The Comparative AI Index Figs. 5 (panels a–d) illustrate the AI development trends of 52 countries from 2010 to 2024, ranking them from lowest to highest based on their average AI index. Figures 5 offer a comprehensive comparison, highlighting AI growth and adoption disparities across different economies. The four panels categorize countries into quartiles, with panel (a) displaying nations with the lowest AI index scores, while panel (d) represents the top-performing countries. By visualizing AI index trends over time, these figures provide insights into the global AI landscape, technological advancements, and policy effectiveness across diverse regions. In panel (a), countries with low AI development, such as Peru, Pakistan, and Uzbekistan, exhibit relatively stagnant or slow-growing AI index values. These nations face technological and infrastructural barriers, including limited investment in AI research, a lack of a skilled AI workforce, and weak policy support. The gradual increase in the AI index for some countries suggests growing efforts in digital transformation and ICT infrastructure development, though they still lag behind global leaders. Panel (b) includes emerging AI adopters, such as Brazil, South Africa, and Turkey, who show moderate AI growth, benefiting from policy reforms and gradual integration of AI in industries.

Panel (c) highlights countries with significant AI progress, including China, France, and Canada, which have made substantial investments in AI research, innovation, and digital infrastructure. These nations have strengthened their AI ecosystems through government initiatives, university-led AI research, and private-sector collaborations. Meanwhile, panel (d) showcases AI powerhouses—the United States, Japan, and South Korea—which consistently rank at the top due to strong AI R&D, advanced computational power, and leading AI-based industries. These countries dominate AI innovation, automation, and machine learning applications, reinforcing their competitive edge in the global AI landscape. The comparative analysis across the four panels underscores the worldwide uneven distribution of AI capabilities. While leading nations continue to push AI frontiers, developing countries are making gradual advancements driven by digital transformation policies. The widening AI gap suggests that strategic investments in AI infrastructure, education, and R&D are crucial for enhancing AI capabilities in emerging economies⁴³. This figure provides valuable insights for policymakers, researchers, and industry leaders to identify AI growth patterns, investment needs, and future AI development strategies at both regional and global levels.

The results in Table 3 provide insights into the impact of AI on SD using the FMOLS model. Column (1) shows that AI has a positive and significant effect (1.2632, p < 1%) on the SD Index, confirming that AI adoption plays a crucial role in fostering sustainability. Among the AI dimensions, AI Research and Development Innovation (AI_RDI) has the most substantial impact (0.7880, p < 5%), followed by AI Infrastructure (AI_INF) at 0.5675 and AI Market (AI_MKT) at 0.1595. This suggests that countries investing in AI-driven R&D and infrastructure experience greater progress in sustainable development. Furthermore, Foreign Direct Investment (FDI), Industrial Development (IND), and Trade Openness (TRD) all contribute positively to sustainable development, indicating that global investment, industrial expansion, and trade activities enhance AI’s role in achieving sustainability goals.

Control variables such as GDP, Education, and Carbon Emissions also show significant relationships with sustainable development. Higher GDP (0.5183) and Education (0.2745) correlate positively with sustainable growth, reinforcing the importance of economic prosperity and human capital development. However, Carbon Emissions (-0.3278) negatively affect sustainable development, emphasizing the environmental challenges associated with industrialization. The R² values range from 0.68 to 0.82, indicating strong explanatory power for the models. The SD Index model’s highest R² (0.82) suggests that AI and its associated factors significantly contribute to explaining variations in sustainable development. These findings highlight the necessity of policy interventions to enhance AI infrastructure, innovation, and global collaboration, ensuring that AI catalyzes sustainability while mitigating adverse environmental impacts⁴⁴.

The results presented in Table 4 using the Two-Stage Least Squares (2SLS) model provide valuable insights into the impact of AI on SD. The first key finding is that AI itself has a significant positive effect on the SD Index, with a coefficient of 1.1784 (p < 1%) in column (1). This confirms that AI adoption contributes meaningfully to sustainable development, underlining its role in advancing environmental, economic, and social sustainability.

