Introduction
Climate change is considered the most threatening danger to human life. There are countless direct effects of climate change that we are already beginning to feel. Studies indicate that about 3.6 billion people are vulnerable to its destructive impacts. Climate change threatens a chain of disasters that escalate in severity over time. If humans fail to address it soon, millions, and possibly billions, of people could lose their lives.
It is crucial to recognize that climate change, an existential threat to humanity, is a direct consequence of human activities. Since the late 18th century, when the Industrial Revolution sparked a rapid expansion of technology to support economic growth, humans have played a significant role in shaping the current climate crisis.
The critical challenge of climate change lies in its inherent complexity and resistance to control. The factors contributing to this phenomenon are intricately intertwined with modern human lifestyles, characterized by intensive industrialization and excessive energy consumption. These practices deplete Earth’s resources, leading to substantial emissions that disrupt the composition of our atmosphere. Consequently, the global temperature continues to rise, representing the paramount and most perilous environmental impact of industrial development.
While global economic growth has increasingly relied on digital technology since the turn of the millennium, its environmental impact remains less clear compared to traditional industries. Some might believe that digital transformation mitigates the environmental impact of economic growth, but this perception may only be partially true at best. The ongoing evolution of digital technology use generates both direct and indirect environmental effects, which, over time, may surpass those caused by traditional technologies.
The most dangerous aspect in this context is the trend of digital technological advancement leading to sudden mutations that grow rapidly, often outpacing awareness of their various impacts, including environmental ones. A prime example is the exponential growth of generative artificial intelligence (AI) technologies in recent years, which is expected to continue in the near future. At the same time, concerns are mounting over the environmental impacts of excessive energy consumption required to operate generative AI models.
This paper addresses the key features of digital technology’s environmental impact, illustrated with relevant examples. The paper is divided into two sections. The first section discusses the environmental impact of digital technology through three main aspects: carbon emissions, electronic waste and its environmental impact, and the hidden costs of digital technology industries. The second section presents perspectives on dealing with the environmental impact of digital technology and possibilities for reducing it through interventions at the level of national and industry policies, as well as through individual practices.
Digital Technology and its Environmental Impacts
There are several approaches to understanding the environmental impact of digital technology. Digital technology is a key factor in a massive transformation in human lifestyle patterns, encompassing social, economic, and political aspects. When approaching the environmental impact of digital technology from this perspective, the challenge lies in attempting to monitor how the digital transformation alters various aspects of human life, thereby influencing its environmental impact.
For example, the shift towards storing data, information, and documents digitally has a positive impact by reducing the need for physical paper storage. This decreases the demand for paper production, which historically has been a major consumer of trees worldwide. This example illustrates several cases where it is difficult to support a logical view of the environmental impact of digital transformation with precise statistics. Therefore, the paper does not discuss such cases in the following sections.
What can be concluded from this example is that digital transformation intricately intervenes in various daily practices, whether individual, collective, social, or economic. While it can be asserted that this intervention leaves an environmental impact, in many cases, this impact cannot be measured with sufficient accuracy.
On the other hand, digital technology is responsible for the emergence of associated industries. Some of these industries do not differ significantly from traditional industries before the advent of this technology. Industries involved in producing electronic devices and equipment, such as computers, smartphones, servers, and data centers, do not differ significantly in their environmental impact from traditional industries like automobile manufacturing. They consume energy and other resources, leaving behind environmental impacts.
Digital technology has also created industries of a different kind, often referred to as digital data industries. These industries can be likened to traditional service industries but remain unique. Providing access to the internet and offering various services and applications that can be accessed through it, these industries revolve around the transmission, storage, and processing of data. With the exponential growth in data volume and increasing global demand for its analysis and processing, a standalone industry has emerged, driving the growth of another sector—artificial intelligence.
Even though ordinary internet users may feel that the data they interact with is intangible and non-physical, digital data certainly has a material existence. All operations involving digital data are indeed physical and are conducted using tangible devices and equipment.
The cloud is not hanging in the ether, as some might feel when dealing with it. This cloud is actually the sum of operations performed by several physical elements, the most prominent of which are data centers that contain servers for storing and processing data. The network itself consists of cables, some of which are copper, but today, most of which are fiber optics. Cables extend underground and at the bottom of seas and oceans to connect the world’s continents. Landline phone networks still play a major role in data transmission. Still, wireless and satellite networks transmit the most data, with most people relying on their smartphones to connect to the Internet and use its applications.
