Artificial intelligence will not only threaten formal jobs. In a high-informality economy, it will unsettle livelihoods, earnings and access to work—unless social protection is redesigned now.
Vincent Dania
The first sign of artificial intelligence disruption in Nigeria will not be a robot in a factory. It will be quieter than that: a bank relying more heavily on chatbots and automated back-office checks; a logistics platform changing how riders are ranked, paid or deactivated; a small business using generative AI to do the work once assigned to an entry-level assistant; a freelance designer or writer competing with cheaper AI-assisted output; a call-centre worker discovering that “first-line support” has become a software function.
This is not a prediction of mass unemployment by tomorrow morning. That kind of panic is usually unhelpful. The more serious risk is that AI will change the terms of work before Nigeria’s institutions are ready: fewer routine tasks, more unstable earnings, greater pressure on young workers, and algorithmic systems that decide who gets opportunities without clear accountability.
For a country where many people already survive outside stable payroll employment, this matters. Nigeria’s social protection system cannot wait until AI disruption becomes a full-blown labour crisis. It must prepare now.
The central policy question is not simply, “How many jobs will AI destroy?” It is more precise: which tasks are exposed, which workers can adapt, which households will absorb the losses, and what public system will protect people while the economy adjusts?
Research on automation offers an important warning. Daron Acemoğlu and Pascual Restrepo argue that automation can displace labour from tasks, while new tasks can also reinstate labour demand if institutions and markets create opportunities for workers to move into them. In other words, technology does not produce one automatic outcome. It creates a contest between displacement and adaptation. Debra Howcroft and Phil Taylor make a related point: the future of work is socially shaped, not technologically predetermined. Policy choices, business incentives, labour institutions and social relations influence how automation affects workers.
That distinction is crucial for Nigeria. The country does not need to resist AI. It needs to govern the transition.
AI disruption will look different in Nigeria
Much of the global debate imagines AI disruption through the lens of advanced economies: large firms, formal contracts, unemployment insurance, measurable layoffs and retraining systems. Nigeria’s labour market is different. It includes formal professionals, civil servants and bank staff, but also traders, artisans, transport workers, self-employed graduates, domestic workers, smallholder farmers, freelancers and platform workers. Many earn income without formal contracts. Many are “working” but still economically insecure.
That means AI will not show up first as formal job loss. It will appear as falling commissions, weaker bargaining power, lower demand for routine services, or exclusion from digital work systems. A worker can lose income without being officially unemployed. A platform worker can be penalised by an opaque rating system without knowing why. A young graduate can find that the entry-level tasks that once provided a first step into employment are now automated or outsourced to AI-assisted tools.
The exposure is not limited to low-skilled work. Recent research on perceived vulnerability to technological change among higher-educated workers found strong concern in finance and information technology—sectors closely exposed to AI. Banking is especially relevant. Studies of AI adoption in banking point to the use of automation in customer service, fraud detection, risk management, personalised banking and repetitive process automation. These systems will improve efficiency and customer experience, but they also change the demand for routine clerical, support and back-office roles.
Nigeria should therefore avoid two mistakes. The first is complacency: assuming AI is a distant problem because the country still faces electricity, infrastructure and industrialisation challenges. Cloud-based AI does not require a fully automated factory. It can enter through phones, apps, banking platforms, customer-service systems, recruitment tools, logistics software and office workflows.
The second mistake is fatalism: assuming technology will automatically wipe out work. Evidence from developing-country contexts cautions against simple forecasts. Christian Parschau and Jostein Hauge, examining automation in South Africa’s apparel industry, show why country-specific and industry-specific adoption barriers matter. Not every forecast of job loss materialises in the same way everywhere. But this is not a reason to ignore risk. It is a reason to build better local intelligence.
Nigeria needs to know where AI is likely to substitute tasks, where it can augment workers, and where it may deepen existing inequalities.
The missing link: social protection for labour-market transition
Social protection is often discussed in Nigeria as a poverty programme or emergency response tool. That is too narrow for the AI era. A modern social protection system must also help people manage transitions: between jobs, between sectors, between old skills and new skills, between informal income and more secure livelihoods.
