Perspective Towards a just AI-assisted energy transitions for vulnerable communities Laurence L. Delina * , Yuet Sang Marie Tung Division of Environment and Sustainability, The Hong Kong University of Science and Technology, Clear Water Bay, Kowloon, Hong Kong A R T I C L E INFO Keywords: Energy transition Artificial intelligence Vulnerable communities Energy justice ABSTRACT This Perspective overviews the role and potential impact of artificial intelligence (AI) in accelerating the tran- sition from fossil fuels to renewable energy technologies and systems. We pay close attention to and speculate the probable impacts of AI on communities which are already vulnerable to the effects of energy transition and discuss how these impacts can be mitigated. In addition to a short literature review, we employed the assistive capabilities of a Generative AI chatbot through hypothetical roleplays to provide feedback on proposed miti- gative measures. We highlight society-positive and society-negative impacts, emphasising the implications for just energy transition. We suggest potential measures to address these issues using the energy justice framework, including developing accurate training datasets, community-based mitigation policies, and establishing partic- ipative decision-making channels. We argue that human discretion must remain paramount, particularly in ensuring participative policymaking that safeguards social equity protection for vulnerable communities in AI- assisted energy transitions. 1. Introduction Fossil fuel use is responsible for over three-quarters of global greenhouse gas emissions and nearly 90 % of CO 2 emissions globally, contributing significantly to the climate crisis [1]. Therefore, tran- sitioning from fossil fuel-based energy to renewable sources such as solar, wind, and hydropower is crucial in addressing climate change. Additionally, the rapid advancements in artificial intelligence (AI) technologies have sparked discussions on their potential role in energy transitions (e.g., [2–4]). Furthermore, there is a growing emphasis on just energy transitions, meaning these sociotechnical shifts must consider distributive, proce- dural, and recognition justice implications [5,6]. Energy justice en- compasses these three elements and aims to provide safe, affordable, and sustainable energy to all individuals, regardless of location [7]. Recognition justice entails acknowledging different stakeholders, particularly vulnerable groups, and considering their unique needs and circumstances [8]. Distributive justice involves sharing benefits and burdens among various stakeholders [9]. Procedural justice relates to the ‘equitable and democratic involvement of all stakeholders in energy decision-making,’ relying primarily on complete information disclosure and public participation [10]. Despite advances in energy transition and AI technologies, little attention has been given to studying the potential impacts of AI-assisted energy transitions on vulnerable communities and ensuring alignment with the elements of energy justice. The differentiation between the various relationships is crucial for a nuanced understanding of the dy- namics of energy transitions and their impacts on vulnerable communities. The relationship between AI and energy transitions centres on opti- mising energy production, distribution, and consumption, thereby fostering more efficient and sustainable energy systems. As the inte- gration of renewable energy sources, such as solar and wind, increases, AI plays a crucial role in addressing the inherent intermittency of these sources. AI can effectively balance supply and demand by enhancing predictions of weather conditions and consumption patterns [2], leading to greater energy stability. Additionally, AI facilitates site selection for renewable energy infrastructure and optimises grid performance through smart grids, which require two-way communication between utility providers and end-users [11–13]. Moreover, AI supports demand response schemes by analysing con- sumption and production patterns, allowing users to adjust their energy consumption based on real-time data [11,12]. This capability not only promotes energy efficiency but also contributes to the overall stability of the energy system. In summary, integrating AI into energy transitions offers significant opportunities for optimising energy systems and enhancing the utilisation of renewable energy sources while also necessitating careful consideration of the implications for all * Corresponding author. E-mail address: lld@ust.hk (L.L. Delina). Contents lists available at ScienceDirect Energy Research & Social Science journal homepage: www.elsevier.com/locate/erss https://doi.org/10.1016/j.erss.2024.103752 Received 6 March 2024; Received in revised form 3 September 2024; Accepted 6 September 2024 Energy Research & Social Science 118 (2024) 103752 2214-6296/© 2024 Elsevier Ltd. All rights are reserved, including those for text and data mining, AI training, and similar technologies.