Every time you click thumbs up on a ChatGPT response, you participate  in an invisible labor system. Hundreds of times every day, there's traumatizing content that doesn't reach you because someone on the other side of the world made that possible. And as we marvel at how conversational AI has become, we think of computers and white guys in Silicon Valley, not the people in the Global South working long hours for pay that bears almost no relation to the value of what they're producing.

The AI industry does what every extractive industry does: makes the supply chain invisible so you don't have to think about it while you use the product. And the attempts to fill that gap – the compliance checklists, the governance frameworks, the "5 ways to be a responsible AI user" listicles – feel insufficient to us. It’s like they’re written for brands that want to appear responsible, not for founders and CEOs who actually want to be responsible.

We use AI tools in our business and we also donate 1% of every F* The Algorithm! 2.0 Workshop sale to Techworker Community Africa (TCA), an organization that works to protect the rights, dignity, and health of techworkers in Africa. 

Those two facts: using AI and donating to TCA, sit next to each other in our company without a clean resolution, and that's exactly why we're writing this. We want to be honest about the tension we’re all facing when it comes to the ethics of AI. So we invited cofounders Mophat Okinyi and Kings Korodi- who have worked inside these systems and are fighting them from within - into Strong Brand Social Club for a version of this conversation that we think most business and marketing content is actively avoiding.

Who’s Behind Techworker Community Africa?

Mophat Okinyi is an AI and human rights advocate, and the founder and CEO of Techworker Community Africa. He is a former content moderator for OpenAI’s ChatGPT and has first hand knowledge of the human labor and psychological toll behind digital platforms and AI training systems that many businesses rely on everyday. Mophat works to bring visibility to the hidden labor supply chains behind AI and social media platforms while advocating for fair working conditions, mental health protections, and ethical technology practices.

Kings Korodi is a globally-recognized AI governance researcher, labor rights advocate, and AI2030 Global Ambassador representing the Global South. He has deep expertise in AI risk, accountability, and human-centered governance. As Co-Founder and Senior Researcher at Techworker Community Africa, his work focuses on ethical AI labor practices, AI incident accountability, and the governance of transnational AI production systems. 

Together, Mophat and Kings have petitioned the Kenyan Parliament to investigate the working conditions of data labelers and content moderators employed across major tech companies such as Meta, Google, and OpenAI. Through TCA, they have also supported the organizing of thousands of workers across Africa and stood alongside collective actions, including legal efforts led by Meta workers in Nairobi, some of whom are TCA members. They understand the AI supply chain and its human impact from the inside.

Here's what they had to say.

The Part of the AI Story You're Not Seeing

We came into the conversation knowing we'd hear about labor conditions, underpaid workers, and difficult working environments. What we walked away with was a complete picture of an industry that has been deliberately architected so that the people powering it stay invisible, and the people profiting from it never have to look.

Mophat opened with something foundational: Behind every AI model, there are thousands of data workers making sure the model actually functions. Data labelers, data curators, and content moderators – people whose entire job is teaching these systems what's true, what's safe, and what's human. "There is no artificial intelligence without data," he told us. "Data in AI is like blood in human beings, it cannot really work without data, and this data has to be trained by people who have human judgment."

The most visible version of this is content moderation. On social media platforms, the reason you don't encounter terrorist content, child exploitation material, or graphic violence in your feed is because a person watched it first, made a judgment call, and pulled it down before it reached you. That happens every day, at enormous scale, largely in the Global South, and without acknowledgment, while most entrepreneurs think “an algorithm” is doing that. 

Kings then walked us through the layer most people know even less about: reinforcement learning from human feedback, the process that makes AI feel conversational and almost alive. After a model is trained, its outputs are ranked by workers who assess which responses are best, and that ranking is what shapes the model to feel like it actually understands you. "It involves an enormous amount of human judgment," Kings said, "and those workers are often very poorly compensated, at rates that bear no relation to the value that the technology actually generates."

