Kenya Plans New Workplace Rules for AI Data Annotators and Moderators

Kenya has unveiled a draft policy that would introduce new labour protections for thousands of people working behind artificial intelligence systems, placing greater responsibility on global technology companies and outsourcing firms operating in the country.

Kenya’s proposed AI worker protection policy sets out minimum standards for pay transparency, mental health support and workplace conditions for content moderators, data annotators and AI quality evaluators whose work helps train and improve systems such as ChatGPT, Gemini and Microsoft Copilot.

If adopted, the framework would become one of Africa’s most comprehensive attempts to regulate working conditions across the AI value chain, addressing long-standing concerns about low wages, psychological wellbeing and employer accountability.

The draft policy, published by the Ministry of Information, Communications and the Digital Economy, calls for occupational protection guidelines covering written employment contracts, psychosocial support, grievance procedures and minimum workplace standards.

It also proposes duty-of-care obligations for employers handling AI-related work. Companies would be expected to protect workers from the effects of prolonged exposure to disturbing material, provide access to mental health services and maintain transparent employment practices.

The framework applies to both local AI companies and international firms operating in Kenya through outsourcing partners.

A major proposal is the creation of a fair-pay reference framework that would establish transparent wage benchmarks for data annotation, content moderation and AI quality evaluation roles.

Rather than leaving compensation entirely to outsourcing contracts, the government wants pay levels to be measured against international rates for comparable work. Employers would also be required to disclose how their wage structures compare with those benchmarks through a compliance reporting process.

The draft policy also addresses workplace surveillance, calling for monitoring practices that are proportionate and respectful of workers’ rights while strengthening grievance and redress mechanisms.

Content moderators and data annotators perform work that remains largely invisible to the public but is essential to modern artificial intelligence.

Data annotators label text, images, audio and video so AI models can recognize patterns and improve their responses. Content moderators review material uploaded to digital platforms, removing posts that violate platform rules and helping prevent harmful content from spreading online.

The work often exposes employees to graphic violence, child exploitation, sexual abuse, self-harm, murder and other disturbing material that can have lasting psychological effects without adequate support.

Workers involved in projects linked to OpenAI and Meta through outsourcing company Sama have previously raised concerns about trauma, low pay and sudden job losses. Sama has disputed those allegations.

The policy acknowledges these occupational risks, stating that existing labour protections do not adequately address mental wellbeing, transparency or employer accountability for AI workers.

Kenya has become one of the world’s leading destinations for AI data annotation and content moderation, supported by a large English-speaking workforce and an established business process outsourcing industry.

Global technology companies rely on human reviewers to improve the accuracy and safety of generative AI systems. Human-labelled data helps AI models identify harmful prompts, refine responses and reduce the risk of producing unsafe or misleading content.

Many of these services are delivered through specialist outsourcing firms rather than direct employment by the technology companies whose AI products depend on the work.

That business model has helped position Kenya as an important part of the global AI ecosystem while also raising questions about worker protections, compensation and corporate responsibility.

The proposed framework would require AI companies and outsourcing providers to align their employment practices with locally defined labour standards instead of relying solely on commercial agreements.

If the policy is adopted, employers operating in Kenya would need to demonstrate compliance with requirements covering fair pay, occupational safety, mental health support and transparent contracting.

The proposal also reflects a broader effort by Kenya to strengthen governance around artificial intelligence while supporting the country’s ambition to remain a leading destination for AI investment and digital services.

For global AI companies, the changes could increase compliance obligations in one of the industry’s most important outsourcing markets. For Kenyan workers, the framework offers the prospect of clearer employment standards and stronger protections for work that has become fundamental to the development of modern artificial intelligence.

Download the free Kaspersky SMB Cybersecurity Guide here to learn how businesses can move beyond traditional antivirus and build a more resilient approach to cybersecurity.