11.3.1 Digital Transformation and Sustainability

Digitalisation entails business procedures that take place in digital format. It therefore involves digitisation, to the extent that information stored in non-digital format – including images, words, sound and signals – is fundamentally captured as a digital representation. This is based on binary numbers, facilitating processing by computer machines. Digital business resulting from digitalisation involves both digital and physical activity, with the digital format at its core. Thus, for example, social media, manufacturing and monitoring take place in dual format.

Although other features are also important, it is predominantly the Internet, blockchains, and the Internet of Things which are crucial to the digital world. Digitalisation involves several advantages, as it accelerates activities, facilitating data processing, storage, and transmission. Digital data can thus be processed and transmitted indefinitely without degradation, at high speed and negligible cost. The Internet can be defined as a global system of interconnected computer networks, that is, a set of networked electronic, wireless and optical technologies.Footnote ²⁸ Within the Internet framework, blockchains are distributed ledgers involving expansive lists of records (that is, the ‘blocks’) that are interconnected via cryptographic hashes.Footnote ²⁹ In a broad sense, the Internet of Things defines the material technology, including items such as computers and mobile devices, that are necessary to the functioning of the digital world, while in a narrow sense it indicates the connection of the Internet to the real world via sensors.Footnote ³⁰ Everyday items such as home appliances, cars and electricity grids are indeed connected to communication networks, delivering a range of services.Footnote ³¹ Naturally, the Internet facilitates the exchange of digitalised data by connecting private, public, academic, business and government networks on a local and global scale. For instance, a document transmitted in digital format can be communicated via the Internet to the other side of the world much quicker than in hard copy via courier. It is considered that Internet traffic has tripled over the past five years and more than three and a half billion people – that is, essentially half of the global population – currently use the Internet, in an upward trend from only half a billion in 2001.Footnote ³²

For the reasons above, among others, the digital transformation has unique and significant potential to foster sustainable development. According to the World Economic Forum, digital technologies can reduce global greenhouse gas emissions by 15%, that is, a third of the reduction needed by 2030 to achieve the objectives of the Paris Agreement.Footnote ³³ In the agricultural sector, mobile devices and drones are used to facilitate the monitoring of soil conditions and field spraying, as well as in collecting data on promising crop varieties and field-specific sustainable agricultural practices.Footnote ³⁴ In the area of transport, digital technologies improve efficiency in cargo transportation and facilitate managing traffic data, thus, for example, alleviating or obviating traffic jams, inefficient fuel consumption and polluting greenhouse gases. Moreover, the implementation of digital conferences and emails significantly contributes to reducing travels and thus carbon emissions.Footnote ³⁵ As paper is made out of wood, digitalisation can help to save paper consumption, and thus forests as a core carbon sink. Blockchains and the Internet of Things can meaningfully contribute to protecting biodiversity. For instance, blockchains are useful to track practices such as deforestation and overfishing. Big data collected via Earth remote sensing through satellites can be relied upon to monitor phenomena such as marine, soil and air pollution, fishing, bushfires, floods, earthquakes, hurricanes and droughts, thus facilitating the planning of mitigation measures and the implementation of adequate adaptation policies.Footnote ³⁶ For example, sensors connected to the Internet (‘Internet of the environment’) can contribute to controlling deforestation, which is responsible for around 15% of greenhouse gas emissions worldwide.Footnote ³⁷ Last, but not least, digitalisation facilitates disseminating information about sustainability – it contributes in this way to creating and sharing knowledge among businesses and individuals.Footnote ³⁸

11.3.2 Digitalising the Energy Sector

The most promising sector for matching digital transformation and climate sustainability is that of energy. Fundamentally, the current energy cycle is based on the extraction of fossil fuel, which directly powers the bulk of global industry, transport, and buildings. Indirectly, fossil fuel or low-carbon energy sources – particularly solar and wind energy – are used to generate electricity, which further powers industry, transport and buildings.Footnote ³⁹ Since the 1970s, the energy sector has been a pioneer in digital technologies and the pace of digitalisation is increasing, with investment in digital technologies by energy companies on the rise. Indeed, digitalisation can improve the efficiency and cost-effectiveness of energy production, transmission, distribution and consumption.

