Human Labour
The hidden workforce behind AI: ghost work, content moderation trauma, data labelling exploitation, and the growing worker organising movement.
Image: Leo Lau & Digit / Better Images of AI / Knowledge Sweatshop / CC-BY 4.0
Articles
Teaching AI Ethics: Human Labour
Hidden beneath the rhetoric of the jobs AI will destroy is an unseen narrative of the jobs it currently requires to function. Ghost work powers the AI industry at the cost of exploited workers.
Teaching AI Ethics 2026: Human Labour
The human cost of AI has become one of the most pressing ethical issues. Major lawsuits, union formation, and investigative journalism have brought the exploitation of workers into public view.
Case Studies
OpenAI and Sama: The Workers Who Made ChatGPT "Safe"
TIME magazine revealed that OpenAI contracted Kenyan workers through Sama to label toxic content for ChatGPT safety systems. Workers earned $1.32 to $2 per hour reading horrific material, despite OpenAI believing they were paying $12.50 per hour.
Scale AI and Remotasks: The Billionaire's Sweatshop
Scale AI, valued at $13.8 billion, scored just 1 out of 10 on fair labour practices. Its subsidiary Remotasks abruptly shut down operations in Kenya and Nigeria, stranding thousands of workers.
Meta, TikTok, and Content Moderation in Kenya
Daniel Motaung, a Facebook content moderator in Nairobi, became a key whistleblower after developing severe PTSD. Over 140 Kenyan content moderators were diagnosed with severe PTSD, and the case against Meta proceeded to trial in 2024.
OpenAI's Kenyan Data Labelling Operation
OpenAI contracted Sama to send tens of thousands of text snippets containing extremely harmful content to Kenyan workers earning $1.32 to $2 per hour, to build a safety mechanism for GPT-3.
Teaching Ideas
Analyse how the language of "automation" and "intelligence" obscures human labour. Compare corporate press releases with investigative journalism.
Map how AI reproduces colonial patterns of extraction. Examine push and pull factors in AI labour outsourcing.
Examine the Meta/Kenya lawsuits and jurisdictional questions for multinational accountability.
Analyse Scale AI's business model and the true cost of AI.
Research PTSD and secondary trauma in content moderation work.
Audit AI systems to identify where human labour is involved.
Draw parallels to textile industry exploitation and historical labour movements.
Examine what responsibilities consumers have for the conditions in which products are made.
Is the gig economy model sustainable? Examine how platform-based work distributes risk and reward.
How do tech companies externalise costs? Analyse how labour, environmental, and social costs are shifted to workers and communities.
What is the true cost of AI? Investigate the hidden human and material expenses behind AI products.
What responsibilities do consumers have for the conditions in which products are made? Consider consumer ethics in AI supply chains.
How should democratic societies regulate AI labour? Explore legislative approaches to protecting AI workers.
What role can citizens play in demanding ethical AI? Investigate grassroots advocacy and digital rights movements.
What human labour is hidden behind AI systems? Audit the invisible workforce that powers machine learning.
How could AI be developed more ethically? Propose frameworks for responsible AI development.
What trade-offs exist between AI capability and worker welfare? Examine how performance gains may come at human cost.
Whose voices are missing from the AI narrative? Examine representation gaps in how AI stories are told.
What persuasive techniques do tech companies use to deflect responsibility? Analyse corporate rhetoric around AI labour.
How does AI reproduce colonial patterns of extraction? Analyse how resource flows mirror historical exploitation.
What are the push and pull factors that make certain countries attractive for AI labour outsourcing? Map the geography of digital labour markets.
How does geographic arbitrage enable exploitation? Examine how wage disparities across borders are leveraged by AI companies.
How does AI labour compare to other historical forms of outsourced and exploited work? Draw parallels across eras.
What can historical labour movements teach us about current struggles? Apply lessons from past worker organising to AI labour.
How have previous technological revolutions affected workers? Compare the impact of industrialisation with AI automation.
Can multinational corporations be held accountable in the countries where they cause harm? Examine jurisdictional challenges in AI labour cases.
How do labour laws apply to gig workers? Investigate the legal status of platform workers and AI labourers.
What legal protections should exist for AI workers? Propose frameworks for safeguarding those who train and moderate AI.
What are the psychological effects of repeated exposure to traumatic content? Research the mental health impact on content moderators.
How should employers protect worker mental health? Explore duty-of-care obligations for companies using human moderators.
What support systems should be available for workers in harmful occupations? Investigate best practices for psychological support in AI labour.
Can knowledge production be separated from its conditions of production? Examine how the labour behind AI shapes the knowledge it generates.
What ethical obligations do knowledge-users have to knowledge-producers? Explore the responsibility of AI users toward hidden workers.
How does the invisibility of AI labour shape our understanding of AI? Investigate how concealed human work distorts perceptions of machine intelligence.