Drivers and effects of digitalization on energy demand in low-carbon scenarios

ABSTRACT

The world is currently facing two socio-technical transitions: shifting to a low-carbon society, and a digital revolution. Despite some claims to the contrary, evidence suggests that spread and adoption of information and communication technology (ICT) does not automatically lead to reduction in energy demand, if this stimulates new energy-using practices or wider economic growth. Despite this policy challenge, the two transitions are often considered separately. This study examines potential drivers of reductions or increases in energy demand due to digitalization identified in recent leading global and UK net-zero transitions scenarios. We analyse the scenarios in terms of effects of digitalization on energy demand by identifying specific direct effects of ICT; indirect and rebound effects in transport and home energy use; and wider effects via economic growth. This analysis implies that the future pathways adopted for digitalization will have a significant impact on future energy demand and hence on the feasibility and acceptability of achieving net-zero goals. We find there are different assumptions and development pathways between scenarios. We also identify a need for better inclusion of behavioural effects and other social science understanding in scenarios on the one hand, and recognition that policy can affect digitalization pathways on the other. Overall, our work suggests opportunities for further research and potential for improving policy interactions between these two transitions, and stimulating greater public debate on the different framings for an ICT-driven low-carbon transition.

Key policy insights

• Modelling and scenario development of low-carbon pathways need to pay more attention to drivers of energy demand, including direct and indirect effects of digitalization.

• Social science understanding of the behavioural change effects of digitalization needs to be considered in assessing energy demand changes.

• Different pathways of digitalization result from individual and collective choices, including technology development, business models and data protection.

• Policy makers should seek to promote pathways that deliver social wellbeing and local environmental benefits, not just economic gains.

KEYWORDS:

• Digital
• energy demand
• low-carbon targets
• long-term scenarios

1. Introduction

The world is currently facing two socio-technical transitions: shifting to a low-carbon society, and a digital revolution. Despite some claims to the contrary, evidence suggests that spread and adoption of information and communication technology (ICT) does not automatically lead to reduction in energy demand, if this stimulates new energy-using practices or wider economic growth (e.g. Lange et al., 2020). Despite this policy challenge, the two transitions are often considered separately.

A small but growing literature considers economy-wide impacts of digitalization on energy demand. In terms of energy and emissions, the International Energy Agency (IEA, 2017) considers the major changes digitalization will have on energy use throughout the economy and concludes that both policy and market design are critical in steering the digital transformation towards a sustainable path. While ICT can reduce energy demand in various sectors, there is potential for big rebound effects throughout the economy (Lange et al., 2020). A report based on expert interviews suggests alignment of policy and (digital) innovation is needed to reduce emissions (Belostotskaya, 2019). A review of indirect energy effects of ICT (Horner et al., 2016) found that the magnitude and even sign of energy impact from ICT is difficult to assess and suggests that more empirical studies are needed on how ICT systems are deployed and used in practice to better identify the parameters driving energy use. Taking a more technical approach, a review of indirect ICT energy impacts found that user-related effects such as rebound needed to be included in life cycle analysis (LCA) (Pohl et al., 2019).

Looking at broader change, a modelling study on the impact of new societal trends on energy demand in Europe (Brugger et al., 2021) notes digitalization as a ‘trend cluster’. All of the trend clusters had the potential to decrease energy demand significantly with guiding policies, with the digitalization cluster responsible for 28% of the energy savings, in addition to influencing other trends. However, all trends could also increase energy demand with a strong rebound effect and a lack of policies; consumer awareness of their energy use or carbon footprint does not automatically translate into behaviour change.

Sector or product specific studies are more common. For example, a review of the studies on the energy impacts of e-materialization (Court & Sorrell, 2020) found optimistic assumptions of perfect substitution while ignoring rebound effects, concluding it was unclear whether e-materialization saved significant energy or would do so in the future. Overall, the literature highlights that digitalization has great potential for energy savings, through both technical change and enablement of social and behavioural change, but realizing these energy savings will depend on actions by users, industry and policy-makers in relation to pathways of development.

We report from work researching the importance of attention to energy demand in the transition to a low-carbon economy, as part of the Digital Society theme of the UK Centre for Research on Energy Demand Solutions (CREDS). We examine how the adoption of information and communications technologies (ICTs) and associated business models and user practices could affect energy demand, as part of the transition to a low-carbon economy. Our work focuses on UK policy, but has implications globally.

In this paper, we investigate key effects of digitalization on energy demand, by examining a number of climate-focused (and sustainability-focused) transition scenarios, UK-specific, Europe-wide and global (see Table 1). In previous work, we compared these scenarios in terms of how they represent digitalization in relation to different dimensions: (1) decarbonizing energy supply or managing energy demand; (2) green growth or shifting to a focus on wellbeing; (3) dominant business models led by large ICT firms, or alternative business models which empower communities and users; and (4) automation for optimizing energy supply and demand or on empowering agency of users (Foxon & Bergman, 2021). In this paper, we further consider the effects of digitalization on energy demand in these scenarios, in relation to: direct effects, indirect effects and rebound – in home energy use and transport; and effects on economic growth, as identified by Lange et al. (2020).

