Abstract
Technology stands at the centre of human development. With it comes numerous problems, including how humanity has changed and continues to change its environment. This article explores some of the problems surrounding energy security, specifically those associated with Industry 4.0 (the fourth industrial revolution) as well as the field of Artificial Intelligence and Machine Learning (AI/ML). While there are few definitive solutions to the problems created by improving old technologies and implementing new ones, change is dependent on sparking conversations about sustainability and ultimately the type of future humanity chooses to create for itself.
Introduction
Technology has been at the core of humanity’s development throughout history. From wheels and levers to steel, combustion engines, radios, spaceflight and the internet [1]. The last 150 years of human progress, however, has led to rapid and, in some cases, permanent changes to the environment we live in. This article will highlight some of the implications relating to energy security, and touch on issues of digitalisation and the future of technology in the field of AI/ML. The goal of this article is to raise awareness about the environmental costs of technology and inspire readers to question what sort of future humanity should create for itself in the dawn of a new digitalised era.
Issues of Energy and Environment
The recent United Nations Climate Change Conference (COP26) put into sharp focus the threats and risks posed by climate change. Energy consumption and energy security are intrinsically linked with this. Two of the issues tied to energy and the environment are irreversible change to the environment due to energy consumption, and the knock-on effects that energy insecurity causes. As many of the sessions at COP26 highlighted, some damage that has already been done cannot be repaired [3]. China, India, and the United States are responsible for 70 percent of the world's coal consumption, while the energy sector remains the largest contributor to climate change across the board [4, 5]. Energy insecurity, which is the lack of constant energy supply to a country or region, has a wide range of impacts including economic, physical, and behavioural. These have ripple and cascade effects as communities and societies find ways to cope with adverse situations [6]. The contemporary global economy and society are dependent upon energy like never before as humanity relies on the internet, smart devices and wireless transactions for everything from buying groceries to transportation, communication, and education. Previous warnings that society would collapse within 100 hours of a catastrophic lack of energy were dire [7]. The past ten years have seen a huge leap in dependence upon new tech and an ever more interconnected world.
Sustainability in Industry 4.0
These issues are not to say that technology is bad or that advances in technology should not be pursued. Instead, they need to be pursued with sustainability and the environmental cost in mind. An example comes in the form of what is called the digital rebound effect. This is the unintended reversal of benefits gained through new technology. For example, in the early 2000s, e-commerce was thought to reduce energy consumption, but instead the ease of shopping and home delivery as well as increased packaging negated the benefits of online shopping [8]. Energy consumption is only likely to increase as humanity moves towards the fourth industrial revolution, often referred to as Industry 4.0, which relies heavily on the Internet of Things (IoT) and machine to machine communication [9]. While there are roadmaps for navigating this new revolution, such as the 2030 Vision for Industry 4.0 [10], this vision for the future is grounded on the pillars of autonomy, interoperability, and sustainability. However, not all visions for the future can account for rebound effects or other problems that will only be discovered while implementing new technologies. There is a lack of policy driven critical assessments on the impact of new technologies [11]. Despite promises of efficiency, the digitalisation of industry is unlikely to be sustainable unless there is pressure on governments and consumers for it to be that way. Instead, the increasing pace of technological and digital development is likely to exacerbate these problems even further by providing cheaper goods and new digital services [12].
Green AI
Another offender when it comes to the environmental costs of the future is Artificial Intelligence and Machine Learning (AI/ML). The field of AI/ML has rapidly developed in recent years and has even become quite advanced, although it is far from the types of AI imagined in Sci-Fi classics such as 2001: A Space Odyssey or The Hitchhiker’s Guide to the Galaxy. Much of this progress is a result of exponentially large amounts of computational power being devoted to deep learning processes. Deep learning is a type of automated predictive analysis based on a hierarchical model of increasing complexity [13]. These deep learning processes that rely on extraordinary amounts of energy and computational power are classified as Red AI, because they do so without much regard for the energy used [14]. Green AI on the other hand refers to “AI research that yields novel results without increasing computational cost, and ideally reducing it” [15]. Researchers into the future of AI/ML indicate that the neural networks and models used for AI should not only be measured by a standard of accuracy of results, but also by a standard of efficiency [16, 17]. Within the AI/ML community there are guidelines for best practices towards creating a Green Artificial Intelligence Standard [18], as well as guides for industries and companies to consider taking steps towards Green AI as their industries become increasingly digitised and automated [17, 19]. These attempts are a good step in the right direction towards shaping a sustainable future.
