IntroductionThe digital era is changing norms in communication, business and investment, policy reform and social interaction. Digitalisation is the use of data and digital technologies1, ranging from the internet, cloud computing, and social media platforms to automation, digital twins, and artificial intelligence. Digitalisation improves process and system efficiency, enables data-driven decision-making, and enhances effective monitoring and reporting of climate-related action2. On the flip side, digital technologies may also amplify climate-related and other environmental risks, as well as societal risks. These include the displacement of workers due to automation, the spread of misinformation, heightened cybersecurity vulnerabilities, increased electricity consumption from expanding information and communication technology (ICT) infrastructure and a rapid increase in electronic waste and other environmental footprints (energy, water, critical minerals)3.Digital transformation means the increasing penetration of, and dependence on, digital technologies across economic sectors, government and public services, and societal functions1, including information dissemination. The speed and extent of digital transformation varies globally, but with a clear general trajectory towards both widening applications and deepening dependence4,5.Scenario assessments help explore and understand potential futures for digital transformation and its implications for climate change mitigation and adaptation6. In turn, this facilitates proactive climate risk mitigation, helping society adapt to new realities while ensuring the benefits of digitalisation are widely accessible. This aligns with the Intergovernmental Panel on Climate Change (IPCC) Working Group III report7,8, which emphasises that digitalisation supports decarbonisation only if appropriately governed. The EU’s twin transition agenda9 further reinforces the entwined green and digital transitions while taking key actions to prevent tensions between them. Similarly, the German Advisory Council on Global Change (WBGU)10 underscores the urgent need for formative political action to ensure that digitalisation serves sustainable development and that a growing digital divide is contained. The importance is also reflected in the Global Digital Compact’s emphasis on aligning digital cooperation with sustainability goals (e.g. addressing digital divides)11.Projecting alternative futures for digital transformation is more complex than in other clearly defined sectors with tangible outputs. As a general-purpose technology with applications across sectors, activities and scales, digitalisation poses unique measurement challenges. There is no single consistent way to quantify physical output, performance or levels of digital transformation. A range of different proxy variables have been proposed in current and historical assessments including: ICT infrastructure (e.g. the number of secure internet servers, installation of industrial robots or robot intensity); access and use (e.g. individuals using the internet, number of devices, broadband and mobile subscriptions); innovation and economic activities (e.g. the number of patents, ICT service exports); knowledge and literacy (e.g. ICT skills); and composite metric-based indicators such as the E-Government Development Index (EGDI) (Supplementary Table 1). Though introduced for other purposes, these variables are also useful for exploring potential future digital transformation within the framework of Shared Socioeconomic Pathways (SSPs).The SSPs define different reference scenarios for global and regional socioeconomic development based on five narratives that span a range of plausible futures and examine resulting challenges and opportunities for climate mitigation and adaptation. The five SSPs are defined by different assumptions on high-level drivers of change: economic growth (GDP), demographics (population), and urbanisation rates. A wide range of additional elements are characterised in both narrative and quantitative forms by numerous studies12,13,14 that interpret the implication of the five SSP narratives for the energy system, land use change, pollution and health, and greenhouse gas (GHG) emissions (Table 1).Table 1 Summary of basic SSP elements24 relevant to digital transformation and proposed extension to include digital transformation explicitlyFull size tableMore recently, several studies have extended the SSP narratives and their constituent elements to assess implications for governance15, extreme poverty16, gender inequality17, income inequality18, net migration and remittances19, structural change20, the human development index21 and armed conflict risk22,23, which involve highly nonlinear phenomena. Resulting ‘SSP extension variables’ are typically based on statistical relationships observed historically with SSP drivers and elements that can then be used to generate future projections based on how these drivers and elements change under the different SSP storylines. All these SSP extensions are designed to enrich the SSP narratives and strengthen their usefulness without altering their fundamental architecture or quantitative interpretation.Despite how widespread digitalisation is today and its anticipated role as a transformative force shaping economic and social life, it is not explicitly mentioned in the SSP narratives. (One small exception is a mention of digitalisation in the SSP5 narrative with reference to global institutions and coordination24).This remarkable omission is evident both in the SSP storylines and in their quantitative elements and interpretations, including the thousands of scenarios using the SSP framework that were reviewed in the IPCC’s recent 2022 assessment25. As a dimension of technological change, digitalisation is implicitly included in the broader context of technological development that varies across the SSPs (Table 1), but this is concerned primarily with energy-related innovation that has a more direct bearing on GHG emissions.Digital transformation introduces significant GHG mitigation opportunities and challenges through the economic and social activities it enables, as well as its direct energy consumption footprint. The lack of quantitative elements in the SSP framework makes it difficult to assess the enabling or exacerbating effect