The Evolution of AI Applications in the Energy System Transition: A Bibliometric Analysis of Research Development, the Current State and Future Challenges

Abstract

The transformation of energy markets is at a crossroads in the search for how they must evolve to become ecologically friendly systems and meet the growing energy demand. Currently, methodologies based on bibliographic data analysis are supported by information and communication technologies and have become necessary. More sophisticated processes are being used in energy systems, including new digitalization models, particularly driven by artificial intelligence (AI) technology. In the present bibliographic review, 342 documents indexed in Scopus have been identified that promote synergies between AI and the energy transition (ET), considering a time range from 1990 to 2024. The analysis methodology includes an evaluation of keywords related to the areas of AI and ET. The analyses extend to a review by authorship, co-authorship, and areas of AI’s influence in energy system subareas. The integration of energy resources, including supply and demand, in which renewable energy sources play a leading role at the end-customer level, now conceived as both producer and consumer, is intensively studied. The results identified that AI has experienced notable growth in the last five years and will undoubtedly play a leading role in the future in achieving decarbonization goals. Among the applications that it will enable will be the design of new energy markets up to the execution and start-up of new power plants with energy control and optimization. This study aims to present a baseline that allows researchers, legislators, and government decision-makers to compare their benefits, ambitions, strategies, and novel applications for formulating AI policies in the energy field. The developments and scope of AI in the energy sector were explored in relation to the AI domain in parts of the energy supply chain. While these processes involve complex data analysis, AI techniques provide powerful solutions for designing and managing energy markets with high renewable energy penetration. This integration of AI with energy systems represents a fundamental shift in market design, enabling more efficient and sustainable energy transitions. Future lines of research could focus on energy demand forecasting, dynamic adjustments in energy distribution between different generation sources, energy storage, and usage optimization.

1. Introduction

The global imperative for sustainable energy systems has positioned artificial intelligence (AI) as a transformative force driving energy transition. Recent unprecedented advances in AI technologies, particularly in machine learning (ML) and deep learning (DL), have revolutionized our ability to manage and interpret massive volumes of energy-related data with exceptional accuracy [1]. While AI applications span numerous sectors, their impact on energy systems has been particularly significant [2,3], enabling sophisticated solutions from grid optimization to demand forecasting [4]. These developments align with Industry 4.0 principles, where AI technologies are increasingly integrated with traditional energy infrastructure, creating new energy management and control paradigms. The integration of AI in energy planning and operations has demonstrated remarkable potential in enhancing system efficiency and facilitating the transition toward sustainable energy frameworks [5], establishing itself as a cornerstone technology in the global effort to achieve clean energy objectives.

In light of the AI capabilities outlined above, the global energy sector faces mounting challenges, with the International Energy Agency (IEA) projecting electricity demand to increase by 3.4% annually until 2026 [6] This unprecedented growth, coupled with ambitious decarbonization targets [7], necessitates a fundamental transformation of energy systems. While advanced economies, particularly China, are implementing large-scale electrification projects [8,9], current initiatives in residential and transport sectors [10] fall short of meeting the IEA’s 2050 net-zero emissions scenario for limiting global warming to 1.5 °C [11]. While the IEA’s net-zero scenario presents technical pathways, the economic feasibility of this transition relies heavily on the successful implementation of AI-driven solutions that can optimize infrastructure investments and operational costs. AI-enabled energy market design tools are emerging as critical enablers of this transformation [12,13], though they face implementation challenges, including their own energy consumption footprint [14]. However, when combined with increasingly cost-competitive renewable technologies [15,16], particularly wind and solar [17], these AI solutions provide a viable pathway toward achieving both rapid decarbonization and economic sustainability.

Building on these AI-enabled solutions, empirical research has extensively documented AI’s transformative impact on energy transition and decarbonization goals [18]. Current applications encompass the entire energy sector, from strategic market planning to comprehensive system operations [19], supporting the technical and economic feasibility of 100% renewable energy systems by 2050 [20]. Two recent studies particularly illuminate AI’s potential in this transformation. Iglesias-Sanfeliz Cubero et al. [21] systematically evaluated how AI-powered neural networks optimize variable renewable energy integration through advanced architectural frameworks and methodological innovations. Complementing this technical perspective, Wang et al. [22] provided quantitative evidence of AI’s economic impact, demonstrating that a 1% increase in AI implementation yields a 0.032% growth in high-quality energy development (HED), thus establishing a clear pathway between AI adoption and clean technology advancement.

Expanding upon these empirical findings, systematic reviews have confirmed renewable energy’s fundamental role in energy transition [23], with AI serving as a key enabler through its advanced analytical capabilities [24]. The practical implementation of AI-driven solutions demonstrates remarkable scalability, successfully addressing challenges from isolated community microgrids [25,26] to complex national energy networks [27,28]. These implementations rely on sophisticated algorithms [29] that have transformed energy storage and production management through precise monitoring and control systems [30,31]. Machine learning applications have proven especially valuable in addressing multi-temporal challenges, from immediate operational decisions to long-term strategic planning while integrating crucial economic and financial parameters [32,33,34]. This integration extends to advanced analytics for price optimization, risk assessment, and macroeconomic forecasting [35,36,37]. Recent bibliometric analyses have particularly highlighted the emergence of artificial neural networks (ANNs) and genetic algorithms (GAs) as dominant technological drivers [38,39], reinforcing AI’s critical role in enabling sustainable energy transformations.

While AI adoption in energy transition strongly correlates with economic development levels [40], significant research gaps remain in understanding its full potential. Although technologically advanced nations are leading sustainable energy implementations and achieving faster progress toward GHG reduction targets [41,42], the systematic integration of AI in energy planning requires further investigation [43]. The current literature predominantly focuses on individual AI tools rather than their holistic application in energy markets [44,45]. To address this limitation, this study conducts a comprehensive bibliometric analysis of AI applications in energy transition from 1990 to 2024, using VOSViewer software (version number 1.6.20) to visualize and analyze research patterns. This systematic review aims to (1) map the evolution of AI applications in energy transition research, (2) identify emerging trends and research hotspots, and (3) provide evidence-based insights for developing long-term roadmaps toward sustainable energy futures. The study’s significance lies in its contribution to understanding how AI can be effectively deployed to accelerate the global energy transition.

