We may live in a revolutionary era of technology, but despite all the benefits, there exists an ethical dilemma: much of our technology depends on the exploitation of people in the Global South, including children. In the Global North, we enjoy smartphones, smart fridges, smart watches, laptops, AI assistants, and so on. But who makes this possible? Who programs artificial intelligence and moderates algorithms? Where do the materials for our batteries come from?

There is a dark side to the digital revolution, largely affecting those in the Global South. Living in modern society means using and relying on technology that comes from underpaid microworkers, many of them children, and child artisan miners in countries like the Democratic Republic of the Congo (DRC). Is the convenience of technology worth such a human cost? Our relationship with machines cannot be separated from our relationship with exploited human labour. There would be no “intelligent” technology without those who create, manage, and supply its foundations.

“Of all the shameful & infamous expeditions
whereby man has preyed upon man…this vile
thing dares to call itself commerce.”

— Roger Casement
The Casement Report (1904)

The Ghost Employees of AI

Despite common perception, artificial intelligence is not entirely automated: there are real humans behind the data labelling, the so-called “microworkers” who are crucial to the functioning of AI and large language models (LLMs). Their name comes from the “micro-tasks” they perform, including the training, validation, and in some cases impersonation of AI. These workers are paid very little given the scale of the AI economy, and many are in fact minors. There are also circumstances where whole families perform data labelling work; both parents and children.

CREDIT: KHADIJA FARAH / TIME
Kenya has become a major hub for the AI workforce.

In 2017, the mean remuneration for those who worked as a part of Amazon’s Mechanical Turk or “MTurk” marketplace was US$1.37 per action, amounting to $3.63 an hour on average, and this has since decreased. As another example, Samasource Impact Sourcing Inc. outsourced work for OpenAI (the company behind ChatGPT), hiring Kenyan workers as moderators; these workers received less than $2 per hour. The role involved interacting with traumatic content including child sexual abuse, bestiality, murder, suicide, torture, self-harm and incest. A worker employed by Sama responsible for analysing text for OpenAI claimed to have developed post-traumatic stress disorder, after reading descriptions of a person engaged in an act of bestiality in front of a minor.

Because these workers are contractors, their industrial protections are limited. Their access to trade union memberships, proper counselling, leave entitlements, and collective bargaining does not functionally exist.

Human Dignity vs. Technology

Cobalt is an essential metal, powering every lithium-ion rechargeable battery. Without it we would not have smartphones, laptops, or other similar devices, and AI data centres also rely on cobalt. The DRC has a very high concentration of cobalt deposits, holding more cobalt in reserve than the rest of the planet combined. However, child labour and other abusive practices taint all cobalt sourced from the Congo. As a result, every device we use tells a story of human exploitation.

The children of the DRC are plagued by poverty. Unable to afford an education, and with their next meal depending upon it, whole families work at the local mine, digging for cobalt, with no protective equipment and old handheld digging tools.

CREDIT: JUNIOR KANNAH / GETTY
A child breaking rocks at a cobalt pit in the Congo.

The story of 15-year-old Elodie, told by Siddharth Kara writing for The Guardian in 2018, exemplifies how the tech industry erodes human dignity. Kara interviewed the young girl as she dug for cobalt: she earns $0.55 a day, has an infant child strapped to her back, and suffers from an infection brought on by late-stage HIV. Later, Elodie would be found dead under a tree, her infant son also deceased, still
strapped to her back. This is the human cost of our technologically-reliant world.

Manipulating the Consumer: Emotional AI

As for consumer protections, laws have struggled to keep pace with technology, leading to an unprecedented erosion of privacy and dignity. A rather concering example is the rise of “emotional AI”, which can both influence consumer behaviour and collect and store large amounts of personal data. Given the scale of the internet, there is a serious lack of transparency, with consumers having little to no idea how their data might be used.

CREDIT: AFFECTIVA
Affectiva is an American software company that claims its
AI can understand human emotions and thoughts based
on analysis of facial and vocal expressions.

Emotional AI refers to a new, emerging technology that aims to interpret and simulate human emotion. It gauges emotions, attention span, and intentions by analysing a person’s writing, images, speech, voice, facial expressions, eye movements, and body language. These machines can sense, capture, process, and interpret data beyond what human senses can match. This data can then be used to further train machine learning algorithms, which can be exploited for the purposes of advertising, surveillance, and more. Once again, consumers and their data are themselves being turned into the commodity.

To make AI more ethical, everyone must be protected. The question is whether profits will take precedence over people.
Will laws enhance consumer dignity, even if they affect advertising revenue? Will regulations of the AI labour force be enacted, even if it leads to soaring production costs? All the tech giants behind AI lack much incentive for regulation, and in fact have plenty reasons to fight against it.

For centuries, global human society has functioned on a simple formula: you work in order to make someone else rich. If you happen to have inherited wealth, great; if you happen to be a microworker in Kenya or an artisan miner in the DRC, bad luck. However, our system is mutating further into one where the technology we use serves as a kind of feudal lord — as economist Yanis Varoufakis calls it, “technofeudalism”. Currently, AI investment is propelling the US economy, and global tech giants like Google,
Amazon, Apple, Nvidia, and Tesla have unprecedented market dominance.

Just as the feudal lords of the 15th century accumulated wealth on the backs of peasant labour, the tech giants behind AI have created wealth on the backs of exploited labour from the Global South.

This article was first published in Keep Left #1.