of digitalisation in uncertainty analyses of climate mitigation futures, including those used to inform policies for achieving national net-zero commitments under the Paris Agreement.In this study, we assess global trends in digital transformation and explore alternative plausible futures within the SSP framework. We model the historical relationship between socioeconomic and demographic variables (our independent variables) and digital transformation measured by a widely-used UN index26 (our dependent variable), across countries and through time. These statistical models guide our projections of future digital transformation, taking into account variation across the SSP storylines.As an extension to the SSP framework, our quantitative projections show relative differences in the pace and depth of digital transformation between countries in an uncertain long-term future. Coupled to interpretive narratives informed by an expert workshop, our projections enrich the plausible futures characterised by the SSPs and expand their applicability to assess the implications of digitalisation for energy use, material consumption, GHG emissions, and the achievability of climate targets.The fast-moving pace and varied application of digital technologies make it hard to anticipate future digital transformation, particularly in the long-term27. These uncertainties are magnified by the potential for disruptive technological breakthroughs such as generative AI, or even future artificial general intelligence (AGI)28.Despite these uncertainties, the highly consequential nature of digitalisation for climate change invites initial efforts to fill a knowledge gap that can be critiqued and strengthened within the large SSP community29. We contribute our results as an enabling framework, including six defined use cases for further quantitative analysis of GHG mitigation and adaptation challenges affected directly or indirectly by digitalisation under the societal, economic and demographic influences on global development explored by the SSPs. An important stream of further analysis would specifically examine future discontinuities with observed historical trends, building on initial analytical work in this area30.ResultsDigital transformation historicallyGiven its extensive scope and complexity, digital transformation and its dimensions can be conceptualised in many ways. We use the UN’s EGDI26 as our dependent variable. The EDGI combines information from three subindices measuring: ICT infrastructure, access, and use; digital skills proxied by human capital; and online service availability and accessibility (Supplementary Table 4). Compared to other indices or measures summarised in Supplementary Table 1, the EGDI has better representativeness by covering multiple dimensions of digital transformation, and provides broader temporal and spatial coverage.Using the EGDI as a measure of digital transformation, we show historical trends aggregated to 12 world regions common to global SSP modelling. Our aggregations use population-weighted averages for countries in each region (Fig. 1a). Over the period to 2020, the digital transformation level of North America is generally higher than that of the other 11 regions, with the Pacific OECD catching up and Sub-Saharan Africa the lowest.Fig. 1: Relative digital transformation level of 12 world regions.a Historical digital transformation level based on the E-Government Development Index (EGDI) with representative countries shown for each region, and b Projected digital transformation levels up to 2050 for world average (population-weighted) and North America as an example region. Countries shown in grey on the map were not included in the direct analysis. Solid lines show central estimates. The inset highlights digital transformation levels at 3 selected years, along with upper and lower bounds representing 95% confidence intervals for the predicted value based on the uncertainty in estimated coefficients. This shows modest differentiation across SSPs over the long term (see text for discussion).Full size imageWe tested a range of socioeconomic and demographic factors as predictors (Supplementary Table 5) of country-level digital transformation using a panel dataset of 62 countries over the period 2003–2020. Our econometric model shows that digital transformation historically can be explained by GDP per capita, R&D expenditure, and population size as independent variables with fixed entity (country) effects to control for unobserved heterogeneity (see ‘Methods’ and Supplementary Table 6), such as trust and cultures.These independent variables in our model proxy the economic capacity (GDP/capita), technological innovation (R&D) and market demand (population) that drive digital transformation. As well as controlling for the wide variation in country size in our sample, the population variable proxies not just aggregate demand for digital services but also the availability of human capital.Future digital transformation within the SSP frameworkUsing country-level projections for the three independent variables in our model, we project digital transformation levels to 2050 under each of the five SSP narratives. We limit our projections to 2050 as uncertainties amplify over longer time periods, particularly for fast-changing phenomena like digitalisation. However, our model can be used to project out to 2100 if needed to link digital transformation to long-term carbon budgets (see use case discussions and Supplementary Fig. 1).Our projections of digital transformation go beyond the 0–1 range over which the historical EGDI is normalised, so they should not be interpreted as future values of the EGDI (Supplementary Table 4). However, our measure of digital transformation serves the same function in showing variation between countries and over time (‘Methods’ and Supplementary Table 10). As with the EGDI, absolute values of digital transformation are hard to interpret as they are based on a basket of different subindices and indicators. It is more meaningful to interpret quantitative changes in a relative sense: for example, digital transformation in Sweden is 0.94 in 2020, increasing to 1.25 by 2050, which is in the top percentile of all 62 countries worldwide in our sample. Higher values of digital transformation imply more ICT technologies and infrastructure, stronger