information. Such integration can provide a more comprehensive representation of complex environmental and societal systems, but it also introduces challenges concerning data quality, interoperability, uncertainty, privacy, and computational requirements. Recent research similarly emphasizes the importance of contextualization and interdisciplinary sustainability expertise when applying AI to Sustainable Development Goal-related research [3].
Third, robustness, generalizability, and scale-awareness remain critical challenges. AI systems trained under specific geographical, climatic, operational, or social conditions may not necessarily perform equally well when transferred to new contexts. The work on unseen-domain land resource management directly addresses this issue, but the challenge is relevant across sustainability applications. Moreover, sustainability-oriented AI systems must account for potential mismatches between the spatial and temporal scales of observations, measurements, underlying environmental processes, and management decisions. Recent work on climate change risk assessment for heritage demonstrates how such scale mismatches can obscure risk understanding and limit the effectiveness of adaptation and management actions [10]. Future research should therefore place greater emphasis on domain generalization, uncertainty quantification, scale-aware modelling, independent validation, and transparent reporting of model limitations.
Fourth, human–AI collaboration should be considered an important component of sustainable AI. The qualitative climate change study in this Special Issue demonstrates how human interpretation and AI-based analysis can complement one another. Similar principles can be applied to environmental management, urban planning, agriculture, and energy systems, where domain expertise remains essential for interpreting model outputs and evaluating their implications. AI should therefore generally be viewed as a tool for augmenting human capabilities rather than as a substitute for contextual expertise and institutional judgement.
Finally, and perhaps most importantly, the sustainability of AI itself must be considered. Recent studies have quantified substantial energy, water, and carbon implications associated with expanding AI infrastructure and have argued for greater attention to the full environmental lifecycle of AI systems [4,5]. Other research has similarly emphasized the need to evaluate AI through broader sustainability frameworks that encompass environmental, social, economic, and governance dimensions [6,9]. This perspective is particularly important for LLMs and other computationally intensive systems, for which the environmental benefits achieved through applications must be weighed against the resources required for their development and operation.
These observations point toward several priorities for future research. Greater integration of AI with Internet of Things networks, Earth observation, UAVs, real-time sensors, digital twins, and LLMs could enable increasingly adaptive sustainability-management systems. At the same time, AI systems should be designed with sustainability objectives embedded throughout their development and deployment cycles. Model accuracy should be considered alongside computational efficiency, energy and water consumption, data requirements, scalability, and environmental footprint. Recent work on measuring the environmental sustainability of AI pipelines further underscores the need for systematic approaches to quantify and manage these impacts [11].
Explainability, uncertainty quantification, fairness, and trust should also become integral components of sustainability-oriented AI. Decision-makers in agriculture, environmental management, infrastructure, and urban planning need to understand not only what an AI system predicts but also the uncertainty associated with its predictions and the factors influencing its outputs. Responsible AI should additionally address algorithmic bias, data governance, privacy, accessibility, and equitable distribution of benefits. These considerations are particularly important because the effectiveness and consequences of AI adoption can differ substantially across regions, sectors, and socioeconomic contexts [6,7,8,9].
Interdisciplinary collaboration will therefore be essential. Meaningful progress will require sustained interaction among AI researchers, environmental scientists, engineers, agricultural specialists, urban planners, social scientists, policymakers, and local communities. Such collaboration can ensure that AI systems respond to genuine sustainability needs rather than merely demonstrating technological capability. It can also help align AI development with broader Sustainable Development Goals and ensure that technological innovation contributes to measurable environmental and societal outcomes.
Overall, the eight contributions in this Special Issue demonstrate the broad and rapidly expanding potential of AI to advance sustainable development. From LLMs and climate research to precision agriculture, UAV-based management, land resource monitoring, industrial optimization, renewable energy, and sustainable urban planning, the studies illustrate how AI can transform complex data into knowledge and decision-support capabilities. At the same time, they demonstrate that the contribution of AI to sustainability cannot be assessed solely through predictive accuracy or technological novelty. Robustness, transferability, human oversight, ethical governance, accessibility, resource efficiency, and the environmental footprint of AI itself must also be considered. The future of sustainable AI therefore lies not simply in developing more powerful models but in developing more responsible, efficient, interpretable, inclusive, and context-aware systems. We hope that this Special Issue will stimulate further interdisciplinary research and encourage the responsible adoption of AI technologies to address pressing sustainability challenges while ensuring that AI itself evolves along a sustainable pathway. We also hope that this collection will provide valuable insights for researchers, practitioners, policymakers, and other stakeholders working at the intersection of artificial intelligence and sustainable development.
Author Contributions
E.G.: Writing—Original draft preparation; H.B. and S.G.: Writing—Review and editing. All authors have read and agreed to the published version of the manuscript.
Funding
This research received no external funding.
Institutional Review Board Statement
Not applicable.
Informed Consent Statement
Not applicable.
Data Availability Statement
No new data were created or analyzed in this study. Data sharing is not applicable to this article.
Acknowledgments
The Guest Editors of the Special Issue “Advancing Sustainable Development Through Artificial Intelligence (AI)” sincerely thank all the authors whose valuable contributions were published in this Special Issue. Their high-quality work and contributions were instrumental to the success of this edition.
Conflicts of Interest
The authors declare no conflicts of interest.
List of Contributions
- Su, X.; Liu, T.; Pang, P.; Luo, Y.T.; Wong, D. How Can Large Language Models Drive Environmental Sustainability? A Systematic Scoping Review. Sustainability 2026, 18, 4327. https://doi.org/10.3390/su18094327.