Data industries are the most prominent face of digital technology, but their physical side remains mostly hidden from non-specialists, that is, most of the daily consumers of its services. Many users of social media platforms do not think about the number and size of data centers that companies use to provide their services to billions of people.
These massive facilities, housing vast numbers of servers, consume substantial amounts of energy to operate continuously and respond swiftly to billions of data transfer and processing requests. In addition to energy, they consume other resources to operate their systems, notably cooling systems, to mitigate the immense heat generated by their electronic equipment. In all these instances, these facilities leave a significant environmental impact.
Energy Consumption and Carbon Emissions
Digital technology plays a dual role in terms of its potential impact on energy consumption and gas emissions. It contributes to the greenhouse effect and results in global warming. Studies suggest that digital transformation could positively influence reducing carbon emissions from traditional industrial sectors.
On the other hand, numerous studies indicate that the expanding use of digital technology increasingly raises energy consumption rates and carbon emissions.
The paper below sheds light on both the aforementioned aspects. It also discusses the environmental impacts of the growth of Generative AI technology, which may represent a significant shift in the scale of digital technology’s environmental impact.
The Potential Positive Impact of Digital Transformation
Based on a report published by the World Economic Forum (WEF), there exists a significant opportunity to incorporate digital technologies into various conventional industries. This integration, undertaken within the broader framework of digital transformation, has the potential to reduce carbon emissions rates across these industries, particularly in their three primary sectors: energy, raw material extraction, and traffic.
These sectors are responsible for most emissions: energy, raw material extraction, and traffic. The sectors contribute 34%, 21%, and 19%, respectively, to total emissions, according to statistical estimates for the year 2020.
The impact that digital transformation can have in these sectors can be achieved through the use of digital technology in:
- Support human intelligence in decision-making: This helps analyze data and make informed decisions about energy and resource consumption.
- Sensor and control technology: collects information and adjusts various processes to make them more environmentally sustainable.
- Empowering digital technologies: Making the most use of the technologies pivotal to the digital economy.
- Developing core technologies: modernizing existing industrial processes using advanced digital technology.
These pathways leverage digital technologies to optimize energy consumption within conventional industrial sectors. They aim to minimize resource waste, reduce the generation of pollutant waste, and enhance transportation processes that significantly contribute to carbon emissions.
The report indicates that the use of digital technologies in the energy sector alone could reduce 8% of total carbon emissions by 2050. This can be achieved by improving the efficiency of carbon-intensive processes, improving energy consumption in buildings, and expanding the use of renewable energy sources using artificial intelligence, cloud computing and the Internet of Things, which is enabled more widely by the use of 5G technology.
The report confirms that digital technologies can reduce carbon emissions by 7% in the materials extraction industry sector by 2050. This can be achieved by improving mining operations and relying on technologies such as big data analysis.
Finally, the use of digital technologies in the transportation sector can potentially reduce carbon emissions by up to 5% by 2050. Achieving this depends on leveraging sensor technologies integrated into the Internet of Things (IoT) and geographic location technologies to gather updated information supporting decision-making processes and optimizing traffic route determination. This approach can effectively reduce emissions in both rail and road-based transportation systems.
Environmental Impact of Total Digital Technology:
In an article published in 2017, the author estimated that digital technologies consume approximately 7% of the total global electricity consumption. However, studies projected that this percentage would increase to 12% by 2020 and continue to grow at an annual rate of 7% until 2030. The operation of data centers consumes about half of this percentage, while devices like computers and mobile phones account for 34%, and industries producing digital electronic devices represent the remaining 16%.
Some estimates indicate that a desktop computer running for 8 hours daily consumes an average of 600 kilowatts/hours per year and produces 175 kilograms of carbon dioxide annually. These figures decrease for laptops, ranging between 150 and 300 kilowatts/hours per year and producing between 44 and 88 kilograms of carbon dioxide annually. In standby mode, these devices consume about one-third of their electricity when actively running.
A less well-known example of the use of digital technology is the process of “mining” for the digital currency Bitcoin. These processes depend on the processing capabilities of many computers that operate for long periods. Electricity consumption in these operations peaked in the first half of 2022, reaching approximately 204 terawatt/hours annually. This consumption significantly decreased in the second half of the same year to about 73 terawatt/hours. Still, it has since risen again and reached approximately 147 terawatt/hours annually as of June this year.
The Environmental Impact of the Data Industry
In 2019, environmental activists in Luxembourg opposed Google’s plans to establish a massive data center in their country. Estimates indicated that the daily operations of this center would require water consumption equivalent to 10% of Luxembourg’s total consumption and about 7% of its total energy consumption in the first phase, increasing to 12% in the second phase.