human capital and digital skills, and more widespread access to digital services.Our projections of digital transformation show disparities across storylines as expected (Fig. 1b; Supplementary Fig. 1 and Supplementary Excel Sheet 4 for other regional projections). SSP3 (‘Regional Rivalry’) shows the lowest level and SSP5 (‘Fossil-fuelled Development’) the highest, aligning with its storyline of rapid economic and technological progress. North America demonstrates the highest level of digital transformation, generally above the population-weighted world average, based on 62 countries representing 12 regions with available data. Using North America’s 2020 digital transformation level (0.92) as a benchmark, we can use our projections to estimate when other world regions will reach or surpass that level under different scenario assumptions. In an SSP2 future, for example, regions will ‘catch up’ by 2025 (Pacific OECD, Western Europe), by 2035 (centrally planned Asia, former Soviet Union), or by 2050 or later (Latin America, Middle East). The longest lag is for Sub-Saharan Africa, which surpasses North America’s 2020 digital transformation level only by 2085. It is important to note that for Sub-Saharan Africa, the Middle East and North Africa, and Other Pacific Asia, as presented in Supplementary Fig. 1, the subset of countries included in the analysis represents less than 70% of each region’s total GDP when compared to the 180 countries dataset used for out of sample generalisation projections (Supplementary Information—Supplementary Table 14). Within this subset, Sub-Saharan Africa and Other Pacific Asia are more strongly represented by countries with relatively higher GDP per capita, while the Middle East and North Africa are primarily represented by countries with lower GDP per capita. As such, projections for these regions should be interpreted with some caution due to limitations in regional coverage.Ranking the twelve regions by their 2050 digital transformation levels in SSP3 and SSP5 shows both convergence and divergence (Fig. 2). The two highest and the two lowest ranked regions remain unchanged by 2050 (North America and Pacific OECD, and Middle East and Sub-Saharan Africa, respectively). Despite Sub-Saharan Africa being represented by countries with relatively high GDP per capita—potentially leading to an overestimation of its digital transformation ranking—it consistently ranks lowest across scenarios, suggesting that the overall regional ranking remains valid. Some regions, such as South Asia, show a significant change in rank by 2050 (to sixth in SSP3 but tenth in SSP5). This is primarily due to the slower population growth assumed for South Asia under the SSP5 assumptions. These relative changes in digital transformation levels over time between regions can be used in various ways as reference points or modifiers when assessing the implications of digital transformation for GHG mitigation and adaptation challenges (see ‘Discussion’).Fig. 2: Digital transformation by 2050.Regional ranking under SSP3 (Regional Rivalry) and SSP5 (Fossil-fuelled Development) scenarios based on projections for 62 representative countries. The values in parentheses in the legend represent the digital transformation levels corresponding to their respective ranks. For Sub-Saharan Africa, the Middle East and North Africa, and Other Pacific Asia, the countries included in the analysis account for less than 70% of each region’s total GDP when compared to the full 180 country dataset (see Supplementary Information). In this subset, Sub-Saharan Africa and Other Pacific Asia are predominantly represented by countries with higher GDP/capita, while the Middle East and North Africa are represented by countries with lower GDP/capita.Full size imageUneven digital transformation across world regionsWe use percentile categories15 to group countries into five categories from very low to very high digital transformation level based on the distribution of projected values from 2020 to 2050 across all the assessed countries. A country is classified as relatively ‘very high’ if its digital transformation value exceeds the 90th percentile (>1.22)—meaning its level is in the top 10% of all observed values regardless of scenario (Supplementary Table 11). By mid-century, the proportion of countries categorised as having low (blue) or very low (light blue) levels of digital transformation ranges from 9.7% (under SSP5 ‘Fossil-fuelled Development’) to 29% (under SSP3 ‘Regional Rivalry’) (see Fig. 3a and Supplementary Table 12). This translates into 5–45% of the population of the 62 countries in our sample living with relatively low levels of digital transformation in 2050. Conversely, the share of the population in countries with high (orange) or very high (red) levels of digital transformation ranges from 15 to 53%.Fig. 3: Number of countries and population by digital transformation level in 2050 based on our 62-country sample.a SSP1-5 and b SSP3 by GDP per capita. The percentile-based categorisation is based on the 2020–2050 digital transformation level: Very high (>90th percentile), high (>75th and ≤90th percentile), medium (>50th and ≤75th percentile), low (>25th and ≤50th percentile), very low (≤25th percentile), see Supplementary Table 11.Full size imageThe average differences between SSPs shown in Fig. 3 can strongly exacerbate underlying inequalities if these are a salient feature of the SSP storylines—as with SSP3. Figure 3b shows that among the 45% of the assessed population with low or very low levels of digital transformation under SSP3 assumptions, none are in countries with average GDP per capita over $50,000, and the majority are in countries with GDP per capita between $20,000–$50,000. These trends in the digital divide between high-income and low-income countries are generally consistent across different SSPs, although varying in magnitude.These results for our 62-country sample (Fig. 3) are biased towards countries with data availability. To provide a full projected global picture, we extend our dataset to 180 countries, but with caution in interpreting these out-of-sample projections (see Supplementary Tables 13 and 14, Supplementary Fig. 8, for approach and limitations). By mid-century, under SSP3 assumptions, we find that 47% of all countries and 35% of the world’s population (~3.5 