TL;DR
The trillion-dollar AI industry runs on the psychological suffering of tens of thousands of content moderators and data annotators in the Global South — workers who review child abuse, violence, and hate speech for poverty-level wages with almost no protections. With 710 million Africans still not using the internet despite living in areas with mobile broadband coverage, according to the Broadband Commission’s 2025 report, the digital divide creates the very desperation these companies exploit. Expanding digital access without building worker protections and genuine economic opportunity only expands the exploitation.
Table of Contents
- What Does AI Exploitation Actually Look Like on the Ground?
- Why Can’t These Workers Walk Away?
- How Does the Digital Divide Enable This Exploitation?
- How Do Global Tax Frameworks Make It Worse?
- Can Digital Inclusion Prevent Digital Exploitation?
- What Can You Do About It?
- FAQ
The trillion-dollar artificial intelligence revolution isn’t built on sophisticated algorithms and cutting-edge technology.
It’s built on tens of thousands of workers’ suffering.
This isn’t a "bad boss" situation at one or two companies. This isn’t a glitch in the system. It is the system. In cities like Nairobi, Manila, and Casablanca, in Syrian refugee communities in Lebanon, and among immigrant populations in Germany, an invisible workforce endures daily psychological horror to keep our digital spaces "clean" and our AI models "safe." And as the African Commission’s Resolution 630/2025 warns, the scaling back of content moderation safeguards by major platforms now threatens to make these conditions even more dangerous.
The exploitation is a predictable consequence of the massive digital divide gripping the Global South. When entire populations are systematically denied the tools, infrastructure, and education to participate in the digital economy on their own terms, they become captive labor — forced to accept whatever digital work is available, no matter how traumatic.
What Does AI Exploitation Actually Look Like on the Ground?
It looks like workers forced to watch videos of child abuse, bombings, and murder — hour after hour, with almost no breaks and no meaningful mental health support.
To illustrate, we borrow heavily from an article published by Noema that focused on three workers: Kings, Ranta, and Abrha, all data annotators and content moderators in Nairobi, Kenya. They worked for companies including Sama and Meta, reviewing and moderating content fed into social media algorithms and datasets used to train large language models like ChatGPT.
The content these workers review is horrific — violence, hate speech, child sexual abuse, xenophobia, racism, and domestic violence. Kings, who grew up in Kibera, one of the world’s largest slums, shares experiences too graphic for most to contemplate: "I saw videos of child abuse." As he describes it: "It goes beyond what makes people human. It’s like being forced to drink poison every day, knowing it’s killing you, but you can’t stop because it’s your job."
What Are the Working Conditions Like?
Far bleaker than tech companies would have us believe. Workers at Ranta’s job receive just one hour of break time during their entire shift for all basic needs — eating, stretching, using the bathroom. Under constant surveillance through computer tracking, cameras, and even facial expression monitoring — to ensure they weren’t "too emotional" in response to what they were seeing — they can’t even keep their phones nearby for comfort, forced instead to store them in lockers. Non-disclosure agreements silence workers from speaking out about their conditions, while job insecurity keeps them trapped.
Even basic protections aren’t guaranteed. Sama stopped making health insurance contributions for Ranta while continuing to deduct payments from her paycheck. These aren’t oversights.
Some companies that rely on organizations like Sama, such as Meta, claim to provide "24/7 on-site support with trained practitioners" and "technical solutions to limit exposure to graphic material." The workers’ testimonies tell a different story. As Abrha describes it: "Imagine being expected to watch a video of someone being killed, and then immediately move on to the next post. There’s no time to breathe, let alone process what we’ve seen."
In Abrha’s case, this meant watching videos from the Tigray war — the deadliest modern-day conflict — the very war he himself had just fled. Sometimes he watched videos of bombings in his own hometown.
What Happens to These Workers?
The impacts follow them home. Severe psychological symptoms including nightmares, flashbacks, anxiety, and hypervigilance haunt their lives.
At just 32, Kings struggles with intimacy due to intrusive images from work.
Ranta speaks of losing all faith in humanity.
Abrha turned to chain smoking just to cope with the daily trauma.
One worker fled to the countryside, their paranoia about potential violence making city life unbearable.
Who can blame them?
Why Can’t These Workers Walk Away?
Because the digital divide has left them with no viable alternatives — and the companies exploiting them know it.