Countries investing in AI technologies experience better overall sustainability outcomes, suggesting that AI’s broad applications across various sectors are crucial in shaping a more sustainable future. Further disaggregation of the AI components reveals interesting nuances. AI_RDI, shown in column (2), has a significant positive effect of 0.7245 (p < 1%), indicating that R&D activities in AI are a key driver of sustainable development. Investment in AI innovation leads to developing more efficient technologies that optimize resource use and improve energy efficiency, which is fundamental to addressing environmental challenges. In contrast, AI_INF, with a coefficient of 0.5213, also supports AI’s widespread adoption, as robust ICT infrastructure enables AI applications in sectors such as agriculture, health, and environmental monitoring, contributing significantly to sustainability goals. On the other hand, AI_MKT shows a more negligible effect of 0.1389 (p < 10%) in column (4), suggesting that while the market for AI technologies is essential for economic growth and innovation, its direct impact on sustainable development is less pronounced than R&D and infrastructure. Nevertheless, this positive relationship indicates that as AI products and services become more globally accessible, they drive economic benefits and foster technological transfers that contribute to sustainability. Moreover, the influence of FDI, IND, and TRD across all models reinforces the importance of external capital, industrialization, and international trade in supporting AI’s role in achieving sustainable development. These factors facilitate the growth and application of AI technologies, further highlighting the interdependence of technological progress and economic development. Including control variables such as GDP, Education, and Carbon Emissions provides additional context to the findings. GDP shows a strong positive relationship with sustainable development, reflecting that wealthier countries are better equipped to invest in and benefit from AI technologies. Similarly, education enhances a nation’s ability to deploy AI effectively for sustainable growth. However, the negative coefficient for Carbon Emissions highlights the environmental challenges of industrialization, underscoring the need for AI to mitigate such adverse impacts. The R² values ranging from 0.65 to 0.79 indicate that the models effectively explain a significant portion of the variation in sustainable development, with the F-statistics confirming the models’ overall statistical significance. This comprehensive analysis suggests that AI, particularly through its R&D and infrastructure dimensions, is vital in advancing sustainable development goals⁴⁵.

Table 5 presents the results of the GMM model, examining the impact of AI on SD while addressing potential endogeneity through instrumental variables. Table 5 also provides four regression models that assess the relationship between AI (overall index), AI_RDI, AI_INF, and AI_MKT with sustainable development, using FDI, IND, and TRD as additional explanatory variables. The results show significant positive relationships between AI and sustainable development, with varying levels of impact across different AI dimensions. In column (1), the overall AI index has a significant positive coefficient of 1.1032 (p < 1%), suggesting that AI contributes significantly to sustainable development. This highlights AI’s broad and critical role in enhancing environmental, economic, and social outcomes. The coefficient value indicates that countries investing in AI technologies will likely see improvements in their sustainable development scores. The model uses AI Patents and AI R&D Investments as instrumental variables to address potential endogeneity, ensuring that the estimates are reliable and not biased due to reverse causality or omitted variable bias. This model’s R² value of 0.81 suggests that the predictors explain a substantial portion of the variation in sustainable development outcomes. Column (2) focuses on the impact of AI RDI (Research and Development Innovation) on sustainable development, with a significant positive coefficient of 0.6725 (p < 1%). This finding underscores that investments in AI research and innovation are crucial in driving sustainable development. AI RDI, which includes advancements in AI models and algorithms, is a key factor in optimizing resources, improving energy efficiency, and addressing environmental challenges. The FDI variable also shows a positive relationship with sustainable development (0.3627, p < 10%), indicating that foreign investments in AI can drive sustainability by facilitating technology transfer and supporting infrastructure development in developing countries. In column (3), the variable AI_INF (AI Infrastructure) exhibits a positive and statistically significant coefficient of 0.4672 (p < 5%), suggesting that robust AI infrastructure is essential for enabling sustainable development. The results highlight the importance of ICT infrastructure in supporting AI applications across various sectors, such as environmental monitoring, agriculture, and healthcare, which are critical for achieving sustainability.

In this model, Trade Openness (TRD) shows a positive relationship (0.3281, p < 10%) with sustainable development, further emphasizing the role of international trade in facilitating AI adoption and knowledge exchange. The instrumental variable used in this model is AI Infrastructure Funds, which helps mitigate endogeneity concerns related to infrastructure investments. In column (4), the AI_MKT (AI Market Influence) variable shows a positive but relatively more minor effect on sustainable development (0.1213, not statistically significant at conventional levels). This suggests that the global AI market drives economic growth and sustainable development, including exporting and commercializing AI technologies. However, its impact is less pronounced than AI RDI and AI INF. The control variables, including GDP, Education, and Carbon Emissions, also provide important context for understanding the broader drivers of sustainable development²¹. Both GDP and Education show significant positive coefficients, indicating that wealthier and better-educated countries are more likely to invest in AI technologies, contributing to sustainability. However, Carbon Emissions show a negative relationship, reflecting the environmental challenges associated with industrial growth⁴³. The F-statistics for each regression model indicate the joint significance of the predictors. The F-statistic values range from 71.67 to 103.22, with higher values indicating stronger model fit and the ability to explain the variation in sustainable development outcomes. These results confirm the relevance and significance of AI as a key factor in sustainable development while acknowledging the importance of control and instrumental variables in addressing potential endogeneity. Table 5 illustrates that AI, particularly in R&D innovation and infrastructure development, is critical in driving sustainability outcomes supported by global investment and trade policies⁴⁵.