This example illustrates the scale of energy and water consumption by data centers, particularly in a small country like Luxembourg. This consumption poses a threat to citizens’ daily lives amid other environmental and climate issues that have led to water scarcity.
On the other hand, the massive demand for data centers represents a significant factor in increasing carbon emissions. While the energy industry still heavily relies on environmentally unfriendly sources, it remains one of the most responsible industries for total global carbon emissions.
Data centers are considered the foundation upon which the internet, its services, and applications rely. They are the backbone and pulsating heart of the network. With the massive and continuous expansion of internet usage worldwide, coupled with the growth of technologies like the IoT and AI, there is an increasing need to build more data centers.
As of 2022, the number of data centers worldwide reached approximately 8,000. According to current statistics, the United States alone hosts 5,381 data centers, followed by Germany with 521 centers and the United Kingdom with 514 centers.
Data centers consume large amounts of electrical energy and water to perform their tasks, but at least 50% of this energy is used for cooling purposes. This indicates that the energy use in data centers is not highly efficient. The immense heat generated by running servers and other equipment in the data center means that a significant portion of the electrical energy used is wasted as heat energy, necessitating a comparable amount of electrical energy to cool the center.
Even before generative AI, data centers accounted for about 3% of the world’s total electricity consumption. In 2019, they were responsible for 2% of total carbon emissions, a percentage equivalent to the contribution of the civil aviation industry in the same year. Data centers were also responsible for 15% of the total carbon impact of IT industries and 18% of pollution resulting from digital technologies.
Generative AI
By the end of 2022, the launch of the ChatGPT app by OpenAI captured the attention of millions worldwide. The year 2023 marked a peak in interest in generative AI technology, with many other companies launching their chat applications like Google’s Gemini and Anthropic’s Claude. As public interaction with these systems grows, their development continues rapidly. This heralds a new era where digital technologies increasingly drive human lifestyles.
AI systems require specialized data centers equipped with GPUs (Graphics Processing Units) or custom-developed processors tailored for the tasks these systems rely on. These processors consume more electricity and generate more heat, necessitating additional electricity consumption for cooling operations. However, this difference is just the tip of the iceberg.
The energy requirements of generative AI models depend on their size, measured by the number of variables used in their algorithms. For example, the electricity consumption of applications like ChatGPT is estimated to be equivalent to the consumption of an average of 33,000 American homes. Additionally, a lawsuit revealed that training the GPT-4 model in 2022 led to a data center consuming approximately 6% of the total water in the region.
This illustrates that the training stages of AI models require exceptional energy consumption. Google and Microsoft’s increase in data center usage during the training of the Gemini and Bing models by 20% and 34%, respectively, confirms this.
A 2019 study showed that the BERT model, which contains 110 million variables, consumed the energy equivalent of a round-trip transcontinental flight. In contrast, researchers estimated that producing the GPT-3 model, which contains 175 billion variables, consumed 1,287 megawatts/hours of electricity and produced emissions equivalent to 552 tons of carbon dioxide, which is equivalent to the emissions of 123 gasoline-powered cars for an entire year. All of this was done only while preparing the model for work and before its launch and use.
Estimating the environmental impact of generative AI and language models in daily use is far more challenging than estimating their impact during the development and training stages. A closer approximation involves comparing the energy consumption of a task performed using an AI model versus the same task performed without it.
Some estimates suggest that conducting searches using AI models consumes four to five times more energy than traditional searches. This is particularly significant given that Microsoft and Google have started integrating AI models into their search engines. With billions of people using these search engines, everyday search operations may now consume five times more energy than last year without accounting for other operations that heavily utilize AI models.
In addition, many technology companies are already rapidly integrating AI functionalities into their applications. As a result, the surge in generative AI poses a cumulative threat that could double the environmental impact of not only data centers alone but also digital technology across all its facets.
Electronic Waste and Resource Consumption
Electronic waste consists of various electronic devices and equipment that are discarded after everyday use and need to be disposed of. Common examples of these devices and equipment include computers in all their forms and mobile phones. Additionally, many other devices incorporate digital electronic components in their manufacturing, including household appliances such as refrigerators, washing machines, and others.
The hazard and increasing volume of electronic waste have been gaining attention for a long time. In 2014, a report by the International Labour Organization highlighted that electronic waste was the fastest-growing waste stream among solid waste types, growing at three times the global population growth rate.
One factor contributing to the rapid increase in electronic waste is the fast-paced evolution of digital technologies. This leads to the obsolescence of digital devices and equipment faster than traditional electrical appliances.