billion) in 180 countries (see Supplementary Fig. 2a) have access to very low or low levels of digital transformation compared to the other countries.Digital transformation as an extension to the SSP narratives and elementsThe SSPs comprise both qualitative narratives (or storylines) and quantitative drivers, elements, and other variables. Our quantitative projections of digital transformation in each of the SSPs are summarised in Table 1 alongside the most relevant SSP elements, including demographics, economic development and technological change24.For example, SSP3 ‘Regional Rivalry’ is characterised by relatively weak digital transformation and high digital inequality—a digital divide—in which a significant proportion of the population (~45%) remains at low or very low levels of digital transformation. We also show how digital inequality interacts with income inequality across SSPs with different economic growth and convergence assumptions. For example, SSP3 has the highest proportion of the population facing digital inequality even in countries with higher GDP per capita (using USD 20,000 as a threshold). This is consistent with a storyline in which economic growth does not automatically translate into digital progress, with digital stagnation exacerbating socioeconomic inequalities.Drawing on insights from an expert workshop on digital and climate futures (see ‘Methods’), we also provide qualitative interpretations of these projections that emphasise how digital transformation interacts with other aspects of the SSP narratives (Supplementary Table 2). As an example, in the SSP4 (‘Inequality’) narrative, digitalisation amplifies the high-growth global knowledge economy with rapid structural change among ‘winning’ countries, firm types, and population segments, but with strong negative effects on job losses, skills displacement, and income polarisation as well as the concentration of power undermining political agency.DiscussionDigitalisation is not currently represented within the SSP framework. Our quantitative projections of digital transformation levels across countries and time allow the impacts of digitalisation on GHG mitigation and adaptation challenges to be modelled and assessed consistently and comparatively across SSPs. This adds an important new dimension to the characterisation of reference or baseline scenario uncertainty in global development relevant to climate futures. Explicit representation of digital transformation also improves the SSP framework’s wider policy relevance beyond climate, in line with recommendations by O’Neill6. Digital transformation is an integral element of Sustainable Development Goals (SDGs)31 on innovation, labour markets, and accessible infrastructure (SDG8,9), and a cross-cutting enabler of many other goals, including those on education and cities (SDG4,11).Here we set out five use cases for how our projections can enrich understanding of GHG mitigation pathways and policies, with an illustrative sixth use case for GHG adaptation (Table 2).Table 2 Applications of projected digital transformation levels within the Shared Socioeconomic Pathways (SSPs)Full size tableAssessing direct energy and material consumption and resulting GHG emissions from ICT infrastructureThe data centres, networks and devices that comprise ICT infrastructure have a direct energy consumption footprint estimated at 1000 TWh in 2023, equivalent to 4% of global electricity use32. This is projected to increase rapidly in the next 5 years in some locations33 due to energy-intensive training and inference of generative AI, including large language models. The material needs of ICT infrastructure are small relative to bulk material flows globally, but account for significant shares of certain critical minerals, including rare earths like indium, gallium and germanium34. Our long-term projections of digital transformation levels can be coupled to analyses of the energy and/or material consumption of the continuing expansion of ICT infrastructure across countries and regions under SSP storylines. For example, the marginal effect of digital transformation on energy demand growth observed historically (controlling for other drivers of demand) could be estimated across different time periods and regions. Resulting elasticities of ICT energy demand to deepening digital transformation in turn would enable longer-term projections under different scenario assumptions, including those in the SSPs. Existing SSP variables refine this analysis by contextualising the carbon intensity of electricity and manufacturing activities across different regions (Table 2). Insights on future energy consumption and GHG emissions from ICT infrastructure across countries under SSP storylines, as well as the extent of dependence on rare earth materials, water or recoverable electronic waste, inform policymakers concerned with resource uncertainties and resilience linked with digital trends.Assessing the indirect impact of specific digital applications (e.g. teleworking) on energy consumption and resulting GHG emissions through changes in behaviour, economic activity or societal functionsDigitalisation impacts energy consumption and GHGs indirectly by influencing or changing household and firm behaviour and so the structure of social and economic activity. Compared to direct impacts, indirect impacts are larger in magnitude, more uncertain, and harder to model for methodological reasons, including difficulties in clearly defining system boundaries35,36. Impacts can be net energy-reducing (through substitution, efficiency, optimisation) or net energy-increasing (through rebound, induced demand)37. Our projections of the relative levels of digital transformation across different regions provide a means of scaling and extrapolating estimates of indirect impacts from specific digital applications. For example, Hook et al.38 synthesised data showing that teleworking results in a net reduction in overall energy use in transport and buildings sectors ranging from −15% to −0.01% accounting for variation in study design, context, and geography. These estimates of indirect impacts of energy can be combined with SSP variables capturing heterogeneity in relevant adoption conditions across countries (Table 2) to project future indirect impacts of specific digital applications on energy, consistent