Table 6 assesses the role of AI in promoting resource utilization sustainability (RUS), with AI dimensions such as AI_RDI, AI_INF, and AI_MKT as independent variables. The results show a positive and statistically significant relationship between AI and RUS, particularly through AI_RDI (coefficient of 0.7428, p < 1%) and AI_INF (coefficient of 0.5312, p < 5%). The findings suggest that AI’s focus on innovation and infrastructure development significantly enhances resource efficiency and sustainability outcomes. The AI_MKT dimension, while positive, has a negligible effect on resource utilization sustainability.

The control variables, including FDI, IND, and TRD, show varying effects on RUS across different models. FDI and IND are positively related to RUS, with significant coefficients in most columns, emphasizing that foreign investments and industrialization play key roles in driving sustainable resource utilization. The F-statistics for all models are high, indicating a good model fit and that the included predictors jointly explain a significant portion of the variation in resource sustainability. The R² values range from 0.70 to 0.83, reflecting the strength of the models in explaining Resource Utilization Sustainability outcomes. The inclusion of GDP, Education, and Carbon Emissions as control variables further supports the analysis, suggesting that wealthier, more educated countries with lower carbon emissions tend to have better outcomes in resource utilization sustainability. These results emphasize the interconnectedness of AI, economic development, education, and environmental factors in achieving sustainability goals. Overall, Table 6 underscores the critical role of AI RDI and AI INF in driving Resource Utilization Sustainability, highlighting the importance of external factors like FDI and IND. Using instrumental variables ensures that the relationships are not biased by endogeneity, reinforcing the robustness of the findings. These insights can help guide future AI policies and investments to enhance resource sustainability.

Figure 6 illustrates the evolution of AI, AI_INF, and its relationship with AI _RD and the overall AI index across four key years: 2010, 2015, 2020, and 2024. The first subplot presents the growth of AI_INF over time, highlighting a steady increase in AI infrastructure development.

The evolution of AI, AI_INF, and its relationship with AI _RD and the overall AI index.

This upward trend suggests that nations are increasingly focusing on enhancing their ICT infrastructure, which is essential for the widespread deployment and application of AI technologies⁴⁴. The growth in AI_INF indicates a positive shift in countries’ capabilities to adopt AI, particularly through advancements in hardware, networks, and digital services that support AI applications across various sectors⁴⁵. Figure 6 (b) compares the development of AI_RDI and the overall AI index over the same period. The plot reveals that AI_RDI has also steadily increased, showing the growing importance of research and development in driving AI advancements. The higher levels of AI_RDI are strongly correlated with the increase in the AI index, demonstrating that nations investing more in AI-related research are likely to achieve higher overall AI development. This relationship underlines the critical role of innovation in AI technologies, such as the development of new algorithms and models, which can lead to improved efficiency and sustainability in various sectors. Figure 6 collectively highlights how AI infrastructure and R&D contribute to the broader growth of AI and its impact on sustainable development. Therefore, Hypothesis 2 is considered valid because Fig. 6 effectively demonstrates the significant role of AI infrastructure and research and development (R&D) in fostering the overall growth of AI. This, in turn, influences sustainable development. The data presented shows a clear connection between the development of AI infrastructure and R&D activities and their positive impact on the broader goals of sustainable development⁴⁴. The evidence provided through the figure supports the claim that advancements in these areas are key drivers in promoting sustainable practices, thus validating the hypothesis.