So the full supply chain has three layers: workers who build and clean the training sets, workers who reinforce it by shaping how the model responds to feel more human, and content moderators who are the last line of defense once the tool is live to keep dangerous content from your feed. Kings put it in clear terms: "Think of it like the electrical grid in your home. You flip a switch and the light comes on, but behind that there's a grid, cable workers, all kinds of infrastructure you never see. AI works the same way. It's not magic."

Mophat reminded us that we are part of the invisible labor chain, too. Every time you've rated a ChatGPT response, chosen one answer over another, clicked the thumbs up, you've participated in that same labor system, he said "just like data workers who are being exploited on the other side." You're making these tools smarter and helping these companies generate more revenue, while they charge you to use the product. 

A Harm Reduction Framework for Business Owners Who Can't Opt Out

We put the tension to them: most of our community doesn't align with big tech values, but we also can't realistically walk away from these tools. What does responsible use actually look like in practice?

Mophat's answer was to start with questions. Who built this tool, and under what conditions are those people working? Who is doing the content moderation that keeps the platform functional? Is it automated, or are there humans behind it? "When you start asking those questions," he said, "you naturally go deeper. And once you have answers, you can make more intentional decisions." “If you can ask yourself these questions,” he said, “and know more of the truth about how these tools work…thereafter you can make decisions on how you can reduce the harm itself.”

Kings pushed us on two things. First, language. When you sign up for a new AI product, spend five or ten minutes on their website looking for terms like trust and safety or human in the loop. "A company that is actually proud of its ethical practices will most likely talk about it. They won't be silent about it." 

Next: Data. Commercial AI tools are leaky, the volume of information they hold is enormous, and the ways that data gets used aren't always clear or controllable. Your business data, your clients' data, your proprietary ideas: when you put those into commercial tools, you can lose control of them in ways that benefit competitors or expose you in ways you didn't intend. Never paste confidential client information or internal strategy into a commercial AI tool. “It’s good to do some research. Not all these tools are good…There are some platforms that want to leak people’s data,” Mophat informed us.

The questions to ask before adopting any AI tool:

• Who built this, and under what conditions are those people working?

• Does the company's website use language like "trust and safety" or "human in the loop"? Silence on workers is a signal.

• What data are you feeding into this system, and do you understand where it goes?

• What is this company's documented track record on data, workers, and privacy? That information is in the public domain.

For businesses that need tighter data control, it's worth exploring enterprise versions of tools with explicit data privacy commitments, or open-source models that can be run locally. The goal is to build a practice of intentional evaluation, asking the human questions alongside the efficiency questions.

Is There Ethical AI Use? The Honest Answer.

Someone once replied to an email we wrote and said “there’s no such thing as ethical use of AI.” It’s a valid point since, for example, people in Nairobi are struggling for clean water while a tech company from elsewhere wants to build a data center nearby that competes for that same water.

So we asked Mophat and Kings what they thought. 

Kings stated the question of ethics depends on the perspective you’re viewing from. Big tech will tell you their AI is ethical because they're creating jobs and changing lives. Workers look at those same claims and point out that they’re providing jobs but those jobs don't come with good working conditions, and the growth comes with environmental costs that those communities bear directly. Think of electric vehicles: the narrative says they don't pollute, but the power companies generating the electricity often do. "People tend to focus on the part that benefits them, without seeing the whole circle," he said. All the issues are intertwined. When it comes to ethics Kings finishes, "It's not pure, but [the harm] can be reduced."

Mophat reminded us that AI is produced by people, which means we have control over it. A software engineer can choose to build a tool that generates deepfakes, or build something that genuinely helps people. The ethics of the technology depend on the ethics of the people developing them.  

Now, we’re starting to see that the technology itself isn’t necessarily the problem. It’s the people who don’t put humans first, and the system that rewards unbridled exploitation in the name of profit margin. 

The One Mindset Shift That Changes How You Use These Tools

We asked both of them: if you could give founders and marketers one thing to carry out of this conversation, what would it be?

Mophat didn't hesitate: “Let’s try and not think of AI as a tool but we can start thinking of it as a system with human consequences.” He continued to explain that when you think of AI as a tool you think about efficiency and how this tool can make your work easier. You don't think about people. But when you start thinking of it as a system with human consequences, you start to ask different questions: who built this? Who maintains it? Where are they from? What are their conditions? “The moment that we start thinking about humans first before everything else then we are going to have a better world.”

