The Digital Dilemma: The AI Revolution and Its Environmental Impact

Mads Kjoeller Damkjaer
7 min readOct 26, 2023

--

The exponential growth of digital technology and the proliferation of Artificial Intelligence (AI) tools like ChatGPT have ushered us into an era of unprecedented connectivity and convenience. While these technological advancements have transformed the way we live, work, and interact, they have also brought to light a pressing concern — the environmental cost of our digital age. In this article, we will explore the dilemma in using digital technology and the profound impact of AI on our energy consumption. We will delve into startling statistics and facts to shed light on the critical issue of the environmental consequences of our digital lifestyles.

The Rise of AI and Digital Tools

The last decade and especially the last year, has seen an exponential growth in AI and digital tools. ChatGPT, developed by OpenAI, is just one example of the many AI-driven applications that have permeated our daily lives. These AI tools, powered by deep learning and neural networks, have proven to be incredibly versatile, capable of performing tasks from answering questions and generating creative content to driving autonomous vehicles and diagnosing medical conditions. They are no longer confined to the realm of science fiction; they are here, and they are transforming the world around us.

Energy Consumption in the Digital Age

As we revel in the convenience and innovation that AI and digital technology offer, it’s crucial to acknowledge the tremendous energy consumption they require. The internet, the backbone of our digital existence, is a voracious energy consumer.

If the internet were a country, it would be the third most polluting nation on Earth, behind only the United States and China. This startling fact is attributed to the massive data centers, countless devices, and the complex infrastructure that keeps the internet running.

One eye-opening analogy is that watching 30 minutes of Netflix consumes as much energy as driving 5 kilometers in a gasoline-powered car. This comparison highlights the stark environmental impact of our seemingly innocuous digital entertainment choices.

To substantiate this, a 2019 report by The Shift Project, a French think tank, revealed that the carbon footprint of the digital technology industry accounts for around 3.7% of global greenhouse gas emissions and is predicted to double by 2025. The report emphasizes the importance of recognizing the environmental consequences of our digital age. That is more than the global use of cement (3%), the same as Cropland (1,4%), Grass land (0,1%) and Deforestation (2,2%) all together globally, or more than aviation (1,9%) and shipping (1,7%) in the transportation industry, based on 2016 numbers from Our World in Data. Globally, we emit around 50 billion tonnes of greenhouse gases each year.

The Carbon Footprint of AI Tools

AI tools like ChatGPT are no exception when it comes to energy consumption. The training process for such models requires massive computational power, often running on high-performance GPUs in data centers. This process is energy-intensive, and the carbon footprint of AI is significant. According to a study by Emma Strubell, et al., training a single AI model can emit as much carbon as five cars during their entire lifetime. Moreover, using AI tools like ChatGPT is not without its environmental costs. Generating a single prompt with ChatGPT is estimated to produce around 0.19 kg of CO2 equivalent emissions.

To help illustrate the concept of 0.19 kg of CO2 equivalent emissions, some analogies would be e.g. driving a gasoline car for approximately 0.8 km, boiling a standard electric kettle roughly four times non-stop in regard to electricity usage or running a 100-watt incandescent light bulb for approximately 5 hours in terms of energy consumption.

Keep in mind that this is an approximate figure and may change over time as technology and infrastructure evolve, and as providers implement more energy-efficient measures or shift to greener energy sources.

Green Energy & the Rebound Effect

Addressing the environmental impact of AI and digital technology necessitates a shift toward green energy sources. The adoption of renewable energy, such as wind, solar, and hydropower, in data centers and the broader technology industry is a promising solution. Tech giants like Google and Apple have made significant commitments to using 100% renewable energy in their data centers, reducing their carbon footprint.

However, the question remains: Is green energy sufficient to counterbalance the energy consumption of our digital age? A study by the Shift Project suggests that relying solely on green energy may not be enough, as it could incentivize further energy consumption, known as the “rebound effect.” Therefore, a more holistic approach, including energy efficiency measures and changes in consumer behavior, is necessary.