Table 1. The ten selected reports, our shorthand notation, their geographical area, focus and scenarios.

Shorthand	Title, organization or project (reference)	Area	Emissions or climate focus	Digitalization focus	Scenarios
CCC	The Sixth Carbon Budget: The UK's Path to Net Zero (Committee on Climate Change, 2020)	UK	Net-zero emissions by 2050	Enabling net-zero economy	Central scenario and four exploratory scenarios
RSOC	Digital Technology and the Planet: Harnessing Computing to Achieve Net Zero (The Royal Society, 2020)		Net-zero – challenges for digital technologies	Essential role in achieving net-zero	–
CAT	Zero Carbon Britain: Rising to the Climate Emergency (Centre for Alternative Technology, 2019)		Zero carbon UK by 2030	Smart appliances and grid balancing	Single scenario
CREDS	The Role of Energy Demand Reduction in Achieving Net-Zero in the UK, Centre for Research into Energy Demand Solutions (Barrett et al., 2021)		Net-zero emissions by 2050 – role of energy demand	One trend affecting energy demand	Four scenarios by level of ambition
CDBB	Four Futures, One Choice: Options for the Digital Built Britain of 2040, Centre for Digital Built Britain (Broo et al., 2020)	Great Britain	Built environment in 2040 (including 1.5°C target)	Explored in each scenario	Four scenarios (2X2); qualitative
NATGRID	Future Energy Scenarios (National Grid, 2020)		Net-zero emissions by 2050	Enables net-zero through smart flexible grid	Four scenarios (2X2)
INHERIT	Reaching the ‘Triple-Win’, Four Future Scenarios of a Healthier, More Equitable and Sustainable Europe in 2040, INHERIT project (Guillen-Hanson et al., 2018)	EU	Triple Win (health, equality, sustainability)	Explored in each scenario	Four scenarios (2X2); qualitative
BÖLL	A Societal Transformation Scenario for Staying below 1.5°C, Heinrich Böll Foundation (Kuhnhenn et al., 2020)	Global	Staying below 1.5°C through societal transformation	Risk of catalyzing more energy use	Single Scenario
GRUBLER	A Low Energy Demand Scenario for Meeting the 1.5°C Target and Sustainable Development Goals without Negative Emission Technologies (Grubler et al., 2018)		Staying below 1.5°C through low energy demand	Pervasive trend, essential part of net-zero transition	Single scenario
SMARTER	#SMARTer2030: ICT Solutions for 21^st Century Challenges (Accenture Strategy, 2015)		ICT solutions for emissions (and other challenges)	Central role in reducing emissions	Technological (or socio-technical) potential

We find that low-carbon transition scenarios vary in their level of engagement with the digital revolution and the level of interaction between the two transitions. The scenarios include many different, and often implicit, assumptions both about technology and about behaviour and social change in relation to technology, energy and climate change. The role of measures to reduce or mitigate energy demand, in terms of behaviour change and technology uptake and use, is often not represented in detail in these scenarios (Creutzig et al., 2018). This suggests opportunities for improving understanding of interactions between the two transitions in these scenarios and discussion of policy implications of these in order to stimulate greater public debate on the different framings for an ICT-driven low-carbon transition, and for including more work on demand-side policy in scenarios.

The rest of the paper is arranged as follows: section 2 describes our methods and data. Section 3 is our main analysis. Section 4 is a discussion, and section 5 concludes.

2. Methods and data

Here we describe our choice of scenarios and our main analysis, building on the framework of Lange et al. (2020) for investigating effects of digitalization on energy demand.

2.1. Selection of scenarios for analysis

While we are primarily interested in UK energy policy, broader scope scenarios and policies are also of interest. We examined recent papers and reports looking at transitions to low-carbon societies, mostly centred on the UK, and some at the European or global level. Our aim was to interrogate how different perspectives and assumptions towards the low-carbon transition, with different relations towards the digital revolution, can lead to a variety of pathways with different implications for energy demand. This could include different drivers of the transition, different levels of reliance on technological solutions, and different assumptions of behaviour and behaviour change.

This review covers a selection of ten documents for analysis, detailed in Table 1; these are selected to cover a range of perspectives and all but one (CAT (Centre for Alternative Technology, 2019)) include explicit references to the impacts of digitalization. To this end, we chose one considering the full global technological potential of ICT to reduce emissions (SMARTER (Accenture Strategy, 2015)), and a second combining technological potential with current social ‘mega-trends’ (GRUBLER (Grubler et al., 2018)). Another highlights societal change pathways, which they consider lacking in public debate (BÖLL (Kuhnhenn et al., 2020)). The EU project INHERIT (Guillen-Hanson et al., 2018) is of interest for its ‘triple-win’ futures approach – narratives including reducing environmental impacts, improving health, and increasing health equity; we focus on the ‘My life between realities’ scenario, as it emphasizes digitalization and connectivity. Similarly, The Centre for Digitally Built Britain (CDBB (Broo et al., 2020)) has qualitative future scenarios explored across five dimensions: built environment, economy, digital technology, society and natural environment; We focused on the scenario ‘A Legacy of Hope’, a future where the UK is on target to meet its net-zero emission target, while an ageing population leads to a shrinking work force. Both envision future societies that are highly digitalized and interconnected.