with SSP storylines. This helps understand their contribution to mitigation goals or the mitigation challenges they pose.Assessing the indirect impact (enabling effect) of digitalisation as a general-purpose technology on energy consumption and resulting GHG emissionsEconometric models identify relationships between digital transformation and energy demand across the whole economy or in industrial sectors. For example, Briglauer et al.39 found a small but significant negative elasticity of CO2 emissions with respect to broadband connections in OECD countries over the period 2002–2019. Kopp et al.40 found that a 10% increase in firms’ ICT investments was associated with a 0.29% decrease in emissions, with a stronger effect in higher-income countries. These statistical models of the aggregate energy impacts of digitalisation typically control for variation in country size, development stage, trade relationships, and other factors explaining energy demand or GHG emissions. A general finding is that additional digitalisation reduces energy demand at the margins, with a stronger effect in more developed economies. These elasticities identified historically can be integrated with our projected levels of digital transformation across different regions to understand the future economy-wide impact of digitalisation for mitigation goals within different SSPs (Table 2).Assessing the interaction between digital transformation and the stringency of climate policy required to reach emission reduction targetsBy design, SSP reference scenarios assume no additional or new climate policies, and vary widely in GHG emission trends resulting from SSP drivers and other elements. This allows climate policy assumptions to be layered in to identify the stringency required to achieve net-zero or other defined emission reduction or climate stabilisation targets under different SSP uncertainties25. As digitalisation will have significant direct and indirect impacts on energy demand, varying across sector and region, digital transformation will interact with climate policy assessments as both enabler and potential exacerbator. Our projected digital transformation levels allow this to be explored explicitly and quantitatively, including by specifying the relationship between the extent of digitalisation and the carbon pricing levels (or the de-risking of finance) needed to achieve net-zero targets. SSP variables related to governance conditions and effectiveness15 provide additional context, helping to tailor policy insights to the specific capacities of different regions41 (Table 2).Informing global initiatives to ensure universal and equitable access to digital transformation opportunities as part of the SDG agendaDigital transformation plays a fundamental but complex role in progressing towards SDGs31,42. Global modelling analyses quantify the synergies and trade-offs of different strategies across the SDGs43 but have not been able to include the interacting effects of digitalisation. Our projected digital transformation levels allow analysis of relationships between SDG indicators and digitalisation across countries under SSP storyline uncertainties. An important example is the digital divide between countries and over time that is shown clearly in our projections (Fig. 3). Modelling the progression of this divide, and how it can be strategically tackled through digital access and infrastructure investments can improve understanding policy capacity (and coherence) to address specific SDG concerns associated with digitalisation (Table 2).Assessing digital climate services for adaptation planningEffective climate change adaptation depends not only on traditional resilience strategies but also on the use of digital technologies to improve responsiveness and decision-making. Our example use cases for SSP-consistent projections of digital transformation levels have emphasised impacts on GHG mitigation challenges and opportunities (via energy demand and GHG emissions). However, there are many examples of digital applications both strengthening adaptation planning to climate impacts44 (e.g. early warning systems for extreme weather events) but also adversely affecting resilience (e.g. over-dependence on digital infrastructure, digital divides). Our projections can be used to explore how digitalisation influences climate change, adaptive capacity and adaptation planning. For example, empirical studies that show how farmers’ practices are affected by access to accurate weather prediction models or real-time information on market prices for agricultural commodities can be coupled to our projections of how the underlying digital capacities to access and use such tools vary across countries and time (Table 2). This adds a new dimension to SSP-consistent analysis of climate adaptation challenges while emphasising how marked regional variation in digital transformation interacts with the geography of climate impacts. The digital divide potentially adds an additional vulnerability to impact hot spots already subjected to multiple climate stressors45.The digital transformation level projects change in future digital infrastructure, activities, and services, as well as the capacity, including human capital, to support increasing digitalisation. While high digital transformation has the potential to enable sustainable development and net-zero transition pathways, it does not guarantee such outcomes. Digitalisation has its own energy and resource consumption footprint, exacerbated by high carbon intensities of electricity in some futures (e.g. SSP5). Digitalisation also has powerful adverse as well as beneficial impacts across application domains and economic sectors, from governance and societal interactions to industrial processes, jobs and livelihoods.There are two important limitations and considerations for our SSP-consistent projections of digital transformation. The first concerns uncertainty; the second concerns endogeneity.Projections based on econometric models using panel data assume that historical relationships observed between variables will hold in the future. While these models can often account for dynamic changes in the short term, uncertainties compound and amplify over the long term. This is particularly the case for digital transformation given its fast-moving innovation cycles, speed of