Kings pushed it further for creative and marketing teams specifically. The narrative that AI will replace humans in copywriting or creative work is simply not true, he said, and founders who've bought into it are solving the wrong problem. "There must always be a human in the loop. That's what AI is actually meant to do [support humans], not the other way around."

Techworker Community Africa: Change We Can Get Behind

Techworker Community Africa advocates for the most invisible people in the tech supply chain from the inside, led by people who have worked those jobs themselves. Their work runs on three pillars: workers' rights advocacy, upskilling and capacity building, and civic education. And the workers themselves drive it all. 

"We always prioritize workers’ experiences.” Mophat said. Kings added, “We call it the ‘North Star’: the human workers. The most invisible and most exploited people in the tech industry," Kings said. “Focusing on their working conditions and their demands we create a powerful and tangible narrative that connects all the other risks.”

They're also working to break the low-wage chain that intermediary companies create. When businesses hire data workers through companies like Sama or Scale AI, a stated rate of $15 per hour can become less than $1 per hour by the time it reaches the worker. TCA builds direct client relationships so the agreed rate goes straight to the people doing the work.

TCA is working with a small team, limited resources, and on the other side of the table, some of the most well-funded and legally protected companies in the history of capitalism. What they've managed to accomplish anyway is worth paying attention to:

• First organization in Africa to petition the Kenyan Parliament look into big tech practices

• First workers in Africa to take Meta to court in Nairobi. Meta moved to Ghana to avoid accountability, and the same legal fight followed them there

• Helped establish the Global Content Moderators Union through UNI Global Union

• Contributed to platform work policy discussions at the ILO in Geneva

• Pushed companies like Sama to introduce Cost of Living Allowance for workers to boost their salaries.

• Educated workers on their rights and provide skill development especially for high school and early college youth to prepare them for jobs.

• Petitioned the Kenyan government when it attempted to change platform labor laws without consulting workers

• Communicate with workers globally to help pressure big tech to treat their workers well

• Over 4,000 workers in their direct network, many of whom have been blacklisted for hire for speaking up

"The resistance against big tech is coming from Nairobi, Africa, where people used to think ‘these are vulnerable communities. These are people who can go for anything in the name of work.’ We are very proud that the resistance is coming directly from Kenya." — Mophat Okinyi

What You Can Do With This

“AI is not a tool, it's a system with human consequences.” That's the mindset shift Mophat asked us to carry out of this conversation.

Most founders using these tools every day have never asked who built them, who maintains them, or what those people are paid. Now you have, and that can change how you move. 

Every time you adopt something new, every time you upgrade a plan, every time you start relying on something you haven't looked at closely, ask: Who built this? Under what conditions? Does this company talk about its workers or go conspicuously quiet on the subject? Those questions, asked consistently, can help cultivate a permanent mindset shift that shapes a more equitable AI industry. 

We've written before about what it looks like to be an activist and a CEO at the same time, and one of the things we keep coming back to: redistribute surplus on purpose. Not as a gesture. As a governance decision, built into how you operate from the start so it scales with you.

Donating to TCA is where we put that into practice, not because we think it resolves the tension of using these tools - it doesn’t - but because it puts real resources behind the people already in the fight: Organizing workers, taking corporations to court, showing up in Geneva, doing the slow unglamorous work of building power for people the industry has deliberately kept invisible.

If your business uses AI, the people TCA represents helped build what you're using, and donating is one concrete way to close that loop. We give 1% of every F* The Algorithm! 2.0 Workshop sale, and we'd love for you to join us. You can do it directly via their donation button. That's what specific, funded, micro-activism looks like from where we stand, pointed directly at people already doing the work.

 

---

 

Strong Brand Social donates 1% of every F* The Algorithm! 2.0 Workshop sale to Techworker Community Africa. To learn more or donate directly, visit techworkercommunityafrica.com.

Thank you to Mophat Okinyi and Kings Korodi for your time, your work, and your willingness to bring this conversation to our community.

 

< Back To Blog