While most scenarios consider future technologies, we included CAT (Centre for Alternative Technology, 2019), as it focuses on achieving zero emissions with current technologies. We also included reports from prominent UK actors with different focuses, including the Climate Change Committee (CCC (2020)), which focuses on achieving net zero through decarbonizing supply and uptake of low-carbon technologies; CREDS (Barrett et al., 2021), which focuses on reducing energy demand; the Royal Society (RSOC (The Royal Society, 2020)) which describes digital technologies as an essential part of the transition to a net-zero economy; and the National Grid (NATGRID (2020)), which considers societal change alongside technology.

Including this broad range of scenario types – from a variety of types of actors in both research and policy practitioner communities – is helpful for assessing the spectrum of different drivers and effects of digitalization on energy demand within current low-carbon scenarios.

2.2. Digitalization and energy demand framework

Different studies used different classifications for the effects of ICT on energy demand, but all include direct impacts and indirect or higher order impacts (Court & Sorrell, 2020). We draw on the framework of Lange et al. (2020), who map four effects of digitalization on energy demand in their whole-economy mapping. These are described below and in Table 2.

1. Direct effects: This includes energy consumption of the ICT sector from the production, use and disposal of ICTs. It is affected by the growth of share of ICT in overall GDP, and mitigated by energy efficiency improvements in delivery of ICT services.

2. Energy efficiency and rebound effects: This refers to energy efficiency improvements through application of ICT to the rest of the economy, and resulting rebound effects of increasing service demand.

3. Economic growth: This refers to impacts of digitalization on economic growth, via increased efficiency and labour productivity and decoupling growth from energy consumption.

4. Sectoral change: This refers to the rise of ICT services as a share of the ICT sector and the economy as a whole.

   Table 2. Mechanisms leading to increase (+) or decrease (-) in energy demand.

   Category	Mechanism
   Direct effects	Increasing number of ICT devices and increased ICT infrastructure (+)
   	Energy efficiency improvements in ICT devices and infrastructure (−)
   Energy efficiency and rebound effects	Application of ICT improves energy efficiency in other parts of the economy (−)
   	Increase in service demands in other parts of the economy due to the use of ICT (+)
   Economic growth	Adoption of ICT improves labour productivity of economy (−)
   	Adoption of ICT leads to increasing income inequality (+ or −)
   Sectoral change	Digitalization leads to increasing share of lower energy intensive sectors in the economy (−)
   	Digitalization reduces energy intensity of new services in these sectors (−)

   Note: Based on Lange et al. (2020).

Lange et al. (2020) found that effects I and III tend to increase energy consumption, while II and IV tend to decrease it. We found little reference to sectoral change (IV) in the scenarios, so we do not address it. We are interested in energy demand, and potential areas for social and behavioural change. We therefore chose in our analysis of effect II to focus on two sectors considered in all scenarios: (a) home energy use and (b) transport.

We qualitatively analyse these effects in terms of their assumptions and resulting transition pathways, and the impacts on energy consumption. We did this through coding of each scenario. We first looked through sections specifically addressing energy demand impacts of digitalization relating to ICT, transport, home energy use and economic growth. Second, we searched the texts for the keywords: smart, digital/digit*, ICT, IoT, internet, energy, AVs/automated vehicles, automation, demand, growth, dematerial*, substitution. We sorted results into categories by similarity of their content, summarized in Section 4 and the accompanying tables.

3. Effects of digitalization on energy demand

The findings for direct effects, indirect effects – for transport and home energy use – and effects on economic growth are detailed in Tables 3–6, and discussed here.

Table 3. Direct effects of digitalization and ITC.

Effect	Details
Increased efficiency of devices	There is a need to insure continued improvements of devices – design, repair and upgrades (CCC) Longevity and repair – legislation needed for ‘right to repair’, alongside ensuring repairability, repair infrastructure (BÖLL); behaviour change as ‘individual action’ leads to using devices for longer (RSOC) Trend towards smaller, more efficient devices and commitment of leading operators to continue this trend (SMARTER) Efficiency improvements in digital technologies including standardization and economies of scale, rapid innovation cycles and learning effects; In addition, ‘shorter lifetime accelerates the adoption of more efficient models contributing to rise the conversion efficiencies of the appliance stock more rapidly’ (GRUBLER sup. p. 27).
Increased infrastructure	Increased energy demand for operating a growing ITC infrastructure (RSOC)
Increased number of devices	Increased number of devices (and use) – shaped by the nature of ‘competitive digitalisation’ (BÖLL) Huge increase expected, perhaps 100 billion by 2030 (SMARTER) Drop in number of ICT appliances per household (UK) – by 10% due to demand saturation, sharing and cloud computing (CREDS) Integration of various services into a single device, especially smart phones, can yield ‘up to a 100-fold potential power savings when in use’ (GRUBLER); granular technologies are smaller and have a lower unit cost, therefore lower adoption barriers.
Increased usage	A marked increase in activity (and number of devices) due to digitally enabled services leading to better quality of life; mostly in the Global South (GRUBLER) ‘Individual action’ could lead to a reduction in usage, e.g. people ‘streaming responsibly’ (RSOC).
Increased usage of data	There is a need to ensure continued improvement of data use (data centre efficiency) (CCC) Moving to the cloud and server virtualization increase efficiency of data, but there is need for more efficiency (RSOC) Increased communication leads to more data centres, energy needed for servers and cooling (NATGRID)

Table 4. Indirect effects: efficiency and rebound in transport.