The main objective of this study is to evaluate systematically how AI tools are contributing to energy planning and transition through analysis of high-impact articles indexed in Scopus. Our methodology combines bibliometric analysis of research patterns, citation networks, and thematic evolution with qualitative assessment of implementation strategies. This dual approach enables the identification of both theoretical advances and practical applications in the field. The findings aim to provide decision-makers with evidence-based insights for transforming energy systems toward sustainable configurations through advanced digital technologies while addressing current implementation challenges.

This study makes several novel contributions to the field. First, it develops an integrated analytical framework that combines quantitative bibliometric analysis with qualitative evaluation, utilizing VOSviewer to reveal previously unidentified patterns in AI–energy research networks. Second, it identifies emerging trends and underexplored areas in AI applications for energy systems, providing a roadmap for future research directions. Third, it comprehensively assesses AI’s role in accelerating the transition toward 100% renewable energy systems, including implementation barriers and potential solutions. Considering both technical and economic constraints, the findings provide practical guidelines for decision-makers to implement AI-driven solutions in their energy transition strategies.

The contributions of this study are substantial. Through analysis of 342 high-impact documents indexed in Scopus, we map international research collaborations and evaluate national AI strategies in energy transition, mainly focusing on China’s emerging leadership role. Additionally, we develop a systematic classification framework for AI applications across the energy supply chain, categorizing specific AI domains and their practical implementations.

The remainder of this paper is structured as follows: Section 2 reviews the state of the art, examining the current state of AI applications to support energy transition and their challenges. Section 3 outlines the methodology used in this bibliometric analysis. Section 4 presents a comprehensive analysis of the results, including publication statistics and network analysis. Section 5 discusses the implications of AI in energy transition research. Finally, Section 6 concludes with the key findings and future research directions.

2. A Review of the State of the Art

2.1. Current State of AI Applications to Support TE

Industry 4.0 and the different applications that include AI developments are becoming key when it comes to creating innovation; in the electrical sector and the renewable energy field, its influence is notable and will continue to increase [46]. As of December 2024, the AI market is valued at USD 146.1 billion [47] and is projected to reach USD 2740.46 billion by 2032 [48]. AI is breaking patterns and predictions, a new scientific discipline aimed at emulating, expanding, and improving human intelligence. However, the warmth that a human being can transmit is still far from being achieved, and it remains to be seen whether AI can achieve this in the future [49]. The World Trade Organization (WTO) has declared that AI is set to usher in profound changes in business and society itself [50]. By employing methodologies based on case studies, AI seeks to replicate human cognitive functions, ranging from autonomous learning to decision-making, by applying increasingly developed and efficient programming through advanced algorithm processing [51]. This technology can effectively replace certain facets of human labor, both physical and cognitive, while offering users efficient auxiliary capabilities [52,53]. The technological development in full swing will impact the improvement in methodologies and the renewal of energy markets, thus contributing to a significant reduction in polluting fossil fuels, an aspect of transcendental importance for humanity.

The conventional power grid emerged as a need to provide electricity, and related studies were not strongly framed in terms of production being sustainable [54], nor was it designed to receive 100% renewable energy sources (RESs) [55]. The diversification of renewable energy sources and the importance of integrating much larger percentages (especially wind and solar) create challenges to supply adequate and continuous service to meet the electrical demand [17]. Advances in AI, including deep learning, machine learning, and other tools, are energizing the global energy sector as much as in other areas [56,57]. In several countries and regions, the indispensable need to use AI to carry out activities in a way that is more efficient than done so by a human is beginning to be considered, such as controlling, forecasting, and efficiently operating the electrical system. This situation in some parts of the world has brought about deep analysis, discussions, and controversy [58]. AI also allows for the design of more robust systems that include 100% renewable systems in the medium and long term [23]. Other studies have focused on using AI for efficient control of equipment [59], including the inverter and photovoltaic (PV) solar panels under appropriate positioning [60], maximizing energy generation capacity [61].

China’s development in various aspects is impressive, and AI does not escape being a global protagonist [62]. Knowing that this field’s development is exceptional and a transcendental tool, in 2017, the Chinese government issued the “Development Plan for a New Generation of Artificial Intelligence” to make China a leading center of AI innovation by 2030 [63]. According to official Chinese statistics, in 2021, the country developed 70.9% of the world’s AI patents, which means they are an AI giant [64]. Other critics argue that while China wins the race in volume, with approximately 13,000 patents, the United States, with 8609 patents, far outstrips it in terms of impact [65]. On the other hand, also recognizing that China is the largest energy consumer in the world, the transition to renewable energy is being promoted to achieve carbon neutrality [66]. In 2023, the National Energy Administration issued the “Guiding Opinions on Accelerating the Development of Digital and Intelligent Energy”, proposing AI to accelerate the transition to renewable energy [67]. The analysis of China is key since several regions of the world have a notable influence on goods and services coming from this world power, which implies that in the coming years, there will be a dependence on several aspects of development [68].

For a more detailed understanding, Figure 1 integrates the AI strategies of several countries, including the US, China, and the European Union. Each has presented its long-term strategies for AI development and usability [69,70,71]. Talent development, scientific research, and empowerment through energy innovation prevail among all countries.

Among the methodologies used for the design of energy markets in transition, the need for specialized software support stands out, including EnergyPLAN [72], ELENA [73], LEAP [74], Message [75], and HOMER [76], among others. Several authors highlight the use of AI as a link to energy transition tools [77,78,79]. A Višković et al. [80] recognize AI as a key element in the energy transition. After analyzing the methodologies of other research focused on the design of continental and island energy markets, they evaluate historical data before designing the future market [27,81]. Other researchers use satellite systems that identify the energy potential of different territories [82]. Subsequently, detailed analyses of energy resources can be carried out using AI to determine how the energy system should change in the medium and long term [25].

2.2. Challenges of AI Applications to Support TE

To ensure that AI effectively drives the energy transition, policies need to be developed that support its responsible, inclusive, and ethical implementation, as emphasized in [32,83]. One of the priorities should be the regulation of algorithms used in the energy sector, ensuring that they are transparent and do not introduce biases or inequalities in energy distribution [84]. In addition, tax incentives should be established to encourage technological innovation, supporting both large companies and startups working on AI-based solutions to improve energy efficiency and the integration of renewable energy [41]. At the infrastructure level, policies should promote the development of smart grids which leverage AI to manage energy distribution more efficiently [85]. It is essential that these technologies are accessible to all regions, avoiding the exclusion of vulnerable or low-income communities [86]. Likewise, the creation of job training and retraining programs will be key to ensuring that the workforce is prepared to adapt to the new challenges and opportunities that AI brings [87]. Finally, policies should foster international cooperation, ensuring that the benefits of AI in the energy transition are shared globally and contribute to the sustainable development of all countries, especially developing ones [88].