Major technology companies release new generations of their products annually. Consumer demand to replace old devices, especially smartphones and computers, isn’t just driven by a desire to keep up with the latest technologies. Many consumers often find themselves compelled to continuously update their devices due to software advancements that require newer hardware to operate efficiently or even function at all.
The Global Volume of Electronic Waste
According to the 2024 report by the International Telecommunication Union (ITU) on global e-waste monitoring, the world reached a new record in electronic waste production in 2022, totaling 62 billion kilograms. This equates to 7.8 kilograms of e-waste per person globally. Only 22.3% of this quantity was documented to be collected and recycled in an environmentally sound manner.
The report also highlights that electronic waste nearly doubled between 2010 and 2022, rising from 34 billion kilograms to 62 billion kilograms. Despite an increase in the amount of e-waste recycled from 8 billion kilograms to 13.8 billion kilograms, the rate of e-waste generation outpaces the rate of recycling by approximately five times.
The European continent leads in e-waste production at 17.6 kilograms per person, but it also has the highest rate of systematic recycling at 42.8%. In contrast, Africa produces the least amount of electronic waste per person, averaging 2.5 kilograms, with documented formal recycling of less than 1% (0.7%).
Health Hazard of Electronic Waste
Electronic waste contains numerous dangerous and toxic elements. According to the World Health Organization (WHO), many of the chemicals used in electronic devices are among the top ten hazardous substances for human health and the environment.
Electronic device users are not typically exposed to these chemicals directly during device use. However, these dangerous elements can spread into the surrounding environment when these devices become waste and improperly handled.
Some wrong practices in handling electronic waste include:
- Dumping waste in the open or into waterways.
- Disposing of electronic waste with regular garbage in landfills.
- Opening, dismantling, burning, or heating electronic waste.
- Attempting to dissolve electronic waste with acids.
- Removing or tearing the plastic covers of electronic devices.
- Manual dismantling of electronic equipment.
All of the above practices are harmful to the environment and human health. They release toxic pollutants, polluting the air, soil, dust, and water in recycling sites and nearby populated areas.
The World Health Organization classifies electronic waste as hazardous to health. Studies have shown that electronic waste contains many chemicals harmful to human health and the environment. People living in areas exposed to electronic waste have higher levels of heavy metals and persistent organic pollutants.
Children and pregnant women are particularly vulnerable to the negative effects of electronic waste, which can compromise their biological systems and organs. Toxic chemicals also impact growth rates and hormone levels.
Manufacturing and Its Hidden Costs
The industrial processes associated with digital technology, from raw material extraction to the manufacturing of electronic devices and equipment, are among the most complex and resource-intensive. The rapid growth in the use of these technologies has led to increased demand for raw materials, thereby expanding the necessary mining operations to meet this demand.
Mining operations are among the most carbon-emitting and environmentally polluting. They lead to irreversible changes in their environments, altering the landscape, removing vegetation cover, and exposing harmful substances to weathering, facilitating their widespread dispersion.
After extracting raw materials, they undergo successive stages of processing and manufacturing. Obtaining required metals such as copper, cobalt, nickel, and lithium involves various chemical processing operations. These processes are considered some of the most polluting, especially if the resulting waste is not properly handled.
However, since most of the extraction and initial processing operations occur in countries rich in these raw materials, which are often poor or developing countries, assurances for safe chemical waste management are minimal or absent. In addition to environmental pollution, these operations also consume large amounts of energy, contributing to carbon emissions that are responsible for climate change.
Finally, in the manufacturing stages, the key components of devices and equipment used in digital technology require highly complex and resource-intensive processes. Among the components that require significant energy and resource consumption in their manufacturing are semiconductors, which are the fundamental components of processing units and other electronic chips, and lithium batteries, which are used in the production of all laptops and smartphones.
Proposed Solutions and Paths to Deal with the Environmental Impact of Digital Technology
Industry and Policy Solutions
When Sam Altman, CEO of OpenAI, along with other prominent figures in the AI industry, signed a statement warning about the existential threat posed by the rapid advancement of AI, experts pointed out that this could be an attempt to redirect global public and policymaker attention away from more immediate and realistic risks. At the forefront of these risks is the significant environmental impact of the runaway growth of generative AI technology.
After over a year, Sam Altman acknowledged that the growth of artificial intelligence technology could lead to an energy crisis. However, he made this acknowledgment only after investing in nuclear energy production, which he presented as a solution to this crisis. These practices highlight that relying on self-regulation by digital technology industries is not the optimal choice. Countries must take legislative action to regulate the growth and operation of these industries and protect the interests of their citizens, especially concerning environmental impact.