Effect	Details
Reduced travel	Video conferencing and ICT-enabled working from home leads to reduced travel – mentioned in many scenarios. ICT impacts through reduced need to travel for other purposes – health, learning, commerce also reduces travel (SMARTER); virtual substitution reduces travel for various purposes, mostly in the Global North (GRUBLER). Reduced travel leads to reduced level of car ownership, fewer cars manufactured (CREDS).
Improved traffic management and integrated travel	AI can improve significantly improve congestion prediction and smart parking similarly can greatly reduce congestion in busy city areas (RSOC). Similarly, SMARTER find that ICT can reduce travel time, distance and congestion through smart traffic control and fast parking platforms. Apps and other ICT platforms enable ride sharing and car sharing, leading to reduced car km travelled and even a drop in demand for cars, leading to 15% reduction in car production (relative to the baseline) (SMARTER). GRUBLER similarly show an increase in vehicle usage and occupancy halving the number of cars globally by 2050. Integrated urban transport networks include digitalized timetables and ticketing (CREDS)
Digitalization enabled behaviour change	‘Business and operating models that offer ‘mobility as a service’ can leverage digital technologies to support more fundamental changes to how individuals access transport services, reducing the number of personal vehicles on the roads’. (RSOC p. 27) ICT enables significant rise in public transport use through more information, intelligent sensors, transparency and increased personal choice (SMARTER).
Increased efficiency of vehicles and driving	ICT enables more efficient vehicles and smart driving technologies that save energy (SMARTER).
Autonomous vehicles enable change	Ride hailing services provided by autonomous vehicles reduce vehicle ownership and travelled distance as they act as taxis or public transport. GRUBLER also suggest AVs can be part of a shared vehicle fleet, increasing occupancy and reducing car numbers. CDBB similarly suggest that artificial intelligence and AVS enable a smaller vehicle fleet. However, in scenarios where AVs are privately owned, they can increase average miles travelled (NATGRID).
Freight and delivery improvements	‘digital technology can help consolidate demand and facilitate collaboration between delivery and logistics providers’ (RSOC p. 28); Digitalized technologies improve freight logistics (CREDS); Smart logistics for fleets and routes, digital warehouses and operational agility could save 723 billion litres of fuel by 2030 (SMARTER). Dematerialization moderates growth in freight transport (GRUBLER).

Table 5. Indirect effects: efficiency and rebound in home energy use.

Effect	Details
Smart energy systems enable improvements	‘Smart meters could contribute a 25% emissions saving from UK homes by 2035 (compared with 2015 levels), by enabling a flexible, decentralized and decarbonized energy system’ (RSOC p. 26). Energy management and automation could reduce up to 40% of households’ CO2 consumption. (SMARTER) Smart appliances, smart meters enable demand management and load shifting (NATGRID). INHERIT (p. 27) tell us that ‘the behaviour of humans within the house, as well as indoor air quality are carefully monitored and integrated into the smart home. Sensor-based environments adjust to the needs of the inhabitants (light, air quality, heat)’.
Enabled behaviour change	Smart home technologies and data analytics, alongside dashboards and apps allow people to understand their energy usage and emissions, and encourage changes to individual behaviour change, saving energy. (RSOC) Smart meters show how much energy households use, motivating cutbacks in consumption (SMARTER). Further, visualizations help consumers understand their consumption. All in all, this amounts to 8% energy reduction in households. Smart technology integration and ‘digital lighting’ gives consumers more control over energy use, even when away from home (CREDS). Not all scenarios show engagement with smart meters! (NATGRIDS) Cost reflective charges and smart tariffs enable load shifting (CCC)
Increased working from home	Increased working from home increases home energy use (CREDS). CDBB (p. 41) focus on social housing, where ‘homes are designed to enable customization and upgrades, with ICT-connected study and workspaces designed into units to facilitate remote working and learning’
Demand management	‘As many as 8.1 m homes actively manage heating demand with residential thermal storage and load shifting by 2050’. (NATGRID p. 8)

Table 6. Digitalization effects on economic growth.

