Is AI Good for the Planet?. Benedetta Brevini

as 17 per cent in early April 2020. But, as the world’s economy started to recover, emissions rebounded; and the UN showed that 2020 only saw a 4–7 per cent decline in carbon dioxide relatively to 2019 (United Nations News 2020). While transportation and industrial activity declined from January 2020, electricity consumption remained constant, which partly explains the minimal drop in emissions. How, you may ask? According to the World Energy Outlook 2019, globally 64 per cent of the global electricity energy mix comes from fossil fuels (coal 38 per cent, gas 23 per cent, oil 3 per cent: IEA 2019). Since fossil fuels are the largest source of greenhouse gas emissions, without fundamental shifts to renewable resources in the global energy production we shall not be able to prevent incalculable loss of life, as the planet becomes uninhabitable.

What is the connection, then, between the climate crisis and the energy used by AI?

      The chapters of this book help readers to answer this question through a discussion of the following themes:

      1 a definition of AI and its promises for the world and the environment;

      2 why data capitalism is crucial to an understanding of AI and who controls and develops AI;

      3 why AI worsens the climate crisis; and

      4 what we can do about it.

      In chapter 1 you will learn about the hype and awe of AI. From the European Union to the United States and China, governments and global consultancies are urgently signing declarations that promise that the effects of AI are comparable to those of previous scientific revolutions, such as steam and electricity. This belief that AI will rescue humanity, solve the climate crisis and reduce the inequalities of capitalism has a crucial effect: it obfuscates the materiality of the infrastructures and devices that are central to AI’s functioning. Setting aside the mythical discourse on AI, chapter 1 aims instead to shed light on definitions of AI and asks you to think about AI in a different, more material way than most of us have done in the past.

In chapter 2 you’ll discover why data capitalism is crucial to the development of AI. You will understand the reasons for AI’s rapid adoption, since 2010, as a result of vast computing resources and oceans of data. From that time, pushed by digital lords of the West such as Google, Facebook, Amazon, Microsoft and Apple (Brevini 2020), AI has been adopted virtually by all businesses and already extends through almost every sector of the economy and society. This chapter also explores the gatekeepers of AI power and imperialism, from the United States to China, and concludes with an examination of lobbying efforts by the most powerful tech giants to set the terms of public debates on AI and to determine policy outcomes (Benkler 2019, p. 161).

In chapter 3 you will discover the environmental costs of AI and its relationship to the climate crisis. The converged communication systems upon which AI relies generate a plethora of environmental problems, starting with energy consumption and emissions, material toxicity, electronic waste and disposal (Brevini and Murdock 2017). As you will realize, AI relies on large amounts of data, since it works with unsustainable energy demands imposed by algorithm training and cloud computing. Finally, while promising to solve the climate crisis, AI companies are marketing their services to coal, oil and gas companies, thus compromising efforts to reduce greenhouse gas emissions and to divest from fossil fuels.

      The Conclusion argues that without challenging the current myths of limitless economic growth and boundless consumerism, without reconsidering the way in which the structures, the violence and the inequality of capitalism work, we won’t be able to achieve the radical change we need if we are to tackle the climate crisis. So, instead of embracing AI as a new utopia that will fix the world and solve the problems created by capitalism, we should start quantifying and reducing the environmental costs and damages of the current acceleration of algorithm-powered AI. After listing a clear set of practices and policy, including practical changes that are easily within our reach, the book argues that abandoning a limited ethics framework and embracing a green agenda for AI that puts the climate crisis centre stage should be our urgent priority.