deployment, and potential for surprises27. Our main projections only run to 2050 and emphasise uncertainty around our central estimates.Some of these uncertainties relevant to digitalisation are represented in the SSP framework and the quantitative projections of drivers, elements and extension variables in the different SSPs. These include rates of technological change and economic growth, and between-country inequalities (Table 1) that indirectly flow through into our projections via their impact on the independent variables in our model.However, additional uncertainties are specific to digitalisation, particularly those related to breakthroughs in AI46 that may result in discontinuous change or even systemic disruption30, rather than the smooth path-dependent trajectories projected by our historically calibrated panel models using future GDP, population, and R&D intensity trends.Against these smooth trajectories, the effects of AI, generative AI, or other digital advances not captured in historical relationships can be explored using sensitivity analysis, sector-specific projections, or further what-if scenario analysis—all of which can be nested within the SSP framework’s long-term variation of macro uncertainties relating to the economy, demography, urbanisation, and technological change.These issues of uncertainty are arguably less fundamental than issues of endogeneity. Our approach treats digital transformation as an internally consistent outcome of socioeconomic development already characterised by different SSPs. However, it is not only an outcome. Digital transformation is also a driver or amplifier of socioeconomic uncertainty, as our analysis of the digital divide compounding income inequality shows. At the same time, digitalisation can contribute to productivity gains and accelerate economic growth.This two-way endogenous relationship—digitalisation as both driver and outcome of change—was also emphasised in our expert workshop mapping of linkages between digital transformation and SSP elements (Supplementary Table 3).As our approach takes the SSP framework as given—both as a set of future narratives and as derived demographic and economic variables used in our modelling—we do not capture the feedback effects of digital transformation on social and economic development. The same limitation applies (by design) to climate impacts on the socioeconomic development pathways, which determine GHG emissions but do not co-evolve with the resulting changes in climate.Other SSP extensions, such as gender inequality and the rule of law, also face this endogeneity problem as they both impact and are impacted by socioeconomic development. As with our projections of digital transformation, the value of these SSP extensions is in making important uncertainties explicit in order to open up further avenues for policy-relevant analysis (see our use cases). What is distinctive about digital transformation, however, is the speed and magnitude of its potential disruptiveness.As an example, digital transformation as a driver of change may undermine governance institutions and political agency through misinformation, unchecked market power of tech companies, and social polarisation under assumptions of weak global oversight3. Generative AI could turbocharge this challenge to the governance landscape necessary for concerted action on global commons problems like climate change, highlighting risks similar to the tragedy of the commons47. To some extent, this is captured implicitly in the SSP3 storyline of fragmentation, divergence, and weakened global institutions. But could AI accelerate and amplify this effect to the point at which it destabilises the fundamental socioeconomic assumptions and relationships on which the SSP framework is built? This eventuality is explored by Carlsen et al.30 in the case of a breakthrough on AGI28. Resulting systemic disruption could require a reimagining of a wholly new scenario architecture for understanding future uncertainties relevant to climate goals.MethodsOur methodology (Supplementary Fig. 3) begins with the pre-selection of independent variables hypothesised to have a relationship with our digital transformation dependent variable. We then build stepwise regression models testing the predictive ability of each independent variable (Supplementary Table 5), yielding a viable historical model for SSP-consistent future projections. Our emphasis in the specification of this historical model is to identify independent variables for which SSP-consistent projections either already exist (e.g. GDP growth as a core SSP driver) or can be estimated indirectly through their observed associations with SSP drivers or elements (e.g. R&D intensity as a function of GDP growth). We can then couple our viable historical model to SSP-consistent projections of the independent variables to estimate within-country change over time in digital transformation in each of the five SSP pathways. Finally, we propose simple narratives to help interpret these quantitative projections of digital transformation, drawing on stakeholder input collected from an expert workshop.Historical dataOur dependent variable is the EGDI26, which we use as an indicator of digital transformation across diverse countries. Published by the UN, the EDGI has the advantage of broad representativeness (extending beyond ICT infrastructure) and the availability of panel data, both in terms of year and country coverage. It combines three components—the telecommunication infrastructure index, the human capital index and the online service index—into a weighted average of normalised scores (Supplementary Table 4). The EGDI is not designed to capture development in an absolute sense; rather, it aims to provide a comparative performance rating of countries. This is consistent with the purpose of our study. The EGDI consists of components that assess the readiness of the digital economy, including infrastructure, human capital, and the prevalence and acceptance of online services. For model estimation, we use data for 62 countries for which there are no missing observations for any variables over the timeframe 2003–2020.To present historical data (and projected results), we group countries into 12 world regions, matching those commonly used by the global Integrated Assessment Models used in quantitative