Most renewable energy systems that plan to use AI are deconcentrated and decentralized. Part of the optimization of the systems involves avoiding long transmission lines as much as possible so that the combinations of renewable energy sources are closer to the load. Given the interest generated by the high performance of AI, decarbonization processes, and the implementation of emerging technologies, many other energy companies are using AI to ensure a balance between supply and demand [89]. The challenge also consists of limiting the share of energy from fossil fuels, since demand is constantly growing. AI has also focused on looking for customer behavior patterns, and technical and economic analysis helps to analyze a large amount of data produced by the power sector [90]. Other applications of AI have also been identified, such as energy storage and the operation of autonomous networks [87], for example, on islands to guarantee service continuity at maximum load [88]; simulations and creation of scenarios are carried out in different meteorological conditions [91].

Research is expected to focus on the integration of AI throughout the energy transition process, optimizing the generation and distribution of renewable energy through accurate predictions on the availability of sources such as solar and wind [92]. Its integration with smart grids will allow fir efficient management of energy supply and demand, facilitating storage and distribution according to needs [93]. In addition, AI will promote energy efficiency in buildings and cities, automatically controlling consumption and improving sustainability [94]. As it advances, it will empower emerging technologies, such as carbon capture and the development of new battery materials. It will also facilitate energy decentralization, allowing for autonomous management of distributed energy systems and optimizing energy transactions through blockchain [28]. Likewise, AI will support the creation of informed and equitable public policies, by analyzing large volumes of data to predict consumption and emissions trends [95,96,97]. Together, these innovations will contribute to a more agile, efficient, and sustainable energy transition [98].

3. Methodology

This study uses a bibliometric approach to visualize the analysis based on the statistics of the number of published articles and the main background surrounding the documents [99]. Bibliometric methods are widely used to evaluate scientific results, analyze the results by area of knowledge, country, and authors, structure scientific knowledge graphs, and identify the direction of development of specific fields [100]. Vosviewer is one of the most recognized bibliometric analysis tools, but there are also others, such as CiteSpace (version number 6.3.R1) [101], Ucinet (version number 6.806) [102], and RStudio (version number 2024.12.1) [103], used in the research field.

In this study, Vosviewer visualization software was used as a powerful analysis tool [99] to extract and analyze the data. In this way, the databases are systematically analyzed, including keywords and the timeline distribution in artificial intelligence and energy transition fields. VOSviewer uses the VOS mapping technique to build two-dimensional distance-based maps, which can display maps such as authors or journals satisfactorily compared to most bibliometric programs [104].

A comprehensive review of the literature related to AI and the energy transition was conducted; the search databases are Scopus and several search attempts were made, including filtering between the following: TS = (“Artificial intelligence *” OR “AI *” OR “Data Analytics *” OR “Business Intelligence *” OR “Artificial intelligent *” OR “Big Data *” OR “Generative AI *” OR “Machine learning *” OR “Deep learning *”) OR “Multimodal model *” OR “NLP *” OR “Computer Vision *” OR “LLM *” OR “Neural networks *” AND TS = (“Energy Transition *” OR “Energy Planning *” OR “Long Term *” OR “Energy *” OR “Zero Carbon *” OR “Electricity *” OR “Decarbonization *” OR “Renewables *” OR “Energy policy *” OR “Energy efficiency *” OR “Smart Energy *” OR “Wind *” OR “solar *”).

The search in the Scopus database yielded 342 documents, indicating a notable growth in publications related to AI and the energy transition over the last four years. Figure 2 below graphically presents the methodology used in the review.

The study consists of three clearly defined steps. The first step consists of analyzing the source of data obtained according to the words previously defined in the Scopus database. Step 2 consists of analyzing the results obtained, considered in terms of data pattern findings. Finally, in step 3, these results are discussed. The countries that generate the most scientific contributions are highlighted, in addition to identifying the main authors and the research centers that are giving the greatest impetus to the area of AI&ET.

The degree of advancement of the use of AI in the supply chain of the electrical system has been considered, not limited to a specific part of the electrical system. The number of works dedicated to integrating AI&ET and classified by subareas has been determined. Published and not just accepted research has been identified, appearing in an important scientific platform called Scopus. Scopus has a long-standing reputation and offers a variety of features, including a user interface and bibliometric analyzers. Scopus’ impact includes diverse subject areas, years of publication, and document types, and covers more than 5000 publishers, data on funders, and patents. Its strict policies mean that a journal must pass a series of filters of the highest quality to be indexed [105]. Scopus data are incorporated into other Elsevier research products such as ScienceDirect, IEEE, and Mendeley [106]. Customers incorporate Scopus Custom Data (SCD) and Scopus Application Programming Interfaces (APIs) into their tools. Scopus provides abstract and citation data for the most important research worldwide; typically, renowned researchers try to make their works available on this site for prestige [107]. Scopus indexes three main types of scientific content: conference proceedings, research journals, and books [108]. Bibliometric analyses, such as in the current study, and data analysis simplify the work of students, researchers, and research center administrators.

4. Analysis of Results

4.1. Annual Publication Statistics

Publication statistics are of great importance in identifying the level of development of an area or synergies between areas in a period of time determined by the research authors. The analyses carried out will depend greatly on the most recent trends, and the future will be marked by the most recent expectations that have been generated. Figure 3 presents the annual publications developed from 1990 to the present (end of 2024) between the related areas of artificial intelligence in the energy transition.

Interpreting the data in Figure 3, it can be expressed that the number of documents published in Scopus in the field of AI&ET research shows significant growth, especially in the last five years. Consequently, the research related to these two merged areas can be divided into three stages of development. The first corresponds to the stage that goes from 1990 to 2013 and can be called moderate growth with 46 documents; the second stage can be recognized as significant growth between 2014 and 2019 with 73 articles, which exceeds the previous period based on the number of documents and in a shorter time. Meanwhile, the third stage is considered high growth because the increase in publications has exceeded expectations between 2020 and 2024, with 223 documents. The number of articles published in these five years is relatively small, but the number of publications exceeds 305% compared to the previous stage. Research on artificial intelligence has achieved significant developments, and we see the authors of various studies applying them to the energy transition processes that need to be intensified and incorporate the highest levels of penetration of renewable energies.