From the standpoint of energy production and delivery, the oil and gas sector has traditionally relied on digital technologies, particularly in the upstream value chain. Digitalisation also has the capacity to optimise production processes related to coal. For instance, big data sets obtained via seismic surveys of land and oceans contribute to outlining reservoirs and oil, gas and coal extraction. Prospectively, miniaturised sensors can improve the extraction of coal, oil and gas from reservoirs. In fact, companies rely on digital technology for activities such as modelling and automating the exploration and production of oil and gas and to operate grids. Drones and other digital techniques can be deployed to monitor the functioning of remote pipelines and prevent or repair malfunctioning; this is particularly important for low-cost detection of methane emissions potentially occurring at any stage of oil and gas production and transportation. Sensors can also help to provide information on the state of transmission and distribution in electricity networks and blockchains are particularly efficient in simplifying a decentralised management of energy resources. Furthermore, digitalisation facilitates the integration of electricity produced via renewable energy sources into electricity markets. For instance, smart grids allow the matching of demand and electricity, with consumers also exporting electricity to the grid when equipped with solar panels,Footnote ⁴⁰ whereby the use of blockchain technology facilitates peer-to-peer electricity trading.Footnote ⁴¹ It is calculated that improving electricity storage can lead to avoiding around 30 megatonnes (Mt) of carbon dioxide (CO₂) emissions within two decades.Footnote ⁴² Thus, for example, investment in digital electricity infrastructure and software has increased by more than 20% yearly since 2014.Footnote ⁴³

At the intersection between energy and other industrial sectors, on the consumption side, industry accounts for about 38% of global final energy consumption and 24% of CO₂ emissions.Footnote ⁴⁴ It is considered that the industry sector is currently undergoing the fourth industrial revolution (industry 4.0), which relies on digitalisation rather than automation, including the modular production of smart objects via smart factories.Footnote ⁴⁵ Internally, digitalisation can lead to energy savings via improved production process control, supporting a shift towards a sustainable business model.Footnote ⁴⁶ For instance, 3D printing, also known as ‘additive manufacturing’, allows the production of both plastic and metal parts layer-by-layer,Footnote ⁴⁷ leading to a reduction of wastage in the form of excess scrap materials, lower inventory costs and the possibility of delivering manufactured parts in shapes.Footnote ⁴⁸ Externally, digitalisation enhances connectivity between producers in value chains, improving, for instance, the recycling of materials.Footnote ⁴⁹

In the building sector, which accounts for almost one-third of global final energy consumption and some 55% of global electricity demand,Footnote ⁵⁰ digitalisation has the potential to cut energy use by around 10% by 2040.Footnote ⁵¹ Notably, smart energy management can contribute to ensuring that energy is consumed when and where necessary by improving the responsiveness of energy services to the environment, for instance, through reliance on technologies such as the use of smart sensors in lighting, heating and cooling. Smart energy management can also maximise efficiency in energy consumption by effectively predicting consumer behaviour and simplifying the monitoring of the energy performance of buildings, facilitating the identification of where and when maintenance is required, thus maximising investment.

Transport accounts for around 28% of global final energy demand, and roughly 23% of global CO₂ emissions from fuel combustion,Footnote ⁵² with a growth in energy consumption for transport of almost half to 165 exajoules expected by 2060. Digitalising the manufacturing process for electric- and hydrogen-powered modes of transport, including cars, ships and planes, would significantly reduce greenhouse gas emissions. Introducing viable hydrogen engines into widespread, mainstream production would lead to a significant reduction in greenhouse gas emissions and possibly to the end of the oil era – or, at the very least, to its sharp curtailment. Cars, trucks, planes, ships, trains and their supporting infrastructure are indeed becoming smarter and more connected, improving energy efficiency and reducing maintenance costs. For instance, commercial aircraft and ships are progressively being equipped with sensors, generating big data that optimise route planning and reduce fuel consumption.Footnote ⁵³

11.4.1 Digital Pollution

Systemic studies demonstrate that digitalisation has an overall positive impact on climate change and sustainability.Footnote ⁵⁴ However, together with its significant advantages, the digital transformation also entails relevant shortcomings for the green transition. Some scholars indeed speak of the ‘false green label’ of digital transformation, thus reaching the diametrically opposite conclusion that ‘digital consumerism’ is unsustainable.Footnote ⁵⁵

