Is AI Good for the Planet?

Benedetta Brevini

• Publisher: Polity
• Year: 2021
• Print ISBN: 9781509547951, 9781509547944
• eBook ISBN: 9781509547968
• Edition: 1
• Topic: Computer Science
• Subtopic: Artificial Intelligence (AI) & Semantics
• Index: Computer Science

Artificial intelligence (AI) is presented as a solution to the greatest challenges of our time, from global pandemics and chronic diseases to cybersecurity threats and the climate crisis. But AI also contributes to the climate crisis by running on technology that depletes scarce resources and by relying on data centres that demand excessive energy use.

Is AI Good for the Planet? brings the climate crisis to the centre of debates around AI, exposing its environmental costs and forcing us to reconsider our understanding of the technology. It reveals why we should no longer ignore the environmental problems generated by AI. Embracing a green agenda for AI that puts the climate crisis at centre stage is our urgent priority.

Engaging and passionately written, this book is essential reading for scholars and students of AI, environmental studies, politics, and media studies and for anyone interested in the connections between technology and the environment.

ONE
Defining AI
Beyond the Hype

The promises of artificial intelligence

If you read and watch the news regularly, you are certainly familiar with the hype and awe surrounding artificial intelligence (AI). Quite likely you’ll have heard of AlphaGo, the AI-powered Go player? In May 2017, the world proclaimed that a ‘Godlike’ AI player defeated the then world champion, Chinese teenager Ke Jie. The Google-designed DeepMind device was programed for playing the ancient Chinese board game Go.

No human had ever played Go better than Ke Jie, until he was beaten by this US AI-powered machine. The sensationalism that followed the match, as over 280 million people watched it live in China, was unprecedented. Commentators described the match as a ‘Sputnik moment’¹ for the development of AI in China, a moment that accelerated investment in AI in order for China to overtake the United States (Lee 2018). AlphaGo used deep learning essentially to teach itself to play, on the basis of millions of Go positions and moves from games played by humans. This was the first, but certainly not the last time commentators have used space metaphors such as ‘Sputnik moment’ to describe the impact of AI.

If AI can beat humans at Go, perhaps it can save us from the world’s biggest challenges – all the range from treating global diseases to averting the climate crisis?

In 2016, President Barack Obama’s Executive Office published two reports that outlined a comprehensive national plan for AI in the United States. The reports were set as a priority for the administration; one of them was prepared by the new National Science and Technology Council. Both called for key stakeholder involvement not only from defence and intelligence services but also from the departments of commerce, treasury, transportation, energy and labour. The reports, produced, remarkably, in only six months, reaffirmed the enormous potential of AI. On the topic, President Obama employed another space metaphor to convey the idea that the stakes could not be higher: ‘The analogy that we still use when it comes to a great technology achievement, even 50 years later, is a moonshot. And somebody reminded me that the space program was half a percent of GDP. That doesn’t sound like a lot, but in today’s dollars that would be $80 billion that we would be spending annually … on AI’ (Barack Obama, as quoted in Dadich 2016).

‘As the handmaiden of neoliberalism, AI has consistently been hailed as the magic wand to rescue the global capitalist system from its dramatic failures.’

The following year, emphasizing the enormous promise of AI and galvanized by its enormous potential for surveillance and repression, the Chinese government released one of the most far-reaching plans for investment in AI. Released in 2017, this three-step program is working towards the goal of making China a world leader in AI by 2030, with an industry worth US$150 billion. The Chinese government outlined strategies designed to support and promote the increased adoption of AI in various industries, in the military and within smart cities, and invested US$2.1 billion alone in an AI-focused technology research park. Complementing all this, the 2019 Beijing AI Principles were developed shortly afterwards, within a multistakeholder coalition (see ‘In the struggle for AI supremacy, China will prevail’, 2018; Roberts et al. 2019).

Besides the United States and China, many countries in the northern hemisphere have invested heavily in funding for AI technologies and intellectual property. France, Israel, the United Kingdom, South Korea and Japan have all joined the race for AI (Cognilytica 2020).

Although Europe is hardly considered a leader in AI developments, it, too, has invested significantly in AI technologies and employs the same enthusiastic rhetoric. In April 2018, the European Commission (EC) presented the ‘Declaration of Cooperation on AI’ signed by all 28 Member States of the EU including Norway. On 7 December 2018 the EC published a coordinated action plan on the development of AI in the European Union (European Commission 2018a, 2018b). It pledged to increase its annual investments in AI by 70 per cent, under the research and innovation program Horizon, in order to reach €1.5 billion for the period 2018–2020.

In evocative terms, the EC report emphasized the revolutionary character of AI, declaring the latter comparable in this respect to other scientific transformations, such as the steam and the electricity revolution. ‘Like the steam engine or electricity in the past, AI is transforming our world, our society and our industry. Growth in computing power, availability of data and progress in algorithms have turned AI into one of the most strategic technologies of the 21st century’ (European Commission 2018b). Similarly, the Organisation for Economic Co-operation and Development (OECD), in a document it released before adopting the OECD Principles on Artificial Intelligence, stressed that AI

contributes to better lives and helps people make better predictions and more informed decisions. These technologies, however, are still in their infancy, and there remains much promise for AI to address global challenges and promote innovation and growth. As AI’s impacts permeate our societies, its transformational power must be put at the service of people and the planet. (OECD 2019)

In June 2019, the G20 adopted human-centred AI principles that draw from OECD’s AI principles, envisaging major benefits to society. Such benefits are supposed to include solutions to the world’s inequalities: ‘The benefits brought by the responsible use of AI can improve the work environment and quality of life, and create potential for realizing a human-centered future society with opportunities for everyone, including women and girls as well as vulnerable groups’ (OECD 2019).