Figure 4 below shows the documents published by country, which includes the 15 countries with the most publications in IA&ET. In this sense, China leads by far with 104 documents, followed by the United States with 43 documents, India with 37, and South Korea with 20. Subsequently, there is a block of 11 countries that have published less than 20 documents in the area but are undoubtedly of great contribution to the energy sector.

Meanwhile,

Table 1. Total countries in terms of published documents on AI&ET.

Country	Docs	Country	Docs	Country	Docs	Country	Docs	Country	Docs
China	104	Iraq	7	Tunisia	3	Cyprus	1	Thailand	1
United States	43	Mexico	7	Viet Nam	3	Czech Republic	1	Ukraine	1
India	37	Singapore	7	Austria	2	Congo	1	United Arab Emirates	1
South Korea	20	Egypt	6	Chile	2	Finland	1	Venezuela	1
Iran	18	Poland	6	Colombia	2	Greece	1	Yemen	1
Spain	15	Russian	6	Luxembourg	2	Hungary	1	Undefined	16
Canada	14	Germany	5	Netherlands	2	Ireland	1
France	14	Italy	5	Oman	2	Kuwait	1
Taiwan	12	Brazil	4	Pakistan	2	Lebanon	1
Malaysia	11	Indonesia	4	Portugal	2	Macao	1
United Kingdom	11	Morocco	4	Romania	2	Norway	1
Australia	10	Turkey	4	South Africa	2	Peru	1
Saudi Arabia	10	Algeria	3	Belgium	1	Slovakia	1
Hong Kong	8	Nigeria	3	Bulgaria	1	Switzerland	1
Japan	8	Sweden	3	Croatia	1	Tanzania	1

shows 65 countries that have published at least 1 article, and 16 documents do not have a country-specific definition.

Figure 5 shows the authors’ names with the largest number of published documents on AI&ET. It should be noted that the author Puig, Vicenç has five documents, followed closely by three authors with four documents each, who are Anwar K., Deshmukh S., and Homod, R.Z. A large number of authors have published three documents each: Fan X, Herbert T, Jin Y, López Estrada F.R., Pérez-Perez E.J. Rohani A, Santos Ruiz I, Taki M, Valencia Palomo G, and Yan D. Finally, AI-Fattah S.M. published two documents related to AI&ET.

In the context of the review concerning AI&ET, the authors V. V. Elistratov [109] and Spiru Paraschiv [110] emphasize the need to increase the growth of renewable and low-carbon electricity generation. Much of the literature contemplates initiatives for the inclusion of energy storage systems. The designed systems are primarily optimized, and AI application tools are helpful in allowing these procedures to have a great long-term effect. AI has been made visible under the synergy of modern technologies, such as DL, ML, and advanced neural networks. These structures allow us to identify that the renewable energy potential can be better exploited by adequately extracting the energy available in the different areas of analysis. It has been identified that by using AI, oversizing the planned generation plants can be avoided, making the initial investment more attractive and achieving the decarbonization of the systems as a primary objective.

Figure 6 presents the documents by affiliation published on AI&ET, with the Chinese Academy of Sciences having the largest number of documents on AI&ET. The Ministry of Education of the People follows with nine documents, Fudan University with seven documents, and Southeast University and Tsinghua University with six documents, among the relevant ones.

Other advanced operations in different areas are already within reach of society thanks to AI, including predicting failures in different production systems, which means anticipating their occurrence, saving time and money, and, above all, safeguarding human lives [57]. The applications of AI in the energy industry will expand and deepen in the coming years. They will focus on the entire energy value chain, including at the customer service level, so AI is expected to be integrated into adjacent related areas such as mobility, agriculture, and security, among others. Its scope may still be unimaginable, but the human being must prevail, which is being constantly discussed in different international forums [111].

Figure 7 presents the documents by type. Articles account for 70.5%, conference papers for 21.1%, conference reviews for 3.8%, reviews for 2.6%, book chapters for 1.8%, and data papers for 0.3%.

4.2. Network Analysis

In this subsection, the publications are analyzed from four perspectives. First, the databases on AI&ET are analyzed using VOSviewer software through a co-occurrence network, the full counting method, and an analysis unit, which identify keywords according to the literature specified in the Methodology Section. The minimum number of occurrences is five for a keyword. Of the 4230 words, 148 meet the threshold. Figure 8 shows that the most prevalent keywords are those identified with larger and more intense circles, consisting of neural networks, deep learning, artificial intelligence, and forecasting.

It should be noted that VOSviewer aims to visualize bibliometric networks. It also runs in a web browser and can be used to share interactive visualizations. These networks can be built based on citation relationships, bibliographic coupling, co-citation, or co-authorship. VOSviewer also presents a comprehensive view of scientific activities in specific areas, such as the one studied in this document on AI&ET, and can serve as an excellent tool to support strategic decision-making.

For visualization purposes, version 1.6.20 of VOSviewer, updated on October 31, 2023, was used. VOSviewer was selected for its superior capability in visualizing complex bibliometric networks and its proven effectiveness in identifying emerging research patterns in technological fields. This version offers improved functions for creating maps based on data downloaded through API and supports the creation of maps based on data exported from Scopus in the new Scopus file format.

The analysis was then carried out by authorship and co-authorship with a maximum of 25 authors in the articles. The number of authors and co-authors is 1182, as shown in Figure 9. The most relevant researchers are identified as Wang X, Wang J, Kim S, Li X, and Zhang Z. Figure 10 identifies that there are collaborations between the most prominent authors, and they form an extended network to publish their documents.

Table 2. Assessment by supply chain according to the most prominent AI domain identified in the featured publications as of December 2024.