- Sprague, N.L.; Meltzer, G.Y.; Dandeneau, M.L.; De Los Santos, D.; O’Neil, D.B.; Kim, A.K.; Parisi, A.; Araujo, S.; Ekenga, C.C.; Siegel, E.L.; et al. A Role for Artificial Intelligence (AI) in Qualitative Research? An Exploratory Analysis Examining New York City Residents’ Perceptions on Climate Change. Sustainability 2025, 17, 10459. https://doi.org/10.3390/su172310459.
- Koohikeradeh, E.; Jose Gumiere, S.; Bonakdari, H. NDMI-Derived Field-Scale Soil Moisture Prediction Using ERA5 and LSTM for Precision Agriculture. Sustainability 2025, 17, 2399. https://doi.org/10.3390/su17062399.
- Ronzhin, A.; Ryabinov, A.; Shkodina, E.; Saveliev, A.; Cherskikh, E.; Figurek, A. UAV-Assisted Pesticide Application in Potato Cultivation Under Waterlogged Soil Conditions: Orthophotomap-Based Monitoring and Field Assessment. Sustainability 2026, 18, 4567. https://doi.org/10.3390/su18094567.
- Nguyen, M.-H.; Vu, C.-C. F-Segfomer: A Feature-Selection Approach for Land Resource Management on Unseen Domains. Sustainability 2025, 17, 2640. https://doi.org/10.3390/su17062640.
- Huang, J.; Zeng, Z.; Hong, F.; Yang, Q.; Wu, F.; Peng, S. Sustainable Operation Strategy for Wet Flue Gas Desulfurization at a Coal-Fired Power Plant via an Improved Many-Objective Optimization. Sustainability 2024, 16, 8521. https://doi.org/10.3390/su16198521.
- Gavilánez, T.; Zamora, N.; Navarrete, J.; Vega, N.; Vergara, G. AI-Based Virtual Assistant for Solar Radiation Prediction and Improvement of Sustainable Energy Systems. Sustainability 2025, 17, 8909. https://doi.org/10.3390/su17198909.
- Cina, E.; Elbasi, E.; Elmazi, G.; AlArnaout, Z. The Role of AI in Predictive Modelling for Sustainable Urban Development: Challenges and Opportunities. Sustainability 2025, 17, 5148. https://doi.org/10.3390/su17115148.
References
- Gohr, C.; Rodríguez, G.; Belomestnykh, S.; Berg-Moelleken, D.; Chauhan, N.; Engler, J.-O.; Heydebreck, L.V.; Hintz, M.J.; Kretschmer, M.; Krügermeier, C.; et al. Artificial intelligence in sustainable development research. Nat. Sustain. 2025, 8, 970–978. [Google Scholar] [CrossRef] [Scilit]
- Cho, H.; Ackom, E. Artificial Intelligence (AI)-driven approach to climate action and sustainable development. Nat. Commun. 2025, 16, 1228. [Google Scholar] [CrossRef] [Scilit] [PubMed]
- Xiao, T.; Fuso Nerini, F.; Matthews, H.D.; Tavoni, M.; You, F. Environmental impact and net-zero pathways for sustainable artificial intelligence servers in the USA. Nat. Sustain. 2025, 8, 1541–1553. [Google Scholar] [CrossRef] [Scilit]
- Angrisani, L.; D’Arco, M.; De Benedetto, E.; Duraccio, L.; Esposito, I.; Tedesco, A. Environmental sustainability of artificial intelligence: A GUM-based, user-centered measurement framework. Measurement 2026, 268, 120654. [Google Scholar] [CrossRef] [Scilit]
- Wang, Q.; Liu, T.; Li, R. Artificial intelligence and environmental sustainability: Investigating the AI-EKC nexus for SDG 7 and SDG 13. Sustain. Dev. 2026, 34, 1141–1166. [Google Scholar] [CrossRef] [Scilit]
- Waelen, R.; van Wynsberghe, A. Considering the social and economic sustainability of AI. Sci. Eng. Ethics 2025, 31, 19. [Google Scholar] [CrossRef] [Scilit] [PubMed]
- Chen, W.; Victor, G. Artificial intelligence, financial development, and energy efficiency: Synergistic pathways to environmental sustainability in high-income economies. Sustain. Dev. 2026, 34, 4916–4931. [Google Scholar] [CrossRef] [Scilit]
- Zambrano-Monserrate, M.A.; Hernández Soto, G.; Subramaniam, Y. Can artificial intelligence contribute to sustainable development by reducing the impact of energy supply on CO2 emissions? Energy Sources Part B Econ. Plan. Policy 2025, 20, 2558529. [Google Scholar] [CrossRef] [Scilit]
- Bansal, C. AI ethics and sustainability: Accelerating paradigm shifts toward sustainable development. J. Strategy Innov. 2025, 36, 200537. [Google Scholar] [CrossRef] [Scilit]
- Orr, S.A.; Zhao, S.; Thomas, H.; Guzman, P.; Hotchkiss, C.; Seekamp, E.; Xiao, X.; Aquilani, C.; Ghaderpour, E.; Grau-Bové, J.; et al. Challenges of scale in assessing the risks of climate change for heritage. Theor. Appl. Climatol. 2026, 157, 449. [Google Scholar] [CrossRef] [Scilit] [PubMed]
- Whig, P.; Gupta, S.K.; Nadikattu, R.R.; Sharma, P. Application of AI in environmental sustainability. In Artificial Intelligence-Driven Models for Environmental Management; Kulkarni, S., Ed.; Wiley: Hoboken, NJ, USA, 2025. [Google Scholar] [CrossRef] [Scilit]
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. |
© 2026 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license.