This is especially important because AI disruption will interact with existing vulnerabilities. Young people entering the labour market will face fewer routine starter roles. Women concentrated in low-paid service or informal work will face new forms of exclusion if digital systems reproduce bias. Informal workers will struggle to prove income loss. Platform workers will be controlled by algorithmic management but classified in ways that deny them labour protections. Small firms will adopt AI unevenly, leaving some workers more productive and others displaced.
This is why conventional employment policy is insufficient. Training programmes alone will not protect households from income shocks. Cash transfers alone will not help workers move into better tasks. Social insurance tied only to formal payrolls will miss many of the people most exposed. Nigeria needs adaptive social protection: a system that can identify emerging risks, enrol people before crisis, deliver support quickly, protect income during adjustment and connect workers to credible opportunities.
What preparation should look like
First, Nigeria needs an AI labour-market risk observatory. This should not be a symbolic committee. It should be a serious data and policy function involving the National Bureau of Statistics, labour-market institutions, universities, employers, unions, platform companies and civil society. Its job would be to map task exposure by occupation and sector, monitor changes in hiring and earnings, study platform work, and identify groups at risk before distress becomes visible. The goal is not to produce alarmist headlines, but to guide policy.
Second, social protection must become portable. In a high-informality economy, benefits should follow the worker, not only the job. Health coverage, pension contributions, injury protection, maternity support, unemployment-related assistance and skills support should be designed so that informal and platform workers can participate. This will require flexible contributions, public subsidies for low-income workers, and mechanisms that allow platforms and firms benefiting from flexible labour to contribute to worker protection.
Third, Nigeria should pilot income support linked to reskilling and transition. Workers cannot retrain if training means hunger. If AI reduces demand for certain routine tasks, affected workers need temporary earnings support, credible training, job-matching assistance and pathways into sectors where human skills remain valuable. These should not be generic “learn coding” schemes. The skills agenda must include digital literacy, AI tool use, care work, technical trades, creative problem-solving, communication, entrepreneurship support and sector-specific upgrading.
Fourth, algorithmic management should be regulated before abuse becomes normal. Workers affected by automated decisions should have rights to notice, explanation and human review. If a platform deactivates a worker, changes pay formulas or ranks workers through automated systems, there should be a transparent grievance process.
Fifth, public incentives for AI adoption should come with social conditions. If firms receive tax incentives, public contracts, innovation grants or regulatory support for digital transformation, they should be expected to produce worker-transition plans. These plans should include consultation, retraining, redeployment options and responsible data practices. Public money should not subsidise silent displacement.
Sixth, Nigeria must strengthen its social registry and delivery systems without turning them into tools of exclusion. Adaptive social protection depends on data, but data systems can harm people if they are inaccurate, opaque or punitive. Workers should be able to update information, contest errors and understand how eligibility decisions are made. Digital delivery should expand access, not punish those without smartphones, stable connectivity or formal documentation.
AI policy is social policy
Nigeria’s AI conversation is often framed around innovation, start-ups, productivity and competitiveness. Those are important. But AI policy is also social policy. A country cannot pursue digital transformation while leaving workers to manage disruption alone.
The politics of this transition may become difficult. Labour-market risk can shape public trust, social solidarity and support for reform. If people experience AI as another force that enriches the already secure while making everyone else more disposable, resistance will grow. If, however, workers see that the state has systems to protect incomes, support learning and regulate unfair digital power, technological change becomes less threatening and more legitimate.
The choice is not between innovation and protection. The countries that manage AI well will be those that connect innovation policy to labour-market institutions and social protection. Productivity gains must be accompanied by transition support. Efficiency must not become a polite word for insecurity.
Nigeria still has time to prepare, but not unlimited time. AI tools are spreading faster than many regulatory systems can respond. The danger is not only that some jobs disappear. It is that livelihoods become more volatile in a country where millions already live close to economic stress, and where formal safety nets do not yet reach everyone who needs protection.
A serious government would treat AI disruption as a foreseeable shock. Not a crisis to be improvised after households have fallen into poverty, but a transition to be planned for in advance.
The next labour-market shock may be written in code. Nigeria’s task is to ensure it does not become another poverty shock. That requires a social protection system built not only for yesterday’s vulnerabilities, but for tomorrow’s work.
Vincent Dania is a PhD researcher on AI and labour market risks in high-informality economies at the Institute of Social Policy and Strategic Studies, Nnamdi Azikiwe University, Awka.