Part of the Supply Chain	Use Cases	Most Prominent AI Domain	Full Article Title	Journal	Authors	Citations	AI Application Examples
Energy Generation (Wind)	Wind energy forecasting and optimization	Machine Learning (ML)	“Artificial intelligence and machine learning in grid connected wind turbine control systems: A comprehensive review” [112]	Energies	Nathan Oaks Farrar, Ali, M. H., Dasgupta, D	25	- ML-based prediction of wind energy generation.
Energy Generation (Solar)	Solar energy production optimization	Deep Learning (DL)	“A review of the applications of artificial intelligence in renewable energy systems: An approach-based study” [113]	Energy Conversion and Management	Mersad Shoaei, Younes Noorollahi, Ahmad Hajinezhad, Seyed Farhan Moosavian	50	- Solar power generation prediction using deep learning.
Energy Distribution (Smart Grids)	Efficient grid management and fault detection	Optimization Algorithms and Neural Networks	“A Survey on the Electrification of Transportation in a Smart Grid Environment” [114]	IEEE Transactions on Smart Grid	Wencong Su, Habiballah Eichi, Wente Zeng, Mo-Yuen Chow	639	- Automated fault detection in the grid.
Energy Storage (Batteries)	Battery life prediction and energy storage management	Predictive Analytics and Machine Learning	“Optimization of energy storage systems for integration of renewable energy sources—A bibliometric analysis” [115]	Journal of Energy Storage	Hira Tahir	14	- Optimization of battery storage efficiency.
Energy Grids (Power Networks)	Fault prediction and grid maintenance	Computer Vision and Predictive Analytics	“The Current State of the Art in Research on Predictive Maintenance in Smart Grid Distribution Network: Fault’s Types, Causes, and Prediction Methods—A Systematic Review” [116]	Energies	Moamin A. Mahmoud 1, Naziffa Raha Md Nasir, Mathuri Gurunathan, Preveena Raj, Salama A. Mostafa	200	- Predictive maintenance for power grids.
Energy Generation (Hydropower)	Hydropower optimization for efficiency	Process Optimization and Simulation	“A new hybrid Artificial Intelligence (AI) approach for hydro energy sites selection and integration” [117]	Heliyon	F. Chen Jong, Musse Mohamud Ahmed, W. Kin Lau, H. Aik Denis Lee	150	- Prediction of water flow and turbine efficiency.
Electric Vehicles (EVs)	Optimization of EV fleet energy consumption and routing	Optimization Algorithms and Computer Vision	“Artificial Intelligence Optimization for User Prediction and Efficient Energy Distribution in Electric Vehicle Smart Charging Systems” [118]	Energies	Siow Jat Shern, Md Tanjil Sarker, Mohammed Hussein Saleh Mohammed Haram	1	- Charging station demand prediction.
EV Charging Networks	Charging station load balancing	Supervised Learning and Predictive Analytics	“Performance analysis of AI-based energy management in electric vehicles: A case study on classic reinforcement learning” [119]	Energy Conversion and Management	Jincheng Hu, Yang Lin, Jihao Li	210	- Energy optimization for charging stations.
Energy Consumption (Industry)	Industrial energy optimization for manufacturing processes	Neural Networks and Optimization Algorithms	“Energetics Systems and artificial intelligence: Applications of industry 4.0” [120]	Energy Reports	Tanveer Ahmad, Hongyu Zhu, Dongdong Zhang	180	- Optimization of energy consumption in industrial operations.
Energy Consumption (Smart Buildings)	Energy management in smart buildings	Intelligent Control and Neural Networks	“Applications of artificial intelligence for energy efficiency throughout the building lifecycle: An overview” [121]	Energy and Buildings	Raheemat O. Yussuf, Omar S. Asfour	48	- Automated energy consumption adjustments in buildings.
Energy Consumption (Homes)	Household energy usage prediction and optimization	Predictive Analytics and Optimization Algorithms	“Deep Reinforcement Learning for Smart Home Energy Management” [122]	IEEE Internet of Things Journal	Liang Yu, Weiwei Xie, Di Xie, Yulong Zou	278	- Real-time energy consumption prediction and adjustment.
Energy Generation (Biomass)	Biomass energy production efficiency	Optimization and Predictive Modeling	“Production of biofuels from biomass: Predicting the energy employing artificial intelligence modelling” [123]	Bioresource Technology	Manish Meena, Shubham Shubham, Kunwar Paritosh	130	- Biomass energy generation prediction.
Energy Generation (Geothermal)	Geothermal energy production forecasting	Optimization Algorithms and Simulation	“The Geothermal Artificial Intelligence for geothermal exploration” [124]	Renewable Energy	J. Moraga, H.S. Duzgun, M. Cavur, H. Soydan	115	- Predicting geothermal energy output based on geological data.
Energy Storage (Hydrogen)	Hydrogen production optimization for energy storage	Predictive Modeling and Simulation	“Artificial intelligence driven hydrogen and battery technologies—A review” [125]	Fuel	A. Sai Ramesh, S. Vigneshwar, Sundaram Vickram	95	- Optimization of hydrogen storage systems.
Microgrids (Energy Networks)	Microgrid optimization for energy distribution	Neural Networks and Control Algorithms	“Role of optimization techniques in microgrid energy management systems—A review” [126]	Energy Strategy Reviews	Gokul Sidarth Thirunavukkarasu, Mehdi Seyedmahmoudian, Elmira Jamei	120	- Real-time control of microgrid energy distribution.
Energy Generation (Tidal and Wave)	Optimization of tidal and wave energy generation	Predictive Models and Computational Simulation	“Wave energy converter array layout optimization: A critical and comprehensive overview” [127]	Renewable Energy	Bo Yang, Shaocong Wu, Hao Zhang, Bingqiang Liu	110	- Wave behavior prediction for energy optimization.
Energy R&D (Innovation)	AI-assisted development of new renewable energy technologies	Computational Simulation and Machine Learning	“The role of utilizing artificial intelligence and renewable energy in reaching sustainable development goals” [128]	Renewable Energy	Fatma M. Talaat, A.E. Kabeel, Warda M. Shaban	135	- Development of new solar materials using AI.
Energy Consumption (Commercial)	Commercial building energy optimization	Machine Learning and Data Analytics	“Data-driven prediction and optimization toward net-zero and positive-energy buildings: A systematic review” [129]	Building and Environment	SeyedehNiloufar Mousavi, María Villarreal-Marroquín, Mostafa Hajiaghaei-Keshteli	66	- Predictive energy optimization for commercial facilities.
Energy Storage (Supercapacitors)	Supercapacitor optimization for fast energy storage	Neural Networks and Optimization Algorithms	“Recent advances in artificial intelligence boosting materials design for electrochemical energy storage” [130]	Chemical Engineering Journal	Xinxin Liu, Kexin Fan, Xinmeng Huang, Jiankai Ge	24	- Predictive analytics for energy storage efficiency.
EV Charging Infrastructure	Optimizing charging network design and maintenance	Optimization Algorithms and Predictive Analytics	“A novel AI approach for optimal deployment of EV fast charging station and reliability analysis with solar based DGs in distribution network” [131]	Energy Reports	Fareed Ahmad, Imtiaz Ashraf, Atif Iqbal	64	- Optimizing the placement of charging stations.
Aviation (Electric Aircraft)	Optimization of electric aircraft design and operations	Computational Simulation and Machine Learning	“Design and Optimization of Control System for More Electric Aircraft Power Systems Using Adaptive Tabu Search Algorithm Based on State-Variables-Averaging Model” [30]	IEEE Access	Ratapon Phosung, Kongpan Areerak, Kongpol Areerak	125	- AI-assisted design for energy-efficient electric aircraft.
Energy Grids (Load Balancing)	Load balancing for energy grids	Control Algorithms and Optimization	“Leveraging the power of machine learning and data balancing techniques to evaluate stability in smart grids” [132]	Engineering Applications of Artificial Intelligence	Zaid Allal, Hassan N. Noura, Ola Salman, Khaled Chahine	150	- AI-driven load balancing for power grids.
Energy Logistics	Optimization of energy transportation networks	Route Optimization and Machine Learning	“AI-based energy-efficient path planning of multiple logistics UAVs in intelligent transportation systems” [133]	Computer Communications	Pengfei Du, Xiang He, Haotong Cao, Sahil Garg	33	- Optimizing energy delivery routes for transport.
Solar Energy R&D	Research in new solar cell technologies	Evolutionary Algorithms and Neural Networks	“Public willingness to pay for the research and development of solar energy in Beijing, China” [134]	Energy Policy	Jianjun Jin, Xinyu Wan, Yongsheng Lin	110	- AI optimization for solar cell material properties.
Energy Generation (Biogas)	Biogas plant optimization	Control Algorithms and Optimization	“Integrated deep learning neural network and desirability analysis in biogas plants: A powerful tool to optimize biogas purification” [135]	Energy	Mahmood Mahmoodi-Eshkaftaki, Rahim Ebrahimi	90	- Biogas production forecasting.
Energy Networks (Infrastructure)	Planning and development of energy infrastructure	AI-Based Planning and Simulation	“AI in public-private partnership for IT infrastructure development” [136]	The Journal of High Technology Management Research	K. Rajendra Prasad, Santoshachandra Rao Karanam, D. Ganesh	32	- AI-assisted energy infrastructure planning.