Fundamentally, while facilitating information transfer, the global digital data flow affects data protection, in turn affecting the reliability of the system in the eyes of the consumer. In particular, concerns are raised by personal data, which amount to about half of the data traded by companies, notably in the telecommunication, financial and business services sectors.Footnote ⁵⁶ Moreover, the digitalisation of the economy raises the question of large companies, which can create monopolies and the corresponding risk of abusing dominant positions. Besides these problems of trust and market management, there is no agreement on the extent to which the digital transformation effectively contributes to the green transition. Thus, while some estimates determine savings triggered by a paperless environment at around 25%, others are far more sceptical, considering that most logging does not necessarily go into papermaking and is rather driven by mere economic considerations.Footnote ⁵⁷

In the energy sector, it is calculated that digitalisation, in conjunction with advances in electrification, might result in substantial greenhouse gas emission impacts, but estimates are uncertain, so much so that energy consumption in road transport might either drop by about half or double, particularly in the light of the interplay between policy and technology.Footnote ⁵⁸ Notably, in the building sector, considerations of a technical and economic nature affect the consumer’s choice as to whether to install control systems such as smart sensors and lighting, whereby financial incentive schemes play a critical role. The extent to which a company engages in digital production is volatile and depends on issues such as the flexibility of supply chains, the impact of digital technologies on employment and cybersecurity; thus, for example, in 2017 the functioning of a nuclear power plant in the United States was disrupted by phishing attacks. In this respect, governments can play a crucial role by supporting research and developments that mitigate investment risk. In the coal, oil and gas production sectors, projects usually take years to develop, whereas digital technologies evolve rapidly and it is therefore difficult to incorporate changes into conventionally defined, multibillion-dollar projects. Moreover, it is considered that the use of digital technologies has the potential to improve technically recoverable oil and gas by around 5% on a global scale, with particular regard to tight oil and shale gas production, that is, more than 10 years of current world oil and gas consumption.Footnote ⁵⁹ Digitalisation also discloses the possibility of further improving geological modelling and mining for coal.Footnote ⁶⁰ Extended exploitability of fossil fuel fosters linear growth and consumption and naturally contributes to an increase in greenhouse gas emissions.

Furthermore, the digital transformation itself is (invisibly) polluting.Footnote ⁶¹ On the one hand, there are indirect emissions triggered by the production of digital devices, such as mobile phones and computers, that is, the Internet of Things lato sensu, involving the extraction of materials such as cadmium, which contribute to greenhouse gas emissions. For instance, mobile phones have a manufacturing phase that accounts for between 85% and 95% of their carbon footprint.Footnote ⁶² Moreover, the use of digital devices leads to the production of about 50 million tons of so-called ‘e-waste’ on a yearly basis, mostly in developing countries, around only 20% of which is recycled, the remainder being dumped into the subsoil in developing countries.Footnote ⁶³ Recycling is particularly complex because e-waste involves common elements, such as lead and chromium, as well as rare minerals, such as scandium and lutetium, which are highly polluting and demonstrably harmful to health.Footnote ⁶⁴

On the other hand, there are direct emissions from digital activities, as the functioning of the digitalised world entails massive electricity consumption. This leads to a predicted use by ICT networks at about 20% of total world electricity consumption by 2025, and consequent greenhouse gas emissions.Footnote ⁶⁵ Thus, improved product efficiency is largely offset by increased overall use because of the so-called ‘behavioural response’. According to some estimates, digital technologies emit around 2% of global emissions, but with a significant increase expected in the mobile communication sector.Footnote ⁶⁶ Other estimates determine the share of global emissions from digital technologies between 3% and 4% by 2020, increasing from 1.5% in 2007.Footnote ⁶⁷ Indeed, global email usage generates as much CO₂ as seven million cars and each email user in the UK sending one less email a day would lead to reducing emissions by 16,433 tons of CO₂, the equivalent of 81,152 flights from London to Madrid.Footnote ⁶⁸ So too, message services and video-streaming have a carbon footprint: considering that using a mobile phone over one hour generates 1.25 tons of CO₂, it is estimated that using a mobile device for 10 years would equal 85–95% of the carbon footprint in the production phase. On average, though, mobile devices have a serviced life of roughly two to three years, following the marketing of new models, according to a system that is profitable to business and increases emissions in the production phase. The European Environmental Bureau considers that extending the use of electronic devices by one year would save the EU a total amount of carbon emissions equivalent to those produced by two million cars.Footnote ⁶⁹ Even devices on standby have a carbon footprint, which a study sets at around 23% of household electricity consumption,Footnote ⁷⁰ whereby it is calculated, for instance, that the largest share of a computer’s global warming potential is due to non-use phases.Footnote ⁷¹ Indeed, electronic devices such as televisions and computers count for half of the idle load of a household, and appliances such as washers and fridges with electronic controls and improved Internet connectivity add around 1300 kilowatt-hours – that is, US$165 on current exchange rates– to a household’s electricity bills.Footnote ⁷² According to the International Energy Agency, 50% of household electricity demand for appliances by 2040 will come from connected devices – on the one hand, creating opportunities for smart responses, but increasing consumption of standby power on the other.Footnote ⁷³

