Kenya's AI Bill Must Put Workers at the Centre

The AI Bill 2026 risks importing not just good ideas, but also the burdens that come with regulating far more advanced economies.[iStockphoto]

Kenya has taken a bold step by introducing the Artificial Intelligence Bill, 2026, the country’s first attempt at establishing a legal framework for Artificial Intelligence governance. At a time when AI is rapidly transforming economies, workplaces, and societies, this legislative effort signals that Kenya understands the urgency of regulating emerging technologies before they outpace public policy.

The Bill arrives at a critical moment for the country as Kenya is positioning itself as Africa’s digital innovation hub, with increasing investments in AI, data centres, digital labour platforms, automation systems, and technology enabled services. From financial services and healthcare to education, agriculture, logistics, and customer service, AI systems are already influencing decision-making processes and reshaping labour markets.

However, while the AI Bill demonstrates commendable ambition in addressing innovation, transparency, and risk regulation, it falls significantly short in the protection of workers and the inclusion of labour in AI governance. Instructively, the future of AI is inseparable from the future of work and, therefore, any AI law that fails to adequately address labour rights risks deepening inequality, accelerating job insecurity, and undermining social justice.

In considering the memorandums submitted on the AI Bill, the Senate must, first, acknowledge that the numerous concerns raised by trade unions are therefore neither anti-technology nor anti-innovation, but, rather, those that stem from a recognition that technological progress without trade union rights often benefits capital more than labour. Indeed, history has repeatedly shown that industrial transitions, when left unchecked, tend to produce exclusion, exploitation, and social instability.

In Kenya’s growing digital labour ecosystem, gig workers, ride-hailing drivers, delivery workers, content moderators, and data annotators are often managed entirely through algorithms. Worse still, most of these workers may never know how decisions affecting their income are made.

The AI Bill acknowledges some of these risks through provisions on transparency, oversight, and risk classification, but the Bill stops short of translating these concerns into enforceable labour protections for it regulates AI systems without adequately regulating their impact on human labour.

One of the most glaring weaknesses of the Bill is the exclusion of workers and trade unions from formal AI governance structures. For instance, the proposed Advisory Committee includes representatives from government, private sector actors, and civil society, yet it does not mandate representation from workers through their umbrella body, the Central Organization of Trade Unions (Kenya), COTU.

This omission undermines the long-established principle of tripartism, which is central to labour governance both in Kenya and internationally under ILO Convention 144. Tripartism recognizes that governments, employers, and workers must participate equally in decisions affecting labour and economic policy, and, therefore, the exclusion of workers from AI governance effectively means that decisions affecting millions of employees will be made without their direct involvement.

Another major concern is the Bill’s narrow conception of employment, especially because Kenya’s labour market is increasingly characterized by non-standard work arrangements. Indeed, millions of youthful Kenyan workers operate in the informal economy, gig economy, and digital platform sector, yet the Bill largely relies on traditional notions of employment, risking the exclusion of workers whose relationships with employers are ambiguous or deliberately disguised.

This creates the danger of a dual labour regime in which formally employed workers enjoy legal protections while gig workers and digital labourers remain exposed to exploitation.

The AI Bill also fails to establish mandatory consultation and collective bargaining obligations before AI systems are introduced into workplaces, a move that grants employers unilateral power to deploy technologies that fundamentally alter working conditions, wages, and job security.

This is not only a dangerous precedent, as AI deployment should not be treated as a purely managerial decision, but also a move that has the potential to directly affect livelihoods and workplace relations. As such, the Bill must provide that trade unions, led by COTU, must have a formal role in negotiating how AI technologies are implemented, noting that social dialogue is essential if technological transitions are to occur peacefully and equitably.

Equally troubling are the inadequate safeguards against workplace surveillance and data exploitation, considering AI systems rely heavily on data collection, often involving continuous monitoring of worker behaviour, location, productivity, and communications. As such, without strict legal protections, employers could misuse AI systems to intensify surveillance, invade privacy, and erode worker dignity.

Notwithstanding the obligations placed on employers to undertake assessment of workplace impacts, the Bill’s provisions on workforce impact assessments and reskilling are equally too weak and largely procedural, as there are no binding obligations on employers to prevent job losses, compensate displaced workers, or invest in meaningful reskilling programs.

This is particularly concerning given projections that AI could significantly disrupt labour markets globally, and our Kenyan legislators should not wait until mass displacement occurs before developing worker transition mechanisms. If managed properly, AI can create new opportunities for Kenyan youth, but without deliberate planning, the costs of disruption will fall disproportionately on workers, especially youth and those in low-income occupations.

There is therefore an urgent need for stronger amendments to the Bill so that labour rights and decent work principles are given priority and explicitly recognized as core objectives of the law. In doing so, parliament must equally broaden the definition of a ‘worker’ to include gig workers, platform workers, and workers in AI supply chains.

If Kenya truly seeks to become a leader in responsible AI governance, then workers and trade unions must be placed at the centre of the conversation so that technology serves humanity, not replaces its dignity.