Leaders in these industries engage in manipulation to influence decision-makers. In the case of artificial intelligence, legislators tend to focus on the potential political impacts of misuse while often neglecting the environmental impact. This is evident in the European Union’s Artificial Intelligence Act, which completely lacks provisions regulating energy consumption, water usage, and other resources in this industry.
Some countries are beginning to recognize the need for regulatory intervention to reduce digital technology’s increasing environmental impact. In 2021, the French Parliament passed a law to “reduce the environmental impact of the digital world.”
The law focused on achieving four goals:
- Increasing public awareness about the environmental impact of various uses of digital technology. This includes educating individuals and companies about avoiding producing large quantities of electronic waste and unnecessary energy consumption.
- Limiting the continuous renewal of digital end devices (such as computers and mobile phones) by reducing deliberate obsolescence caused by a lack of support from new software.
- Supporting environmentally friendly digital use.
- Establishing environmental regulations to reduce the increasing consumption and emissions of networks and data centers.
Suggestions to Reduce the Increasing Harmful Environmental Impact
There are several potential avenues to reduce the increasing harmful environmental impact of digital technology and related industries. Among the key points that could have a significant impact are:
- Developing extraction and processing processes for materials necessary for digital industries. Enhancing the efficiency of these processes is a broad scope, leading to reduced growth rates and environmental pollution. It is imperative to compel large technology companies to finance these development processes, especially in poor and developing countries.
- Implementing laws to curb the runaway growth of generative AI models. Specifically, interventions should push AI technology developers to abandon scaling up language models to achieve higher software capabilities and explore alternative paths to enhance these models’ capabilities.
- Increasing environmentally friendly processing rates for electronic waste and assisting low-income and developing countries to implement environmentally friendly recycling policies.
- Encouraging digital technology developers to explore environmentally friendly approaches by improving manufacturing processes and designing and operating data centers with reduced energy, water, and resource consumption.
- Exploring the adoption of decentralized network solutions that offer environmentally less damaging alternatives. Distributing data processing tasks across personal computers for many users may obviate the need for substantial capacities in large data centers.
Individual Practices of Users
The direct impact of individual practices of digital technology users on the environmental impact of these technologies is very limited. However, modifying some common practices could have a cumulative positive effect on reducing, albeit slightly, the environmental impact of digital technology.
Electricity consumption practices are among the most prominent individual practices with environmental impact. As mentioned earlier, electronic devices such as computers and mobile phones continue to consume energy even when in standby mode. Therefore, it is advisable to completely turn off devices rather than leaving them in standby mode when not in use. This can save a significant amount of energy, and due to the large number of such devices in use, the cumulative impact of this saving could be meaningful.
The constant pursuit of device upgrades leads to multiple negative effects. It shortens the lifecycle of devices, thereby increasing electronic waste production rates. Moreover, it encourages technology companies to plan incremental updates across multiple generations of products, boosting their sales while often providing users with minimal or insignificant actual benefits from the updates. This unnecessary and continuous upgrading of devices results in increased demand without genuine necessity, leading to higher manufacturing rates and consequently escalating the environmental impact of manufacturing processes, including excessive consumption of raw materials.
In addition, it is important to avoid frequent minor software updates, especially those that necessitate hardware upgrades without genuine justification. In many cases, open-source alternatives such as the Linux operating system can be utilized efficiently on older devices, prolonging their usable lifespan.
Conclusion
As digital technologies permeate various facets of human life globally, their rapid growth, coupled with an increasing demand for energy and resources, has rendered them significant factors impacting climate change mitigation efforts. Notably, climate change poses a substantial threat to human lifestyles and the survival of many, making it imperative to address the intersection of technology and sustainability.
This paper has sought to provide a comprehensive review of the most important aspects of digital technology’s environmental impact. In its first section, the paper highlighted three main aspects of digital technology’s environmental impact. Firstly, it discussed the direct impact of carbon emissions through intensive energy consumption and other resources. The paper specifically emphasized the significant role of data centers in this context and the expected exceptional growth of this role with the advancement of generative AI technology.
The second aspect is the serious implications of the growing volume of electronic waste and the inadequacy of recycling capacities to environmentally friendly levels. The third aspect concerns the hidden impact of many of the manufacturing processes of devices and equipment required for digital technology applications, which have grave environmental consequences.
In the second section, the paper reviewed some intervention possibilities to curb the escalating environmental impact of digital technology, whether at the level of national and corporate policies or the level of individual practices.