Conclusion
In summary, the cost to the environment through digitalisation and emerging technologies like AI/ML is great, and only likely to increase going forward. There are several possible solutions but as the phenomenon of the rebound effect shows, not all can be compensated for ahead of time. Two actions that can be done immediately are better impact analyses of the implementation of future technology with the measures that are readily available as well as committing to and striving towards sustainable development of new technologies such as AI/ML models. This requires a wider societal conversation about the future of technology and how we think about the environmental costs involved.
Sources
[1] “History of Technology,” Encyclopaedia Britannica, accessed November 15, 2021, https://www.britannica.com/technology/history-of-technology.
[2] “From the Industrial Revolution to Climate Change: What Happened and Why?,” 2041 Foundation, accessed November 15, 2021, https://2041foundation.org/climate-change-what-happened-and-why/.
[3] “Climate change brings irreversible harm to poor countries. At COP26, rich ones face pressure to foot the bill.,” The Washington Post, accessed 15 November 2021, https://www.washingtonpost.com/climate-environment/2021/11/08/climate-change-loss-adaptation-cop26/.
[4] “Coal Consumption by Country,” Worldometer, accessed November 15, 2021, https://www.worldometers.info/coal/coal-consumption-by-country/.
[5] “This Interactive Chart Shows Changes in the World's Top 10 Emitters,” World Resources Institute, accessed November 15, 2021, https://www.wri.org/insights/interactive-chart-shows-changes-worlds-top-10-emitters.
[6] Diana Hernandez, “Understanding ‘energy insecurity’ and why it matters to health,” Social Science and Medicine 167, (August 2015): 5-7.
[7] Jeffrey Belk, “96 hours to the stone age: How quickly our connected lives crumble when the power goes out,” Gigaom, accessed November 15, 2021, https://gigaom.com/2011/11/23/96-hours-to-the-stone-age-how-our-connected-lives-crumble-when-the-power-goes-out/.
[8] Stefanie Kunkel and David Tyfield, “Digitalisation, sustainable industrialisation, and digital rebound – Asking the right questions for a strategic research agenda,” Energy Research & Social Science 82, (2021): 3.
[9] Bernard Marr, “What is Industry 4.0? Here's A Super Easy Explanation For Anyone,” Forbes, accessed November 15, 2021, https://www.forbes.com/sites/bernardmarr/2018/09/02/what-is-industry-4-0-heres-a-super-easy-explanation-for-anyone/?sh=75dd47fa9788.
[10] “2030 Vision for Industry 4.0,” German Federal Ministry for Economic Affairs and Energy, 2019.
[11] Carl-Otto Gensch, Siddharth Prakash, and Inga Hilbert, “Sustainability in a digital world, chapter Is Digitalisation a Driver for Sustainability?,” in Sustainability in a Digital World, ed. Thomas Osburg and Christiane Lohrmann (Cham: Springer, 2017): 128.
[12] Vlad C. Coroamă and Friedemann Mattern, “Digital Rebound – Why Digitalization Will Not Redeem Us Our Environmental Sins,” ETH Zurich, (2019): 5-8.
[13] Roy Schwartz, Jesse Dodge, Noah A. Smith, Oren Etzioni, “Green AI,” Allen Institute for AI, Seattle, WA, (July 2019):1-9.
[14] Ibid.
[15] Ibid.
[16] Ibid.
[17] Oskar Eriksson, “Moving from Red AI to Green AI, Part 1: How to Save the Environment and Reduce Your Hardware Costs,” DataRobot, accessed November 15, 2021, https://www.datarobot.com/blog/how-to-save-the-environment-and-reduce-your-hardware-costs/.
[18] David Dao, “Green AI,” Github, accessed November 15, 2021, https://github.com/daviddao/green-ai.
[19] Oskar Eriksson, “Moving from Red AI to Green AI, Part 2: A Practitioner’s Guide to Efficient Machine Learning,” DataRobot, accessed November 15, 2021, https://www.datarobot.com/blog/moving-from-red-ai-to-green-ai-part-2/.