In sum, the current trend of digital consumption in the world is unsustainable.Footnote ⁷⁴ In this context, most ‘digital’ greenhouse gas emissions come from developed countries, which raises questions of human rights implementation.Footnote ⁷⁵ To make this situation more sustainable, we need to think about digitalisation as a constitutive element of transitions and a driver of sustainability in transitions, problematising implications for a just and low-carbon future.Footnote ⁷⁶ Responsible consumption and production is at the core of the transformation and of the implementation of the UN Sustainable Development Goals, calling for industrial processes that implement an efficient use of natural resources, reduce waste generation and adequately manage polluting products.Footnote ⁷⁷ Notably, responsible conduct by digital consumers is required, extending the life of computers and mobile devices as much as possible, turning off videos in online conferences and unsubscribing from mailing lists that are not of interest.Footnote ⁷⁸ Policy and market design are core to achieving these objectives, thoroughly harnessing the potential of digital tools towards sustainability.Footnote ⁷⁹

11.4.2 Data Choices

Where the EU takes climate policy digitalisation from here may seem an entirely practical matter of how far that technical process can be stretched and at what marginal cost. But certainly, if the point is to improve our understanding of climate change and its impact on human concerns, we must at the very least recognise the existence of a broader normative context within which that question arises. Digitalisation, after all, is but one element in a larger process of reducing to numeric form elements of our shared experience that, heretofore, are encountered in some other form. This reductive process is premised upon several assumptions.

First, reducing all of human experience (or, as much of it as possible) to numeric form assumes that, in so doing, we increase humanity’s information database. Of course, this does not mean that we actually know more at that stage of the process. It does, however, mean that more of what we know (or, think we know) can be transmitted electronically and can be ‘processed’ at high speed by artificial minds. This achievement, it is assumed, will ultimately allow us to create new knowledge by combining and recombining what we already know in new ways.

A second assumption underlying numeric reductionism is that the new information it enables us to create has utility – that (in one way or another) it will prove to be of value. Now, there are many who will become wary at the first mention of value. And they will become more anxious still when reminded of Sir Francis Bacon’s observation that knowledge is power.Footnote ⁸⁰ But there is, of course, a third, less obvious assumption buried just beneath the surface of the second one. It is that whatever normative challenges may arise from the quantum leap in size that the human database takes as a consequence of digitalisation can be managed by the same technology that allowed us to do the leaping. After all, our cellular phones already sort out nuisance calls for us and either label them as spam or consign them to a subfile with an appropriately dismissive label – thereby giving us permission to simply ignore them.

Much of the information overload imposed on human beings by digitalisation is managed in just this way – the ‘excess’ material is simply ignored. That coping mechanism imposes costs of its own, of course – a fact with which everyone in complex organisations everywhere is thoroughly familiar. And one aspect of that familiarity is the sneaking suspicion that the organisations to which we ‘belong’ may be plotting our disposal in favour of a mechanised mind that has been taught to do our jobs. This suspicion, and the plotting that inspires it, is based upon the assumed identity of the mind and the brain. It is assumed that ‘computer processes like searching, comparing, classifying, adding, and deleting are used by the brain to process information’ and that what the human brain does is what we generally think of when we speak of making up our minds.Footnote ⁸¹ So, the next logical step in the process of digitalisation and the broader numerical reductionism of which it is a part is the development of artificial ‘minds’ that will compensate for human carelessness and sloth in the handling of our newly created information avalanche.