Effect	Details
Rapid innovation cycles	ICT enables better forecasting and monitoring (of emissions); this could ‘shorten innovation cycles to bring new energy technologies to market faster’ (CCC p. 404).
Data as economic driver	‘Achieving this transition requires business models and technology approaches that can create economic value from the use of data, supported by action at different levels to ensure everyone in society can access these and participate in new forms of economic activity’. (RSOC p. 90). Examples for ‘creating value’ are given, e.g. mobility as a service, distributed energy sales.
Digital skills as economic driver	Digitalization has ‘opportunities for individuals to reskill and upskill as the nature of their job changes due to digitalisation’ (RSOC p. 12) – leading to increased productivity.
Asset-light business models driving growth	‘Innovative new business models are disrupting existing businesses, delivering exponential growth with asset-light business structures. … As digital density increases through rapid smartphone penetration, new business models unimaginable a decade ago have the potential to transform our lives and to drive strong growth opportunities across the different sectors’. (SMARTER p. 27)
Competition driving growth	Competitive digitalization is a catalyst for more growth – and more energy use – if the trend of digital futures shaped by big companies continues (BÖLL)
Productivity driving growth	People can work wherever they are (without travelling to the office) leads to a higher quality of life, and hence productivity: ‘Increase in worker productivity due to a better working atmosphere and living quality’ (SMARTER p. 131).
Job gains or losses, working hours	Some studies suggest 30% of UK jobs are vulnerable to automation, which could have an effect of inequality. However, full time work week has not changed in the past 20 years despite automation – there are various factors involved, including the need for increasing labour productivity. Working less could feasibly be supported by increased automation (CREDS p. 26–7). Other studies highlight the potential for job creation with careful management (e.g. SMARTER).
Automation, AI	Many businesses invest in automation (and other technological solutions) as it becomes cheaper than human labour; this is especially the case with a shrinking workforce due to an ageing population (CDBB). Furthermore, there is ‘Increasing digitalization of many professions to support a smaller workforce in maintaining productivity’ (CDBB p. 48).

3.1. Direct effects of digitalization and ICT on energy demand

As can be seen in Table 3, there is overall an agreed trend of significantly increased usage of ICT, its scope, infrastructure and associated data, and in most scenarios an increase in the number of devices. Different drivers are invoked for the change, such as people seeking better quality of life (GRUBLER), or competition in the market (BÖLL). However, saturation of ICT appliances in the UK or more generally the Global North, integration of multiple devices into one (mostly smart phones), and cloud computing all act to reduce the number of devices (GRUBLER, CREDS).

The rise in data in turn leads to more data centres and higher energy demand, tempered by increasing efficiency in cloud servers and increased data centres (CCC, NATGRID, RSOC).

Devices’ individual energy footprint could be reduced through increased efficiency, including through economies of scale, standardization, and rapid innovation cycles. However, we suggest there is a tension between longevity (through policy or personal responsibility) and repairability (requiring political support), which could act to reduce energy use per device, and rapid innovation cycles, which might encourage shorter device life.

Most scenarios do not quantify direct energy demand of ICT, but consider the effect of efficiency improvements in countering ICT growth. RSOC states ICT energy demand will increase if growth outpaces efficiency, or stabilize if efficiency gains keep pace with growth. SMARTER highlights how ICT growth can enable a reduction of emissions equal to nearly 10 times the sector's direct emissions. GRUBLER looks beyond efficiency to consider how the convergence of multiple services into a single smaller device (e.g. smartphones) can reduce energy demand.

A model of global communication technology (Andrae & Edler, 2015) highlights the importance of the rate of efficiency improvements. Their three scenarios from 2010 to 2030 have different annual improvements in efficiency of production, use, datacentres and network. Their model yields an order of magnitude difference in 2030 ICT electricity footprint between best and worst assumed rates. While the size of the gap between scenarios has been criticized (e.g. Belkhir & Elmeligi, 2018), this nonetheless shows the importance of clear, justified assumptions, including life cycle analysis, about devices, datacentres, other infrastructure, and shifts of energy use between them, as well as a clear narrative about the evolution of the ‘internet of things’.

3.2. Energy efficiency and rebound effects

We consider efficiency and rebound in two key domains where digitalization has promise – transport and home energy use.

3.2.1. Transport

Digitalization could have major effects on transport, as (continued) electrification, automation and connectivity could reshape road transport (IEA, 2017). This matches the suggested impacts of digitalization on transport energy demand in our scenarios, summarized in Table 4.

Nearly all scenarios suggest that teleconferencing and remote working can reduce travel miles and save energy, especially in the Global North. This is invoked most frequently around commuting, but also extended to leisure travel, studying and more. SMARTER is the most optimistic, suggesting remote and flexible working and business meetings, together with ‘E-Banking’ and ‘E-Commerce’, could save 105 billion hours and over 330 billion litres of fuel by 2030.

Traffic management and integrated travel are expected to improve congestion through smarter routing using increased data, sensors, and information provision to phones. ICT could also enable more efficient vehicles and ‘smart driving’ (SMARTER). Freight could gain from smart logistics improving efficiency and enabling collaboration.

Meanwhile, digitalized, flexible, smart ticketing and timetables could make public transport more efficient and desirable. In addition, shared vehicles, and possibly automated vehicles (AVs), could increase vehicle occupancy and usage. All together, these act to provide mobility as a service. Business models are important, however, as privately owned AVs could increase travel (NATGRID). Thus, while INHERIT's qualitative scenario suggest AVs lead to a decrease in car ownership, thus saving energy, the CREDS model cites evidence for AVs having a limited impact on energy demand, and limits their use to niche applications and some fixed long-distance routes (Brand et al., 2021).