While internet access has grown exponentially across Africa — from 4.5 million users in 2000 to 526 million by 2019 — this growth masks a stark reality. According to the Broadband Commission’s 2025 State of Broadband in Africa report, 710 million Africans still don’t use the internet despite living in areas served by mobile broadband. Internet penetration across Africa sits at approximately 39% as of 2023, compared to 87% in Europe — a gap that defines what economic choices are available to hundreds of millions of people.
This digital isolation is compounded by a devastating education gap. African students average just 6 years of schooling compared to 10.6 years in G20 countries. The continent’s entire higher education system includes only 25 universities ranked in the global top 500 for STEM fields, with just eight specializing in engineering and seven in natural sciences. Only South Africa, Egypt, and Uganda are even represented in these rankings. While some countries like Algeria, Mauritius, and South Africa maintain secondary school enrollment rates above 80%, others like Tanzania, Ethiopia, and Madagascar struggle to reach 40%. The gap becomes a chasm at the tertiary level, where African enrollment sits at 15.3% compared to 65% in G20 nations.
How Does This Create a Trap?
The economic consequences are profound and self-reinforcing. Over the past decade, low-skill digital jobs have plummeted from 47.6% to 31.6% of available positions, while higher-skill opportunities have grown. Yet 95% of the African population remains trapped at basic or intermediate digital skill levels, with stagnant growth in highly specialized digital skills. Workers must migrate to connected urban centers for any chance at digital employment — even if that means accepting exploitative conditions.
The financial barriers to self-employment are equally daunting. Only 36% of Africans have access to the infrastructure or personal tools to make digital payments, compared to 75% in G20 nations, and just 45% have basic bank accounts versus 82% in G20 countries. Every form of capital — financial, human, social, physical, and informational — is systematically denied to workers in the Global South.
This systematic disadvantage is further entrenched by what can only be described as coercive inclusion in the digital economy. Workers face direct pressure to accept disadvantageous terms, often tied to access to essential services. The lack of alternatives — driven by poverty and marginalization from formal systems — creates a form of forced participation. Information asymmetries ensure workers often don’t fully understand the risks of their positions until they’re already trapped.
How Does the Digital Divide Enable This Exploitation?
By ensuring that entire populations have no economic bargaining power — while global platforms extract maximum value from their labor, data, and knowledge.
The pattern is clear: platforms extract value from the Global South — through user data, content generation, and local knowledge — while shifting all risks and psychological costs onto vulnerable populations. Across Africa, local hotels and travel agencies lose value to foreign platforms like TripAdvisor that accumulate capital outside the continent. Kenyan farmers are pressured into disadvantageous data-sharing agreements, only to find themselves blacklisted from credit over small defaults. Even traditional knowledge is stripped for profit — consider the Shuar indigenous community in Ecuador, whose centuries of medicinal plant knowledge was digitally captured, transferred to U.S. government systems, and enclosed in private pharmaceutical databases without compensation.
With 98% of African economies falling below the G20 average for digital competence, workers across the continent face an impossible choice: remain in digitally disconnected communities with dwindling economic opportunities, or migrate to urban centers where they must accept whatever digital labor is available, no matter how traumatic or exploitative. The fact that companies like Sama actively recruit workers from across the Global South exceeds opportunism — this is calculated exploitation of a population systematically denied the tools to participate in the digital economy on their own terms.
How Do Global Tax Frameworks Make It Worse?
They systematically favor large tech companies while undermining local economies’ ability to invest in their own digital infrastructure — and 2025 developments haven’t changed the fundamental imbalance.
The OECD’s two-pillar framework sets revenue thresholds so high — $20 billion globally — that most exploitative platforms slip through the cracks. While the African Tax Administration Forum recommends a 20% minimum tax rate and 35% profit allocation to market jurisdictions, the global standard remains at just 15% under the GloBE (Global Anti-Base Erosion) rules. Only 10 African nations even participate in this framework, and they face enormous implementation costs — from technical training and infrastructure development to building monitoring systems and enforcement mechanisms.
While the international community agreed in December 2025 on a way forward for the global minimum tax package, and the U.S. One Big Beautiful Bill Act (OBBBA) signed in July 2025 made targeted changes to international tax rules, the structural disadvantage for African economies persists.
Global tech platforms employ sophisticated tax avoidance strategies through headquarters location optimization, transfer pricing, and intellectual property arrangements. Local firms must comply fully with tax obligations while global platforms minimize their contributions. The result is devastating for small economies: revenue loss from non-resident companies, reduced domestic tax collection, and shrinking public service funding that drives increased public desperation.
Can Workers Organize to Fight Back?