Table 7 presents the results of the Panel Quantile Regression (PQR) model, analyzing the impact of AI on sustainable development across different quantiles. The findings show a significant positive relationship between AI and sustainable development, with coefficients increasing across higher quantiles. AI exhibits a strong and positive impact, especially at the 70% and 90% quantiles, suggesting that nations with higher levels of sustainable development experience a more substantial effect from AI implementation. The coefficients for AI, AI_RDI, AI_INF, and AI_MKT are all statistically significant across various quantiles, indicating the robustness of AI’s influence on sustainable development. Looking at specific aspects of AI, AI_RDI (AI Research and Development Investment) significantly contributes to sustainable development, with high positive coefficients across the 10–90% quantiles. This indicates that AI-related R&D investments have a pronounced effect, particularly in countries with lower sustainable development levels (10-50% quantiles). In countries with higher sustainable development (80-90% quantiles), the impact of AI_RDI is also substantial, showing that R&D plays a critical role in fostering sustainable development at various stages. The significance of AI_RDI further emphasizes the importance of investments in AI research to drive sustainability. The results also reveal that AI_INF (AI Infrastructure) significantly affects sustainable development, with positive coefficients at all quantiles. The impact of AI_INF is powerful at the higher quantiles (50-90%), suggesting that nations with more advanced levels of sustainable development benefit more from establishing AI infrastructure. This aligns with the idea that a robust AI infrastructure enables countries to harness the full potential of AI technologies, driving improvements in social, economic, and environmental sustainability. The continuous growth in AI infrastructure is crucial for enhancing long-term sustainability outcomes. AI_MKT (AI Market Share) has a more varied impact, as its coefficients indicate positive effects on sustainable development at the 10-30% quantiles but an adverse effect at the 60-70% quantiles. This suggests that the AI market has a differential impact depending on the stage of sustainable development a country is in.

AI market development contributes positively to sustainability at lower quantiles, likely through increased economic activities and technology diffusion. However, the negative impact at the 60-70% range at higher quantiles may indicate diminishing returns or a need for more strategic approaches to AI market development to avoid adverse side effects. This highlights that the relationship between AI market growth and sustainability is not linear and may require nuanced policies to optimize benefits at various stages of development. These findings collectively confirm that AI’s role in sustainable development is multifaceted, with different dimensions of AI (RDI, INF, and MKT) playing varying roles at various levels of sustainable development^44,45. The results of this study contribute to a deeper understanding of how AI can be leveraged to promote sustainability across countries at different development stages.

Figure 7 illustrates the variations in the impact of Artificial Intelligence (AI) on sustainable development across different quantiles of the sustainable development index for 2010, 2015, 2020, and 2024. The x-axis represents the quantiles of sustainable development, ranging from the 10th to the 90th percentiles. The y-axis shows the magnitude of AI’s impact on sustainable development. The plot highlights how AI’s influence differs depending on each country’s sustainable development level. Figure 7(a) shows that AI generally positively influences sustainable development, with the most significant impact observed at 60–70% in 2020 and 2024. These countries, with higher levels of sustainable development, benefit more from the advancements in AI, particularly in areas such as AI infrastructure and research and development (R&D). In contrast, countries with lower quantiles (10–30%) show smaller or even negative impacts, especially for specific dimensions like AI market share, suggesting that the positive effects of AI are more pronounced in countries with more developed sustainable practices. Figure 7(a) also shows how different aspects of AI, such as AI_RDI, AI_INF, and AI_MKT, contribute differently across the quantiles, emphasizing that AI’s role in fostering sustainable development is not uniform across all levels of development. Figure 7(b) illustrates the asymmetric impacts of AI on sustainable development across different quantiles of the SD Index. The quantiles range from 10 to 90%, representing varying levels of sustainable development across countries.

The variations in the impact of AI on sustainable development.

The results show that AI significantly positively affects sustainable development in the 10–80% quantiles, with the peak effect between 60% and 70% quantiles. This suggests that countries with moderate to high levels of sustainable development benefit most from AI advancements, as the technology improves resource optimization, energy efficiency, and environmental management. Figure 7(b) also reveals that the impact of AI’s specific dimensions varies across different quantiles. For instance, AI_RDI shows a particularly strong influence in countries with lower levels of sustainable development, particularly in the 10–50% quantiles, where AI-related research and innovation can provide significant technological leapfrogging. On the other hand, AI_INF (AI Infrastructure) has a more prominent effect on countries that are already performing better in sustainability, evident from its higher impact in the 80–90% quantiles. Finally, AI_MKT (AI Market) has a positive effect on the lower quantiles (10–30%) but demonstrates a negative impact between the 60% and 70% quantiles, suggesting that while AI market share contributes to sustainable development in less developed countries, it may face diminishing returns or challenges in highly developed regions. Our findings first reveal a significant positive effect of AI on sustainable development across the 10–80% quantiles, with the most substantial impact observed between the 60% and 70% quantiles, supporting the validity of Hypothesis 3. Furthermore, a more detailed examination of various AI aspects showed that AI_RDI significantly promoted sustainable development in the 10–50% and 80–90% quantiles, with the most significant effect in countries with lower levels of sustainable development. On the other hand, AI_INF positively influenced sustainable development across all quantiles, with a more substantial impact in countries already exhibiting higher sustainable development levels. However, it is essential to note that AI_MKT positively impacted sustainable development in the 10–30% quantiles but negatively affected the 60–70% quantiles. This suggests that different AI dimensions can play varying roles at various stages of sustainable development.