presents the evaluation results by supply chain according to the most prominent AI domain area and the featured publications from January 1990 to December 2024. The use of AI in research is experiencing exponential growth, and its applications in the energy sector follow similar trends, which highlights the growing interest in studying the two integrated fields and, most likely, that shortly they will bring about unexpected outcomes since the AI growth ceiling is still far away.

• Among the most prominent AI domains for the energy sector are the domains of “ML”, “DL”, “Predictive Analytics”, “Computer Vision”, “Neural Networks”, and “Optimization.”
• Although there is a long history of using optimization techniques in the energy sector, the current taxonomy of AI is very restricted to rather metaheuristic algorithms. After analyzing high-impact works, articles of global interest were published in world-renowned journals and had a high impact, according to JCR. Additionally, the number of citations for each article is quite significant despite several recent publications. The development of the publications of the main authors in the area of interest and those who develop collaborative work with other prominent authors is highlighted. New, relatively young authors are emerging in these growth areas; they may be the creators of future developments in AI&ET.

The extensive analysis presented above shows that AI techniques directly impact the energy sector, mainly large-scale renewable energy production systems. The search for a more reliable, intelligent, and secure system is being pursued by improving digital systems. The analysis of generation systems has become essential since the conventional electrical system has undergone significant changes. Nowadays, a consumer can also be a generator, and their systems can operate better thanks to communications, especially with the support of AI. This situation radically changes the analysis of energy systems in energy transition processes.

The analysis of the reviewed literature, particularly the publications presented in Table 2, reveals several significant gaps that require further research:

-

Large-scale integration: While most studies focus on specific AI applications, there is a notable lack of research addressing the holistic integration of multiple AI technologies into entire energy systems at the national or regional scale. Current implementations tend to be siloed and lack integration frameworks. Studies currently available for large-scale energy planning, such as EnergyPLAN, Message, Elena, and others, still lack platforms or features that include artificial intelligence tools, which would make energy planning models much more complex and present opportunities for research and development.

-

Economic aspects: Existing studies predominantly focus on technical feasibility, with a scarcity of research comprehensively evaluating the economic and financial aspects of implementing AI-based solutions, particularly in emerging economies. Cost–benefit analyses and return on investment studies for large-scale AI implementations are notably lacking.

-

Standardization: A significant gap exists in the literature regarding the standardization of protocols and methodologies for AI implementation in energy systems. This lack of standardization hinders the replicability and scalability of proposed solutions across different contexts and regions.

-

System resilience: Current publications inadequately address how AI-based systems can maintain their effectiveness in the face of disruptive events or significant changes in operating conditions. The robustness of AI solutions under various stress scenarios requires further investigation.

-

Developing countries: There is a marked absence of studies examining AI solution implementation in developing country contexts, where energy infrastructures may be less advanced. This gap is particularly relevant given that these countries often face unique challenges in energy transition.

-

Social aspects: The literature lacks deep research on the social impact of AI implementation in energy systems, including aspects such as public acceptance, job market implications, and equity in access. The human dimension of AI-driven energy transitions remains understudied.

-

Human Development: No research has yet been identified in prestigious journals that includes human development indicators resulting from these new technological developments.

These identified gaps represent significant opportunities for future research in the field of AI applied to energy transition. Addressing these gaps will be crucial for achieving a more comprehensive understanding of how AI can effectively support global energy transition efforts while ensuring equitable and sustainable implementation.