INHERIT, CDBB and SMARTER suggest the benefits of digitalization include reduced vehicle ownership and car production, saving energy and material resources. However, behavioural effects are generally not detailed, i.e. what determines people's choice of travel mode, car purchase, or amount of travel. Further, rebound effects, e.g. increased travel as the smart transport system becomes more efficient, are hardly discussed.

There are scenarios which foresee minimal change to current mobility beyond technology, e.g. the CCC's Balanced Pathway shows surface transport emissions reducing by 70% by 2035, primarily through ‘zero emission’ vehicles, with only 17% reduction in car miles by 2050. In contrast, CREDS highlights that alongside new technologies, there is a need for reduced demand for passenger mobility for a credible low energy demand future. Overall, we see diverging pathways here, some building on the promise of technology, focusing on technical emissions reduction; and others more engaged with social and behavioural change, and more likely to consider greater systemic change.

3.2.2. Home energy

The impacts of digitalization on home energy use in the scenarios are detailed in Table 5. There is overall agreement in increased ‘smartness’ of homes, including smart meters, appliances and heating. Most scenarios assume this leads to behaviour change through information provision, increased control (including remotely), smart tariffs and ‘encouragement’. Good engagement with people is thought to increase load shifting, for example through actively managing heating using residential thermal storage (NATGRID).

RSOC optimistically suggests smart meters could contribute a 25% emission savings. However, the motivations for change are not always clear, and as NATGRID points out, people will not necessarily engage with their smart meters. Evidence so far from the UK's smart meter roll out suggests a very modest energy savings in energy consumption, perhaps only 1–3% (Sovacool et al., 2017).

Smart appliances as part of smart energy systems are also predicted to reduce energy demand through automation, with sensors and artificial intelligence adjusting light, air quality and heating for residents’ needs (INHERIT). CDBB suggests smart homes could be designed to facilitate working and studying from home, while CREDS highlights the increased energy usage of working from home.

Overall, there are underlying assumptions that increased ICT and connectivity will improve quality of life while reducing energy use, partly through shaping behaviour and partly through automation. We suggest this is highly optimistic. First, because comfort and convenience are assumed, without considering rebound effects of increased consumption (of heat, lighting, data, or entertainment) due to ease of use and energy efficiency of appliances. Further, an Australian study into smart homes (Strengers & Nicholls, 2017) challenges convenience narratives, suggesting smart homes will engage residents in new forms of household labour, with ‘keeping the home running’ becoming a chore in itself.

Second, it is not clear how much energy can indeed be saved through smart home efficiency measures. Energy savings will need to be prioritized, as current marketing strategies ‘prioritize devices and expectations likely to increase energy demand’ (Strengers & Nicholls, 2017, p. 92). We suggest ‘smartness’ might not be the top priority in saving energy. In the UK, for example, retrofitting homes for insulation and efficient heating systems might save energy more easily than smart heating (and cooling) controls.

3.3. Impact of digitalization on economic growth

We next consider how the scenarios assess impacts of digitalization on economic growth, following Lange et al.’s (2020) third effect: ‘impacts of increasing use of ICT on economic growth, in relation to labour productivity, income inequality and energy consumption’; see Table 6.

We note two things. First, the effects are portrayed primarily as positive. For example, more data ‘creates value’, and drives economic growth (RSOC) although there are challenges related to data privacy (CDBB). Other positive effects include shorter innovation cycles bringing new technologies to market faster (CCC); and ‘material light’ businesses driving growth (SMARTER). Further, increased digital skills and the flexibility and comfort digitalization offers could increase productivity. However, there are also some negative impacts, such as potential job losses due to automation (CREDS). Finally, competitive digitalization is tied with more growth and therefore higher energy use (BÖLL) – an economy-wide rebound effect.

Second, in contrast to the previous categories of digitalization impacts, there is no overall narrative emerging from the different scenarios. Rather, we find a broad range of possible impacts. We suggest further research is needed on the effects of digitalization on economic growth, and through it on energy demand, as the impacts are complex.

4. Discussion

The discussion starts with the results of Section 3 on digitalization and energy demand before considering what this means for pathways, with lessons for policy-oriented scenarios. We then turn to questions of digitalization and economic growth, before considering some implications of our work for scenario building.

4.1. Direct, indirect and rebound effects

The main potential driver of increasing energy demand in the digitalization transition is the massive projected increase in the number and usage of ICT devices and the associated infrastructure. Where energy use represents a significant cost for suppliers and users, this could stimulate improvements in energy efficiency and some changes in user behaviour that act to offset this increasing energy use. Both the increase in demand and the extent of this offsetting will depend significantly on the path of digitalization.

Indirect effects of digitalization on home energy use and transport are also likely to be significant. Smart energy systems within homes offer the potential for users to reduce their energy demand whilst maintaining or enhancing service provision, and to provide services to the system, such as demand management and load shifting. However, the realization of these benefits will depend partly on user behaviour, and on the respective reliance on automation or on user behaviour shifts. The scenarios show different balances between automation of energy services and artificial intelligence to reduce the need for user inputs, and empowerment of users to engage with the technologies and manage their demand more actively (Foxon & Bergman, 2021).