Table 8 presents the results from the dynamic panel threshold model (DPTM), which assesses the impact of AI on sustainable development across different regimes defined by threshold values. These regimes are categorized based on varying levels of AI development and sustainable development, allowing the model to capture AI’s potential nonlinear and asymmetric effects in different contexts. The results are separated into three regimes: low, medium, and high thresholds, each representing various stages of AI development. In Regime 1 (Low Threshold), which means countries or regions with relatively low levels of AI and sustainable development, the coefficient for the AI index is 0.4561, and it is statistically significant at the 1% level. This suggests that AI positively but moderately impacts sustainable development in these areas. The coefficient for AI_RDI is also positive (0.5321) and significant at the 5% level, highlighting the importance of research and development in advancing sustainable development, even in less developed contexts. However, the coefficient for AI_MKT is relatively low. It does not appear to have as significant an impact, suggesting that AI’s market-related aspects are less influential in countries with low AI development. In Regime 2 (Medium Threshold), which represents countries at an intermediate level of AI and sustainable development, the impact of AI on sustainable development increases. The coefficient for the AI index rises to 0.7234 and remains statistically significant at the 5% level, indicating a more substantial positive effect on sustainable development. Similarly, the coefficient for AI_RDI increases to 0.7890 and is highly significant, reinforcing the importance of R&D in driving sustainable development at this stage. The AI_INF coefficient is also significant and positive (0.4823*), reflecting that improvements in ICT infrastructure have a noticeable effect on promoting sustainability. In Regime 3 (High Threshold), representing countries with advanced levels of AI and sustainable development, the impact of AI continues to be strong, but the coefficients reflect diminishing returns. The coefficient for the AI index reaches 1.0123 and is highly significant, suggesting a robust positive relationship between AI and sustainable development at high levels of development. However, the effect of AI_MKT is more noticeable at this stage (0.3124), indicating that the market and economic benefits of AI become increasingly important as countries move toward higher levels of development. While the AI_RDI coefficient remains significant (0.9312*), the positive effect of AI_INF is especially pronounced in countries with higher levels of sustainable development.

Finally, Table 8 includes additional control variables such as FDI, IND, and TRD, which are crucial in capturing external factors that influence sustainable development. These control variables are statistically significant across the regimes, suggesting that economic factors beyond AI, such as trade openness and industrial development, also contribute to sustainable development. The table’s F-statistics and R² values demonstrate that the model fits well across all three regimes. Regime 1 shows the highest explanatory power, indicating that the model captures the key factors affecting sustainable development in countries at different levels of AI adoption.

The Dynamic Panel Threshold Model (DPTM) is used to explore the impact of AI on sustainable development, accounting for potential non-linearities and threshold effects. The results of this model, as illustrated in Fig. 8, are presented across three key regimes: low threshold (Regime 1), medium threshold (Regime 2), and high threshold (Regime 3). Each regime represents different stages of sustainable development, allowing for a more nuanced understanding of how AI influences sustainability outcomes at varying levels of development. This approach addresses that AI’s impact on sustainability might vary based on the country’s current development level, highlighting the relationship’s non-linear nature. The first regime, Low Threshold (Regime 1), typically reflects countries with lower levels of sustainable development. AI’s overall impact on sustainability in this regime is positive but modest. AI-related factors such as AI_RDI (Research and Development Investment) and AI_INF (Infrastructure) show their most significant coefficients in this regime, implying that nations in this group benefit significantly from AI R&D and infrastructure development. These nations might be in the early stages of integrating AI technologies into their systems, and thus, foundational investments in R&D and infrastructure are crucial for sustainable development.

Dynamic panel threshold model (DPTM) results – the impact of AI on sustainable development across different development regimes.

