Decarbonization policies in sustainable development constitute one of Japan's essential policy tools. Japan faces environmental unsustainability, as its ecological footprint exceeds its current biocapacity. In this context, it is clear that Japan needs sustainable policies in today's conditions. It is clear that artificial intelligence technologies, which are becoming increasingly widespread today, are related to environmental factors and their social benefits. Low-carbon energy sources can be a policy tool in transitioning from grey energy to green. In this context, this paper examines the impact of artificial intelligence and low-carbon energies on environmental sustainability in Japan, using the Fourier Augmented ARDL model from 1985 to 2022. The study considers possible macro changes by integrating gradual structural changes into the cointegration model. Based on empirical findings, artificial intelligence plays a significant role in promoting environmental sustainability. Low-carbon energy use increases environmental quality. Ultimately, trade openness contributes to environmental development. The Japanese government should support the development of artificial intelligence technologies for a sustainable environment and introduce legal regulations to mitigate their environmental impacts. Energy security should be ensured by boosting the share of renewables in low-carbon energy. Ultimately, this study provides policy implications that can accelerate the achievement of the Sustainable Development Goals.
Globally, artificial intelligence (AI) is currently revolutionizing different spheres of life (Wang et al., 2024). In particular, there are high expectations that AI could revolutionize environmental sustainability (ENS), thereby strengthening climate change resiliency in most economies (Wang, 2023a). AI is a critical tool in environmental disaster modeling, helping to mitigate natural disasters such as severe storms, hurricanes, and wildfires (OECD, 2022). It has also been explored that AI is shaping ENS in several contexts and dimensions, including forecasting and detection of potential climate distortions (Ulug et al., 2025). Likewise, AI notably reduces greenhouse emissions, energy costs, and variability (Bibri et al., 2024). Notably, the transformative capabilities of AI and its potential to foster ENS have generated prolonged debates among scholars. Notably, scholars believe that AI's revolution presents both opportunities and challenges to environmental progress. Some scholars (Shoha et al., 2024; Tao, 2024; Wang et al., 2024; Zhao et al., 2024) argue that AI consistently enhances ENS. Conversely, others (Qian et al., 2023; Shah et al., 2024) argue that AI exacerbates environmental challenges, as it necessitates the excessive utilization of energy and resources. Additionally, prior literature (Ding et al., 2023; Dong et al., 2024a; Liu et al., 2024) underscores overwhelming ambiguities in the AI-ENS relationship.
Likewise, energy scholars have emphasized the cardinal roles of low-carbon energy consumption (LCE) in ENS (Caglar et al., 2025; Erdas et al., 2025). Accordingly, LCE sources include nuclear and renewable primary energies (Avci and Caglar, 2025; Rasheed et al., 2025). Hence, the choice of low-carbon energies stems from the documented adverse effects of traditional energies in carbon emissions (CE) and climate change adversities (Uche et al., 2024). However, it is worth noting that only a handful of studies have investigated the ENS-LCE nexus. Particularly, amidst countervailing submissions, prior studies only verified the implications of variants of LCE on ENS (Bibri et al., 2024; Prakash, 2025; Yasir et al., 2024). This leaves a significant gap in the empirical literature regarding ENS determinants. It is essential to emphasize that a comprehensive evaluation of the implications of LCE on environmental progress has significant policy implications. This argument hinges on its ability to mitigate climate adversities, as it balances global energy needs with environmental concerns (Yasir et al., 2024).
Building on these established insights, this study's objectives are to enhance the empirical understanding of the implications of AI and LCE on ENS in Japan. Notably, Japan is among the developed countries that are intensifying efforts toward carbon neutrality and climate change resilience through AI (Qing et al., 2024). Japan occupies the third position among the largest economies by nominal GDP and the fourth position among the largest economies. Japan is committed to building a society that is both sustainable and resilient. It is one of the world's leading electronics and the third-largest automobile producers (WDI, 2025). Moreover, the Japanese government established a national strategy (AI Strategy) in 2019, and the number of projects supported by this strategy doubled between 2019 and 2020. AI remains paramount in actualizing Japan's 2050 zero CE target. However, considering Japan's unimpressive load capacity factor (LCF) score (Fig. 1 - green shaded areas), there is an urgent need to evaluate the contributions of AI and LCE to ENS. The green-shaded areas demonstrate a consistent decline in Japan's LCF score after 1961. However, there is evidence of gradual improvements since 2019. Could these gradual improvements be attributable to AI evolution and/or the shift toward LCE? These critical questions are begging for answers, having been ignored in prior studies. Additionally, the study examines the Load Capacity Curve (LCC) hypothesis for Japan, given the lack of such an exposition in existing studies. The LCC hypothesis posits that the short-term negative effects of economic growth (as measured by GDP) can be reversed in the long term when GDP reaches a particular threshold (Daştan et al., 2025). Hence, a U-shaped relationship is expected between GDP and LCF.
It is essential to note that the current investigation enhances the depth of knowledge in several dimensions. First, there are a handful of studies on ENS-AI interactions, yet their submissions are largely contextual. Additionally, existing studies have primarily relied on CE to measure environmental quality (EQ). To enhance the depth of knowledge, this study relied on the ENS, as it accounts for both the supply and demand sides of environmental performance (Demirdag et al., 2025). Hence, this study is the first to assess the implications of AI on ENS in Japan. Second, the option of the LCE variable mitigates the drawbacks of prior studies, given their preference for variants of LCE. Notably, selecting comprehensive metrics against the option in prior studies provides more insights for policy adjustments. Third, it is noteworthy that this paper probes the combined effects of AI and LCE on ENS in Japan. Admittedly, available information suggests that such a step has not been considered in extant studies. This is another critical step with profound policy implications. Accordingly, understanding the combined impact of AI and LCE on ENS is crucial for guiding policymakers on how to harness their potential for enhanced energy security through AI's innovative processes. Moreover, the need for proactive measures to address climate change cannot be overstated. Hence, understanding how AI can contribute to this objective is of policy relevance. Lastly, by employing cutting-edge empirical techniques, including the Augmented Autoregressive Distributed Lag (AARDL) approach modified with the Fourier terms, the study yields new policy insights for policymakers. It opens new research pathways for energy scholars. Additionally, the study conducted a robustness evaluation (Fourier Dynamic Ordinary Least Square (DOLS)) on the benchmark model's outcomes. This extra step, which separates the study from others, enhances its contents and facilitates its practical implications.
The study's other components include a review of related literature in Section "Theoretical background and literature review", research methodology, data analysis, and discussions in Sections "Data, model, and methodology" and "Empirical findings and discussion", and a conclusion in Section "Conclusions and policy recommendations" that outlines critical policy options for improved environments.