Blood and Sweat of AI: Millions of Hidden Workers Behind Every ChatGPT Query

Behind every ChatGPT, Gemini, or Claude answer stands an invisible army of people — hundreds of thousands of annotators and moderators from developing countries, without whom modern AI simply wouldn't work. Their labor is carefully hidden.

Who Stands Behind AI

Large language models don't learn by themselves. They need people — to label data, evaluate the quality of answers, remove toxic content, and teach the model to distinguish a good answer from a bad one. Without this manual work, RLHF — reinforcement learning from human feedback, a key technique that makes models useful — would be impossible. This work is performed by millions of outsourcers through platforms like Scale AI, Remotasks, Appen, and Clickworker. Most are from Kenya, Pakistan, India, Venezuela, the Philippines, and other countries where people can be hired for a few dollars an hour.

• Data labeling — identify objects in images, categorize text, assign labels
• RLHF evaluation — compare two model answers and choose the better one, explaining why
• Moderation — review streams of generated content and remove harmful content
• Red-teaming — intentionally try to break the model to find security vulnerabilities
• Transcription and translation — prepare training datasets for multilingual models

How Much They Earn

Rates vary from $1 to $5 per hour — in countries with low living standards, this is still meager. Most workers have no guarantees whatsoever: neither labor contracts, nor benefits packages, nor protection from sudden contract termination. Today there are tasks — tomorrow there are none.

Scale AI, valued at $13.8 billion, builds its business precisely on this workforce. CEO Alexander Wang openly spoke about creating "the world's largest army of annotators," yet the working conditions of this army do not figure in corporate press releases. Journalists at Time in 2023 discovered that workers in Kenya, hired through contractor Sama to moderate OpenAI content, earned around $2 an hour — and spent hours viewing descriptions of violence, leading to psychological trauma.

"We are the secret that companies prefer not to reveal," says one of the annotators from

Nairobi, simultaneously working for three different platforms.

Why This Is Hidden

AI companies sell a narrative of autonomous systems that learn by themselves. Acknowledging the colossal human labor behind every model destroys this image and raises uncomfortable questions about responsibility. Legally, the scheme is arranged so that there is no direct connection between OpenAI and a specific annotator from Lagos — there is a multi-layered chain of contractors. This allows them to not include workers in official employment statistics and not to take on obligations to comply with labor standards.

Researchers from Oxford Internet Institute called this phenomenon "ghost work" — work that is absolutely necessary for the functioning of modern AI systems, invisible to the end user, and deliberately placed outside the scope of corporate responsibility.

What This Means

The AI boom has created a new class of digital precariat around the world. While investors pour billions into autonomous systems, hundreds of thousands of people at $2–3 per hour daily make these systems possible. Without transparency in data supply chains, the technology sector risks reproducing in the digital economy the same problems long criticized in physical production — with the difference that factories are visible, and data centers are not.