The development of artificial intelligence as a means for handling the information overload in governance is an ambition of longstanding.Footnote ⁸² The course of artificial intelligence’s development has been structured by the problem of finding the things that a human mind can do that an artificial brain currently cannot and then teaching the computer that new skill. So, when confronted with the fact that no computer has ever become world chess champion, artificial intelligence developers will often respond with but a single word – ‘yet’. And it is true enough that to conclude something cannot be done simply because it never has been is rarely warranted. In this field, as in so many others, embarrassing failures are often the precursors of dramatic breakthroughs. In the light of this, it might seem futile to problematise the role of artificial intelligence in complex governance contexts such as climate policy implementation. However, there may be some advantage to turning the artificial intelligence developer’s motivating question on its head. Instead of asking what a human mind can do that an artificial brain cannot, perhaps there is value in asking what a computer can do that human beings cannot.

As the digitalisation of climate policy implementation continues – as we reduce to numeric form more of what we experience and ‘know’ – the temptation will grow to use our information processing capacities not to expand our database so much as to sift, combine and recombine its contents in an effort to interpret that content. In short, our search for knowledge will drive a search for understanding. Here, the abilities that computers have that we poor humans lack can be significant. For instance, humans are generally incapable of ‘processing’ information that has no context. The human ability to interpret our experiences is culturally dependent in innumerable ways, while the computer’s ability to process meaningless data is so well known that it has its own acronym: GIGO (‘garbage in, garbage out’). Our inability to process data without clarity on its context (and, thus, its significance to us) is the primary reason that whereas a computer can always reach a decision, humans instead must often simply choose. This simple fact may mean that artificial intelligence will be incapable of answering our toughest questions ‘because it is not in contact with the reality we all must live within’ and, so, cannot help us with ‘the dilemmas of cultural debate and hard personal choice’.Footnote ⁸³ However, their ability to process gibberish may not be the weakest point in the armour of artificial intelligence. There may be another even more troublesome thing that computers can do which human beings cannot.

In the 1983 motion picture War Games, the defence of the United States has been consigned to an imponderably complex super-computer. The disaffected programmer who developed the computer has failed to communicate to it his central normative commitment on the subject – that nuclear wars have no winners. In this fraught moment, the computer is ‘attacked’ by a teenaged hacker in search of new and more challenging video games. The hacker inadvertently starts a countdown to Armageddon that is forestalled only when he is able to lock the computer into endless rounds of tic-tac-toe with itself. At length, the computer decides to give up both its games of tic-tac-toe and Global Thermonuclear War in favour of chess. The world was saved, and an aspiring young actor launched a successful career. But it remained unclear how the computer managed to grasp the logically difficult concept of unprovable truths. A mathematician confronted with Golbach’s Conjecture (that every even number can be written as the sum of two prime numbers) must ultimately accept it as an act of faith, as an unprovable truth. They may offer any number of justifications for the choice of when to stop looking for a proof, but it is always a choice rather than a rationally explicable decision.

Confronted with such a problem, a computer would only be able to arrive at a decision if it had been provided with some bit of automatic program that imposed a limit on search functions with the substitution of some sort of probability algorithm. As a practical matter, the substitution of that kind of routine runs the risk of creating a ‘technological black hole, in which humans will not be able to understand the reasoning behind computer results that make key decisions’.Footnote ⁸⁴ But from a normative perspective, it reveals perhaps the most important thing that computers can do which human beings cannot. When pressed to explain one of its decisions, a computer would ultimately have to respond with some variation of ‘because you told me to’.Footnote ⁸⁵ Unlike humans, a computer must uncritically take a programmer’s word for it. Humans must never do this. And, in particular, those humans involved in any form of climate governance can never do it because policy implementation in that field (to an even greater degree than most others) is a form of communicative action – a search for consensus among humans who ultimately must choose because they will never be able to genuinely decide. That search produces actionable results only in the form of ‘moments of unconditionality’ that embody ‘criticisable validity claims’ whose legitimacy lies in the ‘conditions of processes of consensus formation’ that produced them.Footnote ⁸⁶ And, this kind of human action cannot be programmed for the same reason that Goldbach’s Conjecture cannot be proven.