It’s impossible not to see, in this utopian discourse employed globally, the same rhetoric of the technocrats of the 1990s (Shirky 2008), who argued that the new communicative opportunities provided by the internet would usher in a new era for democracy and freedom (Gilder 2000; Negroponte 1998) and the end of history (Fukuyama 1992). The same ideological discourse, replicated in current techno-enthusiast claims about the cloud (Nye 1994), was more recently debunked by Mosco (2014) in his book To the Cloud: Big Data in a Turbulent World. In pure Enlightenment tradition, this absolute faith in technology, embraced and supported by cybertarians, Silicon Valley circles, global consultancies and politicians (Dyer-Witheford 1999; Brevini and Swiatek 2020), turns into a powerful apology for the status quo and for the current structure of capitalism, without leaving any real space for critique.

Global consultancies promise: Automating industry, boosting profits

This unfailing rhetoric of optimism is amplified by the major global consulting companies’ forecasts. This is unsurprising. Potential applications for machine-learning systems and automation are vast. However, in the realm of retail and manufacturing, these will come at the expense of workers, clients and contractors. The same rhetoric is also echoed by consultancies and service providers that work alongside the corporate and government policy sector. Rao and Verweij (2017) described AI as a transformative force for the global economy with potential impacts on productivity, GDP and economic gains. According to them, AI could increase product variety, personalization, attractiveness and affordability and by 2030 would derive 45% of economic gains from the stimulation of consumer demand.

Maximizing profits through AI: From the United States to China, excluding the developing world

As noted above, global consultancies tend to cast AI in a positive light, while downplaying any negative consequences. Companies tend to emphasize that AI will deliver massive revenues, essentially in three major ways: by boosting productivity and efficiency gains; by reducing labour costs (this is especially true for manufacturing and transport); and by raising the production of AI-powered services that in turn will encourage uberconsumerism. According to a team at the McKinsey Institute, supply chain management is one of four business operations earmarked for presenting the greatest potential for productivity gains through AI, the others being manufacturing, risk management and product development (Bughin et al. 2018). This assessment expresses the same expectations as the OECD, which in its Digital Economy report describes how AI will significantly cut costs and optimize the use of production factors and the consumption of resources in every sector of the economy (OECD 2020). The report ‘AI: Built to Scale’, by Accenture (2019), also demonstrated that this is clearly the same positive sentiment of most technology CEOs: 84 per cent of business executives believe that they need to use AI to achieve their growth objectives.

This belief in AI as a means to solve problems is also reflected in the proliferation of scientific and scholarly papers devoted to this subject. According to a 2020 report by the World Intellectual Property Organization (WIPO), the number of scientific papers in the field of AI has soared since 2000, and this was followed by a spike in patent applications between 2013 and 2016 (World Intellectual Property Organization 2020). The phenomenon points to a shift from theory to application, as AI seeps into production and learning through its artificial neural networks. The WIPO report found that the most popular patents were in the fields of medicine, transport, telecommunications and interactive personal devices, while growth areas included finance, agriculture, e-government and smart cities.

Another trend highlighted by the WIPO is the growth in patents registered by China, which in 2014 recorded its highest volume of first-patent filings. In the same year China, the United States and Japan, together, accounted for 78 per cent of all patent filings. This is in stark contrast to the situation that prevailed earlier in the century, when European countries dominated patent registrations, accounting for nearly one in five filings between 2000 and 2015 (World Intellectual Property Organization 2020).

Global consultancies emphasize AI’s benefits not only for particular stakeholders, but for the world more broadly. They foster the notion that AI provides ‘global benefits’ that can be spread across the world; of course, the United States and China were expected to gain the most. Meanwhile, the ability of what is called ‘the global South’ to access these benefits remains less clear. In chapter 2 I will sketch out how the biggest corporations of AI serve to illustrate this point, and I will do so by considering the massive advantage that Silicon Valley and the Chinese digital giants have over newly emerging start-ups. The absence of infrastructure crucial to AI development is most profound in the poorest Latin American and African countries. While the lack of resources locks these countries out of AI gains, their economies have the most to lose from AI’s uptake. They have populous workforces of low-income manufacturers and farmers who are most easily displaced by automation. Both China and the United States have announced massive investments in AI in the global South, in a way that some may argue resembles a neocolonial approach.

Why techno-solutionism will not fix the world: AI as ‘mover’ of capital accumulation

Despite all AI’s predicted benefits to society, it is clear that its primary purpose is to maximize profits. As another McKinsey report states bluntly, 51 per cent of all the work done in the US economy could be automated, saving companies a sum of US$2.7 trillion, which represents an equivalent loss in workers’ salaries (McKinsey & Company 2017). Moreover, the report predicts that AI could automate roughly a half of all the work globally by 2055. The COVID-19 pandemic has seemingly accelerated the process towards automation. History shows that, since 1990, every recession has been followed by a recovery that offered fewer jobs for the population affected by it. This tim...

Table of contents

• Cover
• Series Page
• Title Page
• Copyright
• Acknowledgements
• Introduction
• 1 Defining AI: Beyond the Hype
• 2 Controlling AI: Understanding Data Capitalism
• 3 Why AI Worsens the Climate Crisis
• Conclusion: AI and the Climate Crisis
• References
• Index
• End User License Agreement