analyses of SSPs (Supplementary Fig. 4). We use population-weighted averages to aggregate digital transformation from countries to world regions. Some regions are represented by a small number of countries—for example, Sub-Saharan Africa is represented by two countries that tend to have a higher GDP per capita than the regional average. This is further detailed in Supplementary Fig. 4 and Supplementary Information. To expand country coverage, we conducted out-of-sample generalisation based on 180 countries instead of the original 62. However, since the model was trained on only 62 countries, we present these extended 180 country results only as Supplementary Information, with potential limitations of this generalisation discussed and visualised.Our independent variables are a range of country-level economic, technological and sociodemographic characteristics associated with digital transformation, drawing primarily on data from the World Bank (Supplementary Table 5).Historical modelWe use stepwise panel data regression to add or remove independent variables systematically based on criteria including statistical significance (p 50th and ≤75th percentile), low (>25th and ≤50th percentile), and very low (≤25th percentile). A country with a score or level above 90th percentile is classified as ‘very high’, meaning its digital transformation level is higher than 90% of all country-year scores across 2020–2050 period. Countries in the ‘very low’ category fall within the lowest 25%. This categorisation is used to analyse the distribution of countries and their populations under different scenarios, along with their associated GDP per capita.SSP narrativesO’Neill et al.24 is the principal reference point for SSP narrative descriptions. Since the SSP narratives nor derived quantifications capture digitalisation trajectories, we interpret our projections of relative digital transformation level alongside the storylines described in O’Neill et al.24 and extend these storylines specifically for digitalisation (Supplementary Table 2). For this, we draw on insights from stakeholders at an expert workshop on digitalisation and climate narratives (Supplementary Table 3). The workshop50 was held on May 13–14, 2024 and brought together 35 researchers and practitioners working on digitalisation’s impacts on energy, the economy, markets, society, and governance.Data availabilityData are provided within the Supplementary Information files. The Python code (standard regression analysis) used in this study is available from the corresponding author upon request.ReferencesVectors of digital transformation. OECD, https://www.oecd.org/en/publications/vectors-of-digital-transformation_5ade2bba-en.html (2019).Digitalization and energy—analysis. IEA, https://www.iea.org/reports/digitalisation-and-energy (2017).Creutzig, F. et al. Digitalization and the Anthropocene. Annu. Rev. Environ. Resour. 47, 479–509 (2022).Google Scholar Measuring digital development: facts and figures 2024. ITU, https://www.itu.int:443/en/ITU-D/Statistics/pages/facts/default.aspx.Digital Progress and Trends Report: Interactive Charts—Digital Adoption World Bank, https://www.worldbank.org/en/data/interactive/2024/03/04/digital-progress-and-trends-report-interactive-charts.O’Neill, B. C. et al. Achievements and needs for the climate change scenario framework. Nat. Clim. Change 10, 1074–1084 (2020).Google Scholar Intergovernmental Panel on Climate Change. Climate Change 2022 Mitigation of Climate Change https://www.ipcc.ch/report/ar6/wg3/chapter/summary-for-policymakers/ (2022).Pathak, M. et al. Climate Change 2022 Mitigation of Climate Change: Technical Summary (Cambridge University Press, 2022). https://doi.org/10.1017/9781009157926.002.The twin green & digital transition: How sustainable digital technologies could enable a carbon-neutral EU by 2050—European Commission. https://joint-research-centre.ec.europa.eu/jrc-news-and-updates/twin-green-digital-transition-how-sustainable-digital-technologies-could-enable-carbon-neutral-eu-2022-06-29_en.Towards Our Common Digital Future https://www.wbgu.de/en/publications/publication/unsere-gemeinsame-digitale-zukunft.Homepage | Global Digital Compact. https://www.un.org/global-digital-compact/en.Riahi, K. et al. The Shared Socioeconomic Pathways and their energy, land use, and greenhouse gas emissions implications: an overview. Glob. Environ. Change 42, 153–168 (2017).Google Scholar Bauer, N. et al. Shared socio-economic pathways of the energy sector—quantifying the narratives. Glob. Environ. Change 42, 316–330 (2017).Google Scholar Rao, S. et al. Future air pollution in the shared socio-economic pathways. Glob. Environ. Change 42, 346–358 (2017).Google Scholar Andrijevic, M., Crespo Cuaresma, J., Muttarak, R. & Schleussner, C.-F. Governance in socioeconomic pathways and its role for future adaptive capacity. Nat. Sustain. 3, 35–41 (2020).Google Scholar Crespo Cuaresma, J. et al. Will the Sustainable Development Goals be fulfilled? Assessing present and future global poverty. Palgrave Commun. 4, 1–8 (2018).Google Scholar Andrijevic, M., Crespo Cuaresma, J., Lissner, T., Thomas, A. & Schleussner, C.-F. Overcoming gender inequality for climate resilient development. Nat. Commun. 11, 6261 (2020).Google Scholar Rao, N. D., Sauer, P., Gidden, M. & Riahi, K. Income inequality projections for the Shared Socioeconomic Pathways (SSPs). Futures 105, 27–39 (2019).Google Scholar Benveniste, H., Cuaresma, J. C., Gidden, M. & Muttarak, R. Tracing international migration in projections of income and inequality across the Shared Socioeconomic Pathways. Clim. Change 166, 39 (2021).Google Scholar Leimbach, M., Marcolino, M. & Koch, J. Structural change scenarios within the SSP framework. Futures 150, 103156 (2023).Google Scholar Cuaresma, J. C. & Lutz, W. The demography of human development and climate change vulnerability: a projection exercise. Vienna Yearb. Popul. Res. 2015, 241–262 (2016).Google Scholar Hoch, J. M. et al. Projecting armed conflict risk in Africa towards 2050 along the SSP-RCP scenarios: a machine learning approach. Environ. Res. Lett. 16, 124068 (2021).Google Scholar Hegre, H. et al. Forecasting civil conflict along the Shared Socioeconomic Pathways. Environ. Res. Lett. 11, 054002 (2016).Google Scholar O’Neill, B. C. et al. The roads ahead: Narratives for Shared Socioeconomic Pathways