In the Medium Threshold (Regime 2), representing countries with moderate levels of sustainable development, the impact of AI continues to be positive but may become more differentiated. The coefficients for AI’s dimensions (such as AI_INF and AI_RDI) remain significant but start to show decreasing returns compared to the first regime. This suggests that countries in this group have already begun leveraging AI to address sustainability challenges, and further investment in infrastructure or R&D leads to diminishing marginal returns. However, the positive impact of AI_INF suggests that continued development of AI infrastructure remains vital to ensure the efficient application of AI technologies in solving sustainability issues, especially in sectors like energy and agriculture. The High Threshold (Regime 3) represents the most developed countries in terms of sustainable development. In this regime, the positive impact of AI still exists, but with some variations. While AI_RDI continues to have a positive effect, its influence becomes less pronounced than the first two regimes. This can be attributed to the fact that highly developed nations are likely already at the cutting edge of AI technology, and additional investment in AI R&D may reduce sustainability. Interestingly, the effect of AI_MKT (Market Share) in this regime becomes more evident, emphasizing the importance of the ability to export AI products and technologies for further economic and environmental benefits. One of the key observations in the DPTM plot is that the relationship between AI and sustainable development is not linear. As the thresholds change from low to high development levels, the coefficients of AI dimensions such as AI_INF, AI_RDI, and AI_MKT fluctuate, reflecting the different needs and challenges countries face at various stages of development. For example, while countries at lower development levels require more focus on infrastructure and R&D, more advanced nations might focus on market expansion, technological diffusion, and optimizing existing AI systems. This nuanced approach underscores the importance of tailoring AI strategies based on a country’s level of development. Furthermore, the DPTM results emphasize that AI’s impact on sustainable development is highly context-dependent. As countries progress through various stages of development, their requirements and the role of AI evolve. Countries in lower regimes benefit more from foundational AI capabilities, such as infrastructure and R&D. In contrast, those in higher regimes experience the most value from AI market growth and the application of advanced AI technologies. This suggests that AI policy and investment strategies must be adapted to each country’s specific needs and developmental stage to maximize the benefits of sustainable development. This insight is crucial for designing targeted AI-driven interventions in global sustainability efforts. In conclusion, the DPTM provides a valuable framework for understanding the differentiated impact of AI on sustainable development across various development thresholds. It demonstrates that while AI positively impacts sustainability in all regimes, its effects are most pronounced in the early stages of development, where AI infrastructure and R&D investments are critical. As countries advance, AI-driven sustainability efforts focus on optimizing AI systems, expanding markets, and enhancing cross-border knowledge transfer. Therefore, understanding and implementing AI strategies that align with each country’s stage of development is essential for promoting sustainable development globally. All four regressions successfully passed the nonlinear test and achieved significant threshold values, confirming the validity of Hypothesis 4.

Figure 9 illustrates the varying impact of AI on three key sustainability goals: zero carbon emissions, zero poverty, and zero waste. These goals align with the broader objective of sustainable development, highlighting AI’s role in advancing environmental, economic, and social dimensions. The bars for each goal represent the impact of three AI dimensions: general AI (AI), AI_RDI, and AI_INF. This comparison enables us to understand how different aspects of AI contribute to achieving these goals across various sustainable development parameters. Figure 9 reveals that AI, in its various dimensions, strongly influences achieving zero carbon emissions, zero poverty, and zero waste. For example, AI_RDI (research and development) consistently shows a notable positive impact, especially in the context of zero carbon emissions and zero waste. This suggests that the innovation and optimization of AI algorithms and technologies are crucial in reducing carbon footprints and minimizing waste. Moreover, AI_INF, which encompasses ICT infrastructure development, is key in advancing these goals by providing the necessary technological backbone for implementing AI solutions at scale.

AI on sustainable development goals: Zero Carbon, Zero Poverty, Zero Waste.

Interestingly, Fig. 9 also highlights how AI dimensions have different effects depending on the goal. For instance, AI_RDI has the most significant impact on zero carbon emissions, reflecting the role of AI in optimizing energy systems and enhancing environmental sustainability. On the other hand, AI_INF is particularly beneficial for goals like zero waste, where infrastructure investments help facilitate the widespread adoption of AI-driven waste management systems⁴⁶. Figure 9 underscores that integrating AI technologies across different sectors accelerates progress toward achieving sustainability goals and aligns with the “three zeros” framework: zero carbon emissions, zero poverty, and zero waste.