Data workers in Africa often have a hard time. They face job insecurities – including temporary contracts, low pay, arbitrary dismissal and worker surveillance – and alarming physical and psychological health risks. The consequences of their work can include exhaustion, burnout, mental health strain, chronic stress, vertigo and weakening of eyesight.
Data work includes text prediction, image and video annotation, speech to text validation and content moderation.
The world of data work is built on labour arbitrage – exploiting the fact that workers earn less and have less protection in some countries than in others.
Large technology firms often outsource this work to the global south, including African countries like Kenya, Uganda and Madagascar, and also India and Venezuela. The result is complex production networks that are generally opaque and shrouded in secrecy.
Workers and researchers have issued many warnings about data workers’ health. Despite numerous court cases in multiple jurisdictions, nothing much has been done to address these issues either by tech companies or by regulators.
Still, the news of the death of a Nigerian content moderator, Ladi Anzaki Olubunmi, who was found dead in her apartment in Nairobi, Kenya on 7 March 2025, came as a shock. While the circumstances of her death are still unclear, it has renewed calls for wider systemic change. Her death has sparked condemnation from the Kenyan Union of Gig Workers, which demanded an investigation.
Since 2015, we have been studying the central role of African data workers in building and maintaining artificial intelligence (AI) systems, acting as “data janitors”. Our research found that companies rarely acknowledge the use of human workers in AI value chains, thus they remain “hidden” from the public eye. In other words, the world of AI is built on the toil of human workers most people are unaware of.
In this article, we outline key steps needed to protect these data workers in Africa. They include business process outsourcing regulations, ensuring quality rather than quantity of jobs, and providing social protection. There is also a need to name and shame companies that maltreat data workers.
Data work needs tighter regulation.
Regulation
Business process outsourcing is the practice of procuring various processes or operations from external suppliers or vendors. Firms that do this are sometimes trying to evade local regulations (like minimum wages) and responsibility towards workers’ welfare (via sub-contracting and the use of temporary employment agencies).
This is happening in Africa as some data training firms and digital labour platforms circumvent local labour laws.
But there is more to the story.
Data work is also seen by lawmakers and practitioners as a solution to the rampant unemployment and informality across Africa. African governments have actively created regulatory environments that enable these practices to thrive, despite adverse outcomes for workers.
Nonetheless, new regulations have been proposed lately, like the Kenyan government’s Business Law (Amendment) Bill, 2024 targeting the wider business process outsourcing and IT-enabled services sector. Particularly, it makes business process outsourcing firms responsible for any claim raised by employees. It ensures some accountability for firms bringing data work to Africa.
Other governments should follow with similar measures ensuring worker rights are enforceable. Some data workers are hired on contracts as short as five days and get paid less than the local minimum wage. Firms found violating labour standards should be penalised.
In fact, there is an urgent need to create regional or continent-wide regulatory frameworks covering the business process outsourcing sector, limiting the space for firms to exploit workers.
It’s possible, however, that jobs might be lost as firms relocate to places with favourable laws, an everyday reality in the outsourcing networks.
Quality, not quantity
African governments should prioritise the quality of jobs and not quantity. Policymakers should think about wider national economic development plans, particularly structural diversification and upgrading of their economies.
Historically, these strategies have resulted in success in some states, addressing social and economic issues such as unemployment, poverty and inequality.
Another option for African governments is to enhance social protection among data workers. Financing this is a serious issue, so proper taxation and compliance among workers and employers is urgently needed.
Finally, there is a role for naming and shaming firms that treat their data workers poorly. There is evidence that such efforts improve compliance and firms’ behaviour.
Worker movements
African data workers have taken risks in openly speaking about their experiences. But these kinds of approaches work well when combined with collective bargaining.
Workers have historically won their labour and civil rights after long and hard-fought struggles. There is a long history of African worker movements and trade unions resisting the apartheid and colonial regimes across the continent.
While the freedom of association is enshrined in the African Charter on Human and Peoples’ Rights and most governments have legislation committed to collective bargaining, it is rarely implemented in the new outsourcing sectors, particularly data work.
It is also difficult to organise workers in the industry, because of the high churn rate. For instance, data training firms like Sama offer short-term contracts to employees, often as short as five days.
Some firms are hostile to workers’ organising activities.
But numerous data worker-led associations have emerged in Africa recently, some led by the co-authors of this article. Techworker Community Africa, African Tech Workers Rising, African Content Moderators Unions and Data Labelers Association are among them.
These initiatives are crucial to ensure workers have decent remuneration, work-life balance, adequate working hours, protection against arbitrary dismissal, safe working environments, and contributions towards their health and welfare.
Several high-profile court cases are currently being pursued by African data workers against Meta and Sama. There is precedent. In 2021. Meta was ordered by a Californian court to pay US$85 million to 10,000 content moderators.
AI-dependent tools such as ChatGPT or driverless cars would not exist without African data workers. They are tired of being “hidden”. They deserve to be treated with respect and dignity.
Mophat Okinyi, Kauna Malgwi, Sonia Kgomo and Richard Mathenge co-authored this article.
Mohammad Amir Anwar, Senior Lecturer in African Studies and International Development, University of Edinburgh
This article is republished from The Conversation under a Creative Commons license. Read the original article.