describing world futures in the 21st century. Glob. Environ. Change 42, 169–180 (2017).Google Scholar Riahi, K. et al. Mitigation pathways compatible with long-term goals. In Climate Change 2022 - Mitigation of Climate Change Working Group III Contribution to the Sixth Assessment Report of the Intergovernmental Panel on Climate Change (eds Shukla, A. R. et al.) Ch. 3, 295–408 (Cambridge University Press, 2022).United Nations. E-Government Development Index. https://publicadministration.un.org/egovkb/en-us/About/Overview/-E-Government-Development-Index.Koomey, J. & Masanet, E. Does not compute: avoiding pitfalls assessing the Internet’s energy and carbon impacts. Joule 5, 1625–1628 (2021).Google Scholar Suleyman, M. & Bhaskar, M. The Coming Wave: AI, Power and the Twenty-First Century’s Greatest Dilemma (Bodley Head, 2023).Gritsenko, D., Aaen, J. & Flyvbjerg, B. Rethinking digitalization and climate: don’t predict, mitigate. npj Clim. Action 3, 1–7 (2024).Google Scholar Carlsen, H., Nykvist, B., Joshi, S. & Heintz, F. Chasing artificial intelligence in Shared Socioeconomic Pathways. One Earth 7, 18–22 (2024).Google Scholar Vinuesa, R. et al. The role of artificial intelligence in achieving the Sustainable Development Goals. Nat. Commun. 11, 233 (2020).Google Scholar Energy and AI—analysis. IEA, https://www.iea.org/reports/energy-and-ai (2025).What the data centre and AI boom could mean for the energy sector—analysis. IEA, https://www.iea.org/commentaries/what-the-data-centre-and-ai-boom-could-mean-for-the-energy-sector (2024).Malmodin, J., Bergmark, P. & Matinfar, S. A high-level estimate of the material footprints of the ICT and the E&M sector. in 168–148. https://doi.org/10.29007/q5fw.Bieser, J. C. T., Hintemann, R., Hilty, L. M. & Beucker, S. A review of assessments of the greenhouse gas footprint and abatement potential of information and communication technology. Environ. Impact Assess. Rev. 99, 107033 (2023).Google Scholar Wilson, C. et al. Evidence synthesis of indirect impacts of digitalisation on energy and emissions. In Proc. 2024 10th International Conference on ICT for Sustainability (ICT4S) 116–127. https://doi.org/10.1109/ICT4S64576.2024.00021 (2024).Pettifor, H., Agnew, M., Wilson, C. & Niamir, L. Disentangling the carbon emissions impact of digital consumer innovations. J. Clean. Prod. 485, 144412 (2024).Google Scholar Hook, A., Court, V., Sovacool, B. K. & Sorrell, S. A systematic review of the energy and climate impacts of teleworking. Environ. Res. Lett. 15, 093003 (2020).Google Scholar Briglauer, W., Köppl-Turyna, M., Schwarzbauer, W. & Bittó, V. Evaluating the effects of ICT core elements on CO2 emissions: recent evidence from OECD countries. Telecommun. Policy 47, 102581 (2023).Google Scholar Kopp, T., Nabernegg, M. & Lange, S. The net climate effect of digitalization, differentiating between firms and households. Energy Econ. 126, 106941 (2023).Google Scholar Yang, C., Gu, M. & Albitar, K. Government in the digital age: exploring the impact of digital transformation on governmental efficiency. Technol. Forecast. Soc. Change 208, 123722 (2024).Google Scholar The Digital Revolution and Sustainable Development: Opportunities and Challenges. Report Prepared by the World in 2050 Initiative https://pure.iiasa.ac.at/id/eprint/15913/, https://doi.org/10.22022/TNT/05-2019.15913 (2019).van Soest, H. L. et al. Analysing interactions among Sustainable Development Goals with Integrated Assessment Models. Glob. Transit. 1, 210–225 (2019).Google Scholar Balogun, A.-L. et al. Assessing the potentials of digitalization as a tool for climate change adaptation and sustainable development in urban centres. Sustain. Cities Soc. 53, 101888 (2020).Google Scholar Byers, E. et al. Global exposure and vulnerability to multi-sector development and climate change hotspots. Environ. Res. Lett. 13, 055012 (2018).Google Scholar Luers, A. et al. Will AI accelerate or delay the race to net-zero emissions?. Nature 628, 718–720 (2024).Google Scholar Ostrom, E. Tragedy of the commons. in The New Palgrave Dictionary of Economics 13778–13782. https://doi.org/10.1057/978-1-349-95189-5_2047 (Palgrave Macmillan, 2018).SSP Extensions Explorer. https://ssp-extensions.apps.ece.iiasa.ac.at/.Hoy, Z. X., Woon, K. S., Chin, W. C., Van Fan, Y. & Yoo, S. J. Curbing global solid waste emissions toward net-zero warming futures. Science 382, 797–800 (2023).Google Scholar Expert workshop on digitalization narratives and climate change mitigation. IIASA—International Institute for Applied Systems Analysis. https://iiasa.ac.at/blog/jun-2024/expert-workshop-on-digitalization-narratives-and-climate-change-mitigation.Download referencesAcknowledgementsY.V.F. and C.W. acknowledge funding from Horizon Europe grant #101056810 (CircEUlar project) and European Research Council grant #101003083 (iDODDLE project), respectively.Author informationAuthors and AffiliationsEnvironmental Change Institute, University of Oxford, Oxford, UKYee Van Fan & Charlie WilsonInternational Institute for Applied Systems Analysis (IIASA), Laxenburg, AustriaCharlie Wilson & Marina AndrijevicStockholm Environment Institute, Stockholm, SwedenHenrik Carlsen & Somya JoshiAuthorsYee Van FanView author publicationsSearch author on:PubMed Google ScholarCharlie WilsonView author publicationsSearch author on:PubMed Google ScholarMarina AndrijevicView author publicationsSearch author on:PubMed Google ScholarHenrik CarlsenView author publicationsSearch author on:PubMed Google ScholarSomya JoshiView author publicationsSearch author on:PubMed Google ScholarContributionsY.V.F. led the analysis and wrote the main manuscript text. C.W. contributed to the conceptualisation, methodology, discussion of use cases, advisory support and writing. M.A. and H.C. provided advisory support on the analysis, use cases and writing. S.J. contributed to the initial discussion of the conceptualisation and use cases. All authors reviewed the manuscript.Corresponding authorCorrespondence to Yee Van Fan.Ethics declarationsCompeting interestsThe authors declare no competing interests.Additional informationPublisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.Supplementary informationR2 Supplementary Materials_R1 CW_E_CleanSupplementary Excel_CleanRights and permissionsOpen Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.Reprints and permissionsAbout this article