They’re trying — and getting crushed for it. The African Content Moderators Union was formed to unite workers across different African countries, and as of 2025, organizers like Kings continue advocating for mental health resources, fair wages, and ethical AI practices. But companies have moved aggressively to suppress collective action. When workers in Kenya tried to organize their own union, they were told to join the postal and telecommunication union instead — a transparent attempt to dilute their collective power.
The dismissal of Richard Mathenge, the co-founder of Techworker Community Africa, after engaging in union activities sent a clear message about the costs of resistance. Without effective collective bargaining power or regulatory protection, workers find themselves locked into deepening patterns of dependency, unable to challenge the platforms that profit from their trauma.
Can Digital Inclusion Prevent Digital Exploitation?
Yes — but only if access comes with protections, skills, and genuine economic opportunity. Digital inclusion alone isn’t enough.
The evidence for expanding access is compelling. Nigeria saw a 3% increase in labor force participation and a 9% rise in household consumption, with poverty rates declining 7% after just three years of mobile internet coverage. In Tanzania, the impact was even more dramatic — an 8% jump in labor force participation, 4% increases in both wage employment and nonfarm self-employment, and a 10% boost in household consumption. Most importantly, these gains often reach those who need them most: female-headed households, rural communities, and less-educated workers see the largest improvements.
But the content moderation industry shows us what happens when digital opportunities come without safeguards. Workers like Kings, Ranta, and Abrha face severe psychological trauma while platforms profit from their pain. Companies market themselves as engines of economic opportunity while avoiding accountability for worker wellbeing.
This is why organizations like Human-I-T work not just to provide technology and internet access, but to empower people with the skills and knowledge to use these tools for their own advancement — on their own terms. Every refurbished device, every internet connection, every digital skills training session is a step toward a digital economy that serves people, not just corporate interests.
What Can You Do About It?
The question isn’t whether we’re creating a digital underclass — we’re already doing it. The question is whether we have the thoughtfulness and willpower to stop it from expanding.
We at Human-I-T are doing our part, and we need your support.
Join us in making digital inclusion work for everyone. Support organizations like ours so that, as more people come online, they do so with the protections and assistance they need to truly unlock their potential — on their terms. Fill out the technology donation form today and take a step towards closing the digital divide while championing responsible technology practices.
Dive Deeper into the Sources
- Amnesty International: The Human Cost of AI (2024)
- Digital Inclusion and Evolution of African Labor Laws
- Ethiopian Policy Review: Digital Economy Analysis
- AllAfrica: Digital Economy Coverage (2024)
- World Bank: Digital Development Report
- World Bank: Digital Skills and Employment in Africa
- Brookings: Digitalization and Digital Skills Gaps in Africa
FAQ
How does AI exploit workers in the Global South?
Major tech companies outsource content moderation and data annotation to workers in countries like Kenya, the Philippines, and Morocco — where economic desperation and limited digital alternatives force people to accept jobs reviewing horrific content including child abuse, violence, and hate speech. Workers endure constant surveillance, minimal breaks, and severe psychological trauma for a fraction of what companies profit from their labor. Non-disclosure agreements and job insecurity keep them silent.
Why does the digital divide make exploitation worse?
When only about 39% of Africans have internet access and 95% of the population is stuck at basic or intermediate digital skill levels, workers have almost no economic alternatives. They can’t freelance, start digital businesses, or access higher-skill opportunities. Companies like Sama know this — and they actively recruit from these populations because the power imbalance guarantees compliant, low-cost labor.
What is being done to protect AI workers in Africa?
The African Content Moderators Union continues to organize across countries, and the African Commission’s Resolution 630/2025 raised alarms about the human rights impact of platforms scaling back content moderation safeguards. But corporate retaliation against organizers remains aggressive. Meaningful change requires both stronger regulatory protections and expanded digital inclusion that gives workers genuine economic alternatives.
How does Human-I-T help address the digital divide?
Human-I-T tackles digital inequity by refurbishing donated technology and getting it into the hands of income-qualified families, providing low-cost internet access, digital skills training, and tech support. Every device we refurbish diverts e-waste from landfills while giving someone the tools to participate in the digital economy with dignity. Get connected today or donate your old technology to help close the digital divide.
Can digital inclusion alone stop exploitation?
No. Without worker protections, fair tax frameworks, and genuine economic opportunity, expanding internet access can actually expose more people to exploitation. Digital inclusion must be paired with labor safeguards, collective bargaining rights, and policies that ensure tech platforms contribute fairly to the communities they profit from. That’s why Human-I-T’s model goes beyond connectivity — we provide comprehensive digital inclusion that empowers people to use technology on their own terms.