Increases in ICT-enabled working from home are assumed to lead to reductions in travel and overall energy use. However, recent research (Caldarola & Sorrell, 2021) has argued that there may be significant rebound effects that could negate some or all of the energy saving, such as: home-workers making additional trips for shopping or leisure; longer commutes on days in which they do travel to work if they choose to live further away; and additional home energy use. If these types of rebound effect are not fully represented in low-carbon scenarios, they may over-estimate the potential energy savings of digitalization.

Indirect effects in relation to transport include virtual interactions substituting for travel, smarter traffic flow management in cities, and facilitation of new ‘mobility as a service’ business models. These effects could lead to significant reductions in energy demand. However, these energy savings could be offset by rebound effects for users and businesses. Better traffic flows could lead to increased travel, and privately owned autonomous vehicles could lead to new energy-demanding practices. Cheaper and more flexible services could lead to higher energy service demands, such as improved logistics leading to increases in on-demand delivery of goods and services.

4.2. Digitalization pathways

Our findings of different transport futures with different end uses and varying levels of energy demand reflect a recent European study (Noussan & Tagliapietra, 2020), which found that the potential effects of digitalization on passenger transport could lead to diverging effects in both energy demand and emissions, depending on how new technologies and services are rolled out. They describe two opposite scenarios: ‘responsible’ digitalization and ‘selfish’ digitalization. Responsible digitalization involves a shared evolution towards optimizing a digitalized system, considering community benefits, not just additional services to individuals. For transport, this means increased occupancy and better use of public transport, and more localization and flexible working reducing the need to travel; while selfish digitalization implies maximizing individual benefits and reducing the cost of travel by car, supported by AVs (Noussan & Tagliapietra, 2020). This study found that transport energy demand could be 60% higher in the selfish scenario compared to the responsible scenario.

Most of the scenarios analysed here suggest a future closer to responsible digitalization in terms of energy demand reduction and public benefits. However, they do not necessarily discuss governance, business models or social change needed to ensure this outcome. For example, CDBB, INHERIT and SMARTER all suggest AVs (and other measures) will reduce the number of private vehicles, but do not offer governance or policy to address the potential rise in travel (and energy) demand. Similarly, a rise in vehicle occupancy is assumed in several scenarios, but only CREDS backs up this shift with research and suggested policies.

How the different effects of digitalization will play out will depend on the social and technological evolution of ICT systems, as well as governance. This will require policy-oriented scenario analysis to pay more attention to factors including user engagement with technologies, consumer awareness and new user roles as ‘prosumers’; efficiency and longevity of devices, and changes to number of devices, usage and data; and development and adoption of new business models.

In our previous work (Foxon & Bergman, 2021), we characterized these scenarios in terms of their underlying assumptions, and found that high-level framings influence the pathways of the digital transition. For example, the extent to which ownership of key ICT businesses continues to be highly concentrated or is challenged by new business models, or even public ownership, leads to diverging pathways, with implications including the use of data as a public good or as a commodity. Different pathways also emerge from approaches to automation of services versus empowering user engagement, including in terms of users’ influence on digital technologies (as in GRUBLER). These pathways also affect energy demand.

Policy-oriented decarbonization scenarios need to pay greater attention to future pathways adopted for digitalization, considering the importance of energy demand reduction in achieving net-zero. Scenarios should go beyond simplistic assumptions of how ICT usage can reduce energy demand and how uptake can be maximized, and examine direct, indirect and rebound effects associated with different pathways for digitalization. This would help to inform policy focus and decisions affecting the role of digitalization in relation to the feasibility and acceptability of achieving net-zero goals.

4.3. Green growth and energy demand

In most of the ‘green growth’ decarbonization scenarios, higher levels of investment and reducing costs of digitalization are assumed to drive higher labour productivity and overall economic growth (Foxon & Bergman, 2021). This higher level of economic activity is assumed to feed through into overall improvements in social welfare. However, some scenarios, such as CREDS, argue that focusing the advances from digitalization and other social megatrends on measures to improve social wellbeing, health and local environmental benefits, rather than on purely economic gains, could enable energy demand reductions and enhance the feasibility and acceptability of a net-zero transition. The impact of digitalization and automation on quality and quantity of jobs also requires more careful examination. Further, quantitative models show contradictory impacts of climate policy on the macro-economy, depending on their representation of finance, innovation and technological change (Mercure et al., 2019). While innovation is often described as an economic driver, energy innovation will not necessarily fuel economic growth.