Pal Boza and Theodoros Evgeniou [137] gave sufficient guidelines on how artificial intelligence can support the integration of variable RESs into the electrical system. The literature review also identifies that authors mainly apply non-linear models due to the large data volume and complexity [138]. The application of AI in the energy industry is gradually shifting toward a data analytics-based approach as technologies reach significant maturity levels and processing capabilities improve [139]. According to Figure 3 presented in this study, ML algorithms began to be used more frequently in the third period to forecast electricity demand, generation, and energy market price. The future is promising for carrying out energy transition processes in a planned and orderly manner, avoiding taking unforeseen actions on the fly that typically determine higher economic resource requirements [140]. Reducing maintenance or fuel costs or extending the useful life of assets can be measured in laboratory environments and verify that the new order based on clean energy would benefit the energy sector.

Serban and Lytras [141] summarized the applications of AI in large-scale integration of RESs concerning the generator side, the grid side, and the consumer side. This is considered to be possible within a significant time range. Researchers commonly consider a time frame of up to 2050, so we built an adapted model in which countries with reduced development capacity can be inserted, as shown in Figure 11.

5. Discussion

According to the literature review in this study aimed at identifying the main impacts of joint AI&ET developments that focus on global digital decarbonization, a large number of RES systems will be connected to the grid shortly given the great growth in AI as evidenced in the last five years (third stage). Much of the literature indicates that small distributed renewable energy sources can generate and sell electricity to the grid [142,143]. Among the most renowned studies are those promoted by Marcos Tostado, who provides sufficient guidelines on how smart homes will influence the future and has also created a methodology to evaluate grid-integrated energy systems in the community context [26,144,145]. Electric vehicles and associated techniques (such as fast-charging batteries) will show increasing market demand, as stated by David Borge-Diez et al. [146] in their study regarding V2B. Antonio Barragan et al. [147] have also pointed out the main barriers that currently hinder the deployment of renewable energies in countries that seek to transform their energy matrix to a more environmentally friendly one. It is also recognized that smart home devices can be connected to the grid even without the knowledge of the grid operator [148]. All this would significantly impact the energy stability of the local electrical grid, which, through AI techniques, can be controlled and energy efficiency criteria applied [149,150]. Optimizing the operation of the network with the help of AI techniques will further improve the transmission and distribution capacity of existing lines, and, to a large extent, repowering these network segments will be avoided. The useful life of the equipment can be extended thanks to the transformation of the energy sector that includes higher rates of participation in renewable energy supported by state-of-the-art digital systems [151,152]. However, the application of AI techniques in large-scale renewable energy integration still encounters many barriers and limitations that must be intensively addressed and discussed, so an interesting deployment of research is expected to be indexed in the databases and its methodologies will contribute to the development of AI&ET [79,153]. Some of these challenges include the following:

• Integrating renewable energy: The variability in sources such as solar and wind can make integrating them into the grid difficult. AI can help predict energy production and manage demand, optimizing the use of renewable energy.
• Energy efficiency: Improving energy consumption efficiency is critical. AI can analyze consumption patterns and suggest strategies to reduce energy use in buildings and industrial processes.
• Smart grid management: Implementing smart grids requires complex management. AI can facilitate real-time monitoring and control, ensuring efficient and reliable energy distribution.
• Energy storage: Storage systems are crucial to balancing supply and demand. AI can optimize the use of batteries and other storage systems, improving their performance and extending their lifespan.
• Changing consumer behavior: The energy transition involves changes in how consumers use energy. AI can personalize recommendations and incentives to encourage more responsible consumption.
• Innovation in clean technologies: AI can accelerate the research and development of new energy technologies, helping to discover more efficient and sustainable solutions.
• Inequality in energy access: The energy transition must be inclusive. AI can analyze social and economic data to identify areas where energy access needs to be improved.
• Regulation and policies: Designing effective policies for the energy transition is complex. AI can provide predictive analysis that informs policy decisions and regulations.

Collaboration between energy and AI experts is essential to effectively address these challenges and promote a more sustainable and efficient energy future.

As expected, algorithms in AI related to planning are popular in power system planning, especially those designed to achieve the location and sizing of power generation units, transmission, and distribution networks [79]. However, according to the literature review, the use of AI in planning exercises is detected to be much less frequent than in O&M, which may be related to the nature of the task and the AI algorithms currently used [154]. For example, in planning generation systems using renewable energy, efforts are directed toward expansion plans of energy systems using the concept of smart energy [45]. This type of planning structures energy systems as blocks but does not focus solely on parts of the network as the study of smart grids does. Of course, the more distant the reference horizon is, the less visible the AI precision is in short- and medium-term planning. However, long-term studies are conceived as roadmaps that define the path to be followed to achieve, for example, decarbonization processes in a defined territory with renewable energy potentials that possess exploitation possibilities.

Table 3. Comparative analysis of the present study with the latest research related to AI&ET.

Title of the Research	Main Focus	Technologies/Methods Used	Impact on Energy Transition	Challenges	Authors
“Optimization of energy acquisition system in smart grid based on artificial intelligence and digital twin technology” (2024)	Optimizing smart grids using AI for efficient energy flow management.	Optimization algorithms, neural networks, and machine learning.	Improves grid efficiency and stability, facilitating renewable energy integration.	Need for advanced infrastructure.	Zhen Jing et al. [155]
“Artificial intelligence powered large-scale renewable integrations in multi-energy systems for carbon neutrality transition: Challenges and future perspectives”(2022)	Forecasting renewable energy generation (solar, wind) using AI for grid integration.	Machine learning, prediction algorithms, neural networks, and Big Data analysis.	Enhances renewable energy predictability, optimizing grid integration.	Variability in generation and storage.	Zhengxuan Liu et al. [156]
“Predicting energy consumption in multiple buildings using machine learning for improving energy efficiency and sustainability” (2020)	Optimizing energy consumption in buildings and cities using AI.	Optimization algorithms, deep neural networks, and predictive analytics.	Reduces demand peaks and increases energy efficiency.	Scalability in different environments.	Anh-Duc Pham et al. [157]
“AI-driven predictive maintenance for energy infrastructure” (2024)	Using AI for predictive maintenance in energy infrastructure.	Machine learning algorithms, Big Data analysis, and cyber–physical systems.	Improves infrastructure reliability and reduces downtime.	Complexity in integrating heterogeneous data.	Ibrahim Adeiza Ahmed and Paul Boadu Asamoah [158]
“Smart energy systems: A critical review on design and operation optimization” (2020)	Implementing AI in energy management for smart grids.	AI algorithms, optimization, machine learning, and neural networks.	Optimizes efficiency and stability in smart grids.	Requires infrastructure upgrades.	Yizhe Xu et al. [159]
“Application of artificial intelligence for prediction, optimization, and control of thermal energy storage systems” (2023)	Optimizing renewable energy storage systems with AI.	Optimization algorithms, neural networks, and predictive control.	Enhances storage capacity and renewable energy integration.	Cost and efficiency of storage systems.	A.G. Olabi et al. [31]
“Does artificial intelligence promote energy transition and curb carbon emissions? The role of trade openness” (2024)	A comprehensive review of AI applications in energy transition.	Review of technologies, machine learning approaches, Big Data, and predictive analytics.	Provides a holistic view of AI in the energy transition.	Challenges in integrating diverse technologies.	Qiang Wang et al. [22]
“Artificial intelligence techniques for stability analysis and control in smart grids: Methodologies, applications, challenges and future directions” (2020)	AI applications for enhancing the stability and reliability of power grids.	Predictive algorithms, machine learning, and Big Data analysis.	Increases the stability and resilience of power grids.	Integration of real-time control systems.	Zhongtuo Shi et al. [160]
Current research	Evaluation of AI in the energy transition in recent years.	Using interconnection maps of the main development centers.	Identify the main AI developers and the application paths in energy transition processes.	Impact of AI on long-term energy planning.	Icaza Alvarez, D., Gonzalez L., Fernando, Rojas E., J., Borge Diez D., Pulla G. S., Flores C.