The rebound effect, also known as the Jevons Paradox, is a phenomenon in economics and environmental science that describes the situation where increased energy efficiency in a particular technology or sector leads to increased overall consumption of that technology or sector. In other words, when an activity or technology becomes more energy-efficient, people tend to use it more, which can partially or completely offset the anticipated energy savings and environmental benefits.

For example, a study published in “Nature” found that the fuel efficiency of passenger vehicles improved by 50% between 1990 and 2015 in the United States. But researchers from the University of Sussex, found a rebound effect, estimating that a 50% increase in fuel efficiency in cars could lead to a 50% increase in vehicle travel, offsetting the expected fuel savings. This behavior-driven rebound effect highlights the tendency for drivers of more fuel-efficient vehicles to drive more or choose larger vehicles, ultimately using more fuel.

Another example is the U.S. Department of Energy reports that energy-efficient upgrades in homes can result in savings of up to 25% on heating and cooling bills. Though, a study in the journal “Energy Policy” suggests that homeowners who invest in energy-efficient improvements may increase their energy consumption for other purposes, such as heating a larger space or using additional electrical appliances, leading to a situation where the energy savings achieved are partly or entirely reversed.

Data centers, integral to digital technology, have indeed made significant strides in energy efficiency. According to the U.S. Environmental Protection Agency, data center energy intensity has decreased by approximately 12% from 2010 to 2018.

Despite the progress in data center energy efficiency, the convenience and popularity of digital technology have led to a surge in digital activities. For example, global Internet traffic increased by a staggering 47% in 2020, as reported by Cisco. Streaming services, like Netflix, saw a surge in usage, contributing to a notable share of the data center energy consumption. The energy-intensive nature of activities such as streaming, cloud computing, and data storage has resulted in the overall energy consumption associated with the digital world remaining substantial, even with the gains in energy efficiency.

Just adressing that this is not only a technical problem, but very much a human behaviour challenge.

Leadership in the AI Era

The digital dilemma calls for leadership in the AI era. Governments, corporations, and individuals must collaborate to strike a balance between technological innovation and environmental sustainability. Policies that encourage the use of renewable energy and the development of energy-efficient technologies are essential.

On an individual level, we can reduce our carbon footprint by using digital tools responsibly, opting for energy-efficient devices, and supporting companies that prioritize sustainability.

Balancing AI focus and sustainability as a leader involves:

1. Innovate Responsibly: Encourage AI innovation that aligns with sustainability goals, investing in eco-friendly technologies.

2. Efficient Resource Management: Prioritize energy efficiency, green energy adoption, and ethical data practices, reducing the carbon footprint of digital operations.

3. Cultural Shift and Advocacy: Foster a culture of sustainability within the organization, engage employees, and collaborate with stakeholders and policymakers to ensure AI advancements align with broader environmental objectives.

How can we avoid the mistakes of the Industrial Era?

The digital dilemma is real, and it’s essential that we confront it. The growth of AI and digital tools has revolutionized our world but comes at a significant environmental cost. As we ponder the dilemma of our digital age, we must be mindful of our energy consumption and work towards sustainable solutions. The future of AI and digital technology must involve more energy-efficient processes, renewable energy sources, and a collective commitment to reducing our digital carbon footprint.

1. How can we develop a comprehensive strategy that leverages renewable energy sources, energy-efficient technologies, and sustainable practices across industries to decarbonize the economy while avoiding the environmental mistakes of the past?

2. What measures can be taken to ensure a just and equitable transition to a low-carbon economy, considering the lessons from the social and environmental injustices that arose during the industrial era, such as poor working conditions and pollution in marginalized communities?

3. How can we promote sustainable consumption and responsible innovation to prevent a repeat of the overconsumption and environmental degradation that characterized the industrial era, while fostering economic growth and technological progress in the AI and digital age?

To address this issue effectively, we must consider not only the benefits of AI but also the long-term consequences of our digital choices. As individuals, as businesses, and as a society, we have a role to play in mitigating the environmental impact of our digital revolution. Only then can we ensure a sustainable and responsible future for our digital world.

With this article we are not talking against either AI or the digital innovation, but highligting that we should not make the same mistakes again.

--

--