Who Benefits?

Benedetta Brevini – lecturer in communication and media at the University of Sydney, Australia – reveals that AI demands material infrastructure and produces a massive carbon footprint and e-waste that disproportionately damages marginalized communities. Brevini calls for greater transparency and accountability from AI developers and nations that oversee their operations.

Change the World

The majority of business executives believe they must invest in AI to expand their business. Major consulting firms insist AI will enrich industries, particularly manufacturing and retail, by boosting productivity and reducing labor expenses – but they never mention the human cost.

In all its variety of forms, AI relies on large swathes of land and sea, vast arrays of technology, and greenhouse gas-emitting machines and infrastructures that deplete scarce resources through their production, consumption and disposal.Benedetta Brevini

Consulting firm McKinsey estimates that AI will automate more than 50% of all work by 2055, saving industries trillions of dollars in workers’ pay. Proponents of “technological determinism” insist AI will solve inequalities of race, income and gender. But technology, Brevini insists, is inherently social. The society that embraces a technology, she points out, shapes its uses and development.

In 2018, a World Economic Forum report itemized the areas in which AI would leverage the global response to the climate crisis. It emphasized autonomous vehicles, the energy sector, agriculture and food distribution, climate informatics, disaster response, smart cities, and science generally. It also referenced a “digital Earth” in which Microsoft has invested $50 million over five years.

“Ecomodernists” regard climate change as merely a technological challenge. This belief, Brevini writes scathingly, dovetails nicely with capitalism’s incessant drive to grow the economy.

“Digital Lords”

Brevini cites two types of AI: “narrow” and “general” AI. The former seeks to tackle huge amounts of data in a narrow scope. The latter is a superhuman intelligence. Neural networks comprise the backbone of AI, and within AI itself reside machine learning and deep learning.

Machine learning can improve its own performance without human intervention. Deep learning systems possess many layers and process massive amounts of data. But they don’t always reveal their process for making predictions. The infrastructure in which AI operates includes the cloud, datafication and the Internet of Things (IoT).

In one single day we produce 40 times more bytes of data than there are stars in the observable universe.Benedetta Brevini

The world’s data production in one single day comprises 44 zettabytes – or 44 x 1,000,000,000,000,000,000,000 bytes. Corporate entities disproportionately benefit from data extraction, analysis and monetization. AI’s purpose is to maximize profits, especially in surveillance. Google invented “surveillance capitalism” by scraping data from its servers regarding consumer behavior. AI technology attracts so much capital because it organizes data into a tradeable commodity. Facebook, Netflix, Twitter and other “digital lords” monitor and manipulate human interactions – to great profit – with AI.

The Climate Crisis

Brevini has little good to say about any market use of AI and smart devices. She cites data centers that consume huge amounts of energy and water for cooling, which, in the United States and Australia, stresses water resources. The Utah Data Center, for example, occupies 1.5 million square feet and infringes on local wildlife.

Fully 1.52 billion smart devices existed in 2019. This means 19% of the population own a device made from finite resources that continuously connects to the internet. Smart devices contribute to uberconsumerism, with their built-in obsolescence and apps that facilitate frictionless transactions. Technology serves as a destructive “conveyer belt” for endless product iterations.

Infrastructure that sustains AI devices uses approximately 7% of the world’s energy. Mobile internet users will soon increase from 3.8 billion in 2019 to 5 billion in 2025. The Internet of Things will double its connections by 2025. That will boost global emissions, which will grow from 1.6% of the global carbon footprint in 2007 to 14% by 2040. Streaming services fuel this increase.

Manufacturing smart devices produces “e-waste,” such as mercury, flame retardants and fluorocarbons. In 2019, people generated 53.6 million tons of e-waste, mostly in Asia. Only 17.4% of this waste was recycled.

Amazon and Microsoft falsely claim their carbon footprints are zero. Microsoft runs its operations on 60% renewable energy, while claiming to do so on 100%. Amazon’s Virginia data center relies almost exclusively on coal and natural gas to produce energy.

Tech giants partner with the world’s biggest producers of fossil fuels, such as Shell, BP, Chevron and ExxonMobil. Google created a division of gas and oil, to facilitate faster and more efficient fossil fuel extraction. Amazon promises the sector it can apply machine learning to improve oil prospecting.

Legislation and Regulation

Data centers should run strictly on renewable energy, Brevini insists, and governments must exercise regulatory muscle. Smart devices should carry labels indicating what raw materials they contain, their carbon footprint, and likely recycling options. In a pie-in-the-sky moment, Brevini suggests that antitrust legislation would make platforms into public utilities, thus barring platform builders from owning and exploiting the platform’s content. That seems – to understate enormously – unlikely.

Addressing the power asymmetries that current AI development is already displaying necessitates both structural reform and individual behavior change.Benedetta Brevini

Brevini believes transparency and accountability should frame AI’s contribution to the climate crisis and mitigate its most deleterious effects.

Revelatory and Unrealistic

Benedetta Brevini offers revelatory information about the climatic and economic dangers of AI and smart devices. But she also offers highly unrealistic proposals for mitigating those dangers, which Brevini mostly bases on her distaste for global capitalism and a naive faith in legislatures’ abilities and motivation to rein in madly profitable, de facto tech monopolies. The dissonance between these two aspects of Brevini’s treatise sadly undercuts her valid, perceptive – and downright scary – overview of the potential of AI to undo Earth’s climate and undermine any not-for-profit human interaction. Brevini’s righteous anger at the arrogance and unchecked power of tech and oil giants will resonate with readers everywhere – unless they own stocks in the giants.