below compares the differentiating aspects between the most outstanding works in AI&ET and the present study. The key differences in the works in the analysis field in high-impact journals and general trends can be seen. The main approaches are related to renewable energy optimization; algorithms, neural networks, and machine learning have typically been used. These developments are leading to easier integration of renewable energy, improving energy efficiency and the stability of energy systems. The challenge lies in making these applications much more economically accessible.

Research on the application of AI in the energy transition has advanced rapidly in recent years. It focuses on smart grids, renewable energy forecasting, energy storage optimization, and predictive maintenance. These studies use advanced technologies such as machine learning, Big Data, and neural networks to improve energy system efficiency, renewable energy integration, and sustainability.

Key challenges include variability in renewable energy generation, data integration complexity, and solutions’ scalability. Despite this, studies are paving the way for more efficient and resilient energy systems crucial to achieving global sustainability and decarbonization goals.

It is suggested that policies be developed on the incorporation of AI in the energy industry by creating a regulatory framework that guarantees the ethical use of AI, ensuring that biases in algorithms are avoided and that models are transparent and explainable. It would be appropriate to create regulations to require companies to conduct ethical and social impact assessments before implementing AI technologies in their operations. This would include assessing how they might affect the workforce, communities, and the environment.

6. Conclusions

This review article presents the development of research addressing the issue of climate change with a much greater focus on AI&ET. However, the literature is being developed at an accelerated pace that will become much more promising in the coming years and will surely have surprising impacts on the energy industry. There is still a lack of important developments, such as creating roadmaps for countries using artificial intelligence. However, there are already signs that efforts are being made to create tools directly related to these areas of long-term market planning. All the studies found that 100% RESs is technically the most desirable and best applicable. The following can be concluded:

The synergy between AI and ET represents a crucial advancement in the quest for a more sustainable and efficient energy system. First, AI can optimize the management of energy resources by analyzing large volumes of data in real time. This makes it possible to adjust energy production, distribution, and consumption, facilitating the integration of renewable sources such as solar and wind, which depend on variable weather conditions. By predicting demand and balancing supply, AI helps minimize energy waste and reduces operating costs, which is especially important in the context of increasing pressure for sustainability. In addition, the development of smart grids is one of the areas where the synergy between AI and ET shows its potential. Thanks to AI’s analytical capacity, it is possible to implement monitoring and control systems that optimize the operation of the energy infrastructure. These smart grids allow for dynamic energy management, adapting to fluctuations in real time and guaranteeing a more reliable supply. This not only improves the energy system’s resilience but also supports the penetration of renewable resources, helping to meet greenhouse gas emission reduction targets.

Another crucial aspect is how AI can influence consumer behavior and promote responsible energy use. By personalizing recommendations and analyzing consumption patterns, AI can educate and incentivize users to adopt more sustainable habits. For example, energy management apps and platforms can offer real-time information on consumption and suggestions for reducing it. This approach not only fosters environmental awareness among consumers but also contributes to the stability of the energy system by reducing demand during periods of high load.

Among the research centers that are currently creating the most development worldwide are the Chinese Academy of Sciences, the Ministry of Education of the People, Fudan University, Southeast University, and Tsinghua University, with most of these centers located in Asia. For this reason, it is expected that in the coming years there will be a significant deployment of products and services from these centers that will revolutionize the different productive sectors in the world. Future scientific research will be directed toward optimization for the integration of renewable sources such as solar and wind through smart power grids and predictive supply and demand systems. In addition, AI will improve energy efficiency in buildings and industries through automation and real-time analysis. It will contribute to the development of more efficient storage technologies and smart management of charging infrastructure for electric vehicles. It will also facilitate decentralized and more flexible energy markets and enable accurate monitoring of emissions and the life cycle of energy resources. In this way, AI will become essential to accelerating the transition toward a more sustainable and resilient energy system.

Finally, AI is also instrumental in developing more informed and equitable energy policies. Using predictive analytics and data intelligence techniques, policymakers can identify trends, challenges, and opportunities in the energy sector. This includes identifying areas with limited access to energy, which favors inclusive initiatives in the energy transition. Combining AI with a social justice-focused approach can ensure that the benefits of the energy transition are equitably distributed, contributing to a sustainable future for all communities. In short, the synergy between AI and the energy transition will foster sustainability and transform how we interact with energy. The limitations of this study include the fact that it has not been possible to identify the extent to which AI is impacting the energy systems of developing countries. Future work that follows from this study lies in creating decarbonization roadmaps for countries, regions, and islands by applying current AI-based tools. Future research should address critical gaps, including large-scale AI system integration, economic viability in emerging economies, standardization protocols, system resilience, implementation in developing countries, social equity considerations, and impacts on human development indicators. These directions would enhance our understanding of how AI can effectively support sustainable energy transitions. This could be of great interest for future research in the field.