Scenarios that focus on supply-side solutions tend to be framed within a green growth narrative that assumes decoupling of economic growth from carbon emissions, including through the role of ICTs in providing dematerialized services. Scenarios that focus on demand-side solutions tend to be more sympathetic to a ‘degrowth’ narrative, with BÖLL explicitly discussing degrowth, arguing that this focus would enable continued wellbeing of citizens whilst achieving rapid carbon reductions. CDBB explores scenarios where the economy shrinks due to an ageing population and reduced workforce, but wellbeing is maintained. GRUBLER combines an emphasis on demand-side solutions with a narrative suggestive of green growth – although this has been interpreted as a degrowth scenario due to the reduction of material throughput (Keyßer & Lenzen, 2021). We suggest that the green growth scenarios do not sufficiently engage with the tensions between growing the economy and reducing energy demand. Digitalization driving energy efficiency improvements and economic growth could lead to increased demand for goods and services, including demand for energy, causing an economy-wide rebound effect (Brockway et al., 2021).

4.4. Implications for scenario building

We next consider implications for scenarios from our work. First, there is limited engagement and inconsistent representation of how digitalization will affect demand for energy services. This is despite evidence that most measures for a balanced net-zero pathway require societal/ behavioural changes alongside adoption of low-carbon technologies (Committee on Climate Change, 2020). All the scenarios incorporate some action for societal and behavioural changes to address the demand-side, but they differ in the relative importance that they assign to these and the underlying drivers of change. This is in line with the broader literature: questions about demand-side options tend to focus on public acceptability and the feasibility of economic and social structural changes that may be needed to introduce these options, rather than investigating the drivers of increasing service demand (Eyre & Killip, 2019). Further, a study reviewing life cycle analysis (LCA) of ICT (Pohl et al., 2019) found a technical, product-oriented focus, with behavioural effects, including various rebound effects, not considered; they suggest incorporating usage behaviour into LCA.

Second, we suggest some scenarios have a simplistic approach to both individual behaviour and social change. Further, we note that some scenarios assume optimized social engagement with technology that maximizes energy savings. There are explicit or implicit assumptions in some narratives about behaviours compatible with, or even assisting, deep decarbonization. We are in agreement with other research that finds social science approaches to understanding the complex dynamics of policy implementation needed to reduce emissions are not well integrated into climate scenario analysis (Hewitt et al., 2021). We suggest digitalization pathway narratives should draw more on theories of individual behaviour (e.g. behavioural economics) and social behaviour (e.g. practice theory) and empirical research on human interaction with energy-related technology (e.g. smart meters, see section 3.2.2), drawing lessons for policy makers on governance of digitalization of society that will maximize energy demand reduction.

Third, we suggest more attention needs to be given to the plausibility of scenarios. A recent report (Stammer et al., 2021) considers what would make future scenarios plausible in the context of climate change, breaking with ‘the optimism bias that pervades much of existing decarbonization research’ (p 30). Whilst we have not analysed the models underlying these scenarios in this paper, our findings support the insights of recent research that net-zero modelling needs to consider how to incorporate socio-political changes and new economic approaches (Pye et al., 2021); management of uncertainty (Hewitt et al., 2021; Pye et al., 2021); and to differentiate between normative approaches – which recommend what strategies actors take – and positive approaches – describing what actors are actually observed to do (Mercure et al., 2019). Hughes et al. (2013) argue that scenarios and pathways that focus too much on present actors’ roles can miss disruptions and transformation inspired by visions and values. Overall, scenarios would do well to consider the social changes they assume, and justify the plausibility of such changes, including the precipitating events and actors’ changing roles.

5. Conclusions

This paper has highlighted the importance of considering the interactions between the digital revolution and the net-zero carbon transformation of energy and economic systems. We find there is a need for this to be further explored through scenario analysis and participatory dialogue. Assumptions about social, technological, and economic drivers lead to very different futures.

In the net-zero scenarios explored here, we have found that they often include optimistic assumptions about how adoption of digital technologies and related services can help to reduce energy demand, but without discussing the governance, business models or social change needed to ensure these outcomes. This suggests the need for more explicit representation of these factors and how they might influence pathways of digitalization in future net-zero scenarios.

This also suggests that policy support for technology is not enough, as there are different pathways for technological development and adoption. Scenario analysis can inform policy for directionality of development of digital innovation (Mazzucato, 2016) to support environmental goals such as energy demand reduction and decarbonization, alongside social goals such as privacy of data and reducing harmful content on websites. Our findings are in line with IEA (2017) that government policy design is critical for digitalized energy systems to follow a sustainable trajectory. For example, a digitalized transport system providing efficient, affordable travel requires behaviour-targeting measures to prevent rebound through increased travel (and therefore energy demand), especially in low-occupancy vehicles.

A transition to a net-zero society by 2050 or earlier will require many interacting changes in technologies, institutions, business models and user practices, in which ICTs will have a crucial role to play. Achieving wide public consent for these changes and overcoming the resistance of vested interests to changes will require informed public debate on these issues. The further development of more integrated low-carbon and ICT scenarios, explicitly including different drivers and causation patterns, could play an important role in this. The role of digitalization in these debates is important, as ICTs and associated new business models and practices have the potential for reducing energy demand through improving energy efficiency and stimulating economic structural changes, but also the potential for increasing energy demand through direct energy use and stimulating re-spending leading to economic growth and economy-wide rebound effects.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

This work was supported UK Research and Innovation through a grant to the Centre for Research into Energy Demand Solutions (CREDS) (Ref EP/R035288/1).