From Silicon Valley to Nairobi Call Centers – How We Design Policy Now Will Decide Who Wins and Who Loses

Global Inequality: This article explores how artificial intelligence (AI) is transforming work, wages, and inequality worldwide. Drawing on recent research from the International Monetary Fund (IMF), International Labour Organization (ILO), OECD, World Bank, and WTO, it examines how AI affects different types of jobs, who is most exposed in advanced and developing economies, and how generative AI could either widen or reduce global inequality.

The article summarizes key findings on job exposure, labour-demand shifts, and gendered impacts, presents a comparative policy table, and highlights critical voices on “AI colonialism”, worker surveillance, and algorithmic bias. It concludes with policy proposals—skills, social protection, regulation, and global governance—to ensure AI supports a fairer world of work, and suggests further readings in APA style.

1. Two Competing Stories About AI and Jobs

When people talk about AI and work, two narratives collide:

• Doom story: AI will automate millions of jobs, throw workers into permanent unemployment, and turbocharge inequality.
• Hope story: AI will boost productivity, create new occupations, and free humans from drudgery—if we manage the transition.

Recent evidence suggests both stories are partly true, depending on who you are and where you live.

The IMF estimates that about 60% of jobs in advanced economies and around 40% globally are “exposed” to AI, meaning that key tasks could be either automated or augmented. In Latin America and the Caribbean, a World Bank analysis finds 26–38% of jobs exposed to generative AI, with only 2–5% at risk of full automation and a much larger share likely to be transformed rather than eliminated.

The OECD’s Employment Outlook 2023 describes AI as a general-purpose technology that will spread across sectors, changing tasks in most occupations while fully replacing relatively few—at least in the short term.

That nuance is crucial:

AI is not just a “job killer” or “job creator”; it is a task re-shaper whose impact on inequality depends on the rules we write now.

2. How AI Is Changing Work – The Evidence So Far

2.1 Exposure vs. Automation

A useful distinction in the latest research is between:

• Exposure – jobs where AI could perform some tasks.
• Full automation risk – jobs where AI could perform most tasks, potentially replacing the role.

The World Bank’s 2024 and 2025 work on generative AI finds that in Latin America, 30–40% of jobs are exposed, but only a small fraction—2–5%—face a high risk of full automation with current technology.

An IMF staff note on “Gen-AI and the Future of Work” similarly concludes that in advanced economies, many jobs are highly exposed but likely to be augmented (more productive workers using AI tools) rather than simply replaced.

The OECD’s surveys across manufacturing and finance show that workers using AI report higher job satisfaction and sometimes better wages, but they also face higher work intensity, privacy concerns, and fear of job loss.

For background and data dashboards, see:

• IMF blog on AI and the global economy: https://www.imf.org/en/blogs/articles/2024/01/14/ai-will-transform-the-global-economy-lets-make-sure-it-benefits-humanity
• OECD AI & labour markets hub: https://oecd.ai/en/work-innovation-productivity-skills/key-themes/labour-markets

2.2 New Jobs, New Tasks

Across studies, economists agree that AI will create as well as destroy jobs:

• ILO notes that AI may generate entire new occupations in areas such as data curation, AI safety, and human–machine collaboration.
• Early World Bank evidence from job postings suggests AI is already changing labour demand, with some white-collar listings declining in regions like South Asia.
• A 2025 research paper synthesizing global findings concludes that AI will likely reallocate labour across tasks and sectors rather than simply cause net job loss in the short run, though distributional impacts could be large.

The key question for justice is therefore not “will there be jobs?” but “who gets the good jobs, and who bears the cost of transition?”

3. Inequality Within and Between Countries

3.1 Within Countries: High-Skill vs. Low-Skill, Men vs. Women

The IMF’s 2025 working paper on AI adoption and inequality shows that in advanced economies, AI tends to complement high-income, high-skill workers, raising their productivity and wages, while middle-skill routine workers are more at risk of displacement.

ILO and UN work on the “AI divide” stresses that existing inequalities often map directly onto AI impacts:

• Workers with higher education and digital skills are more likely to use AI as a productivity tool.
• Lower-skilled workers may see tasks automated without access to retraining.

Gender also matters. A 2025 study on Africa’s outsourcing sector warns that women’s tasks are, on average, 10% more vulnerable to automation than men’s, and up to 40% of tasks in low-paying roles could be automated by 2030 without targeted support.

3.2 Between Countries: The Emerging AI Divide

Globally, institutions are increasingly worried about a North–South AI divide:

• The IMF notes that advanced economies are more exposed and better prepared, with stronger digital infrastructure and more cognitive jobs; developing countries may be slower to see productivity gains but still face disruption.
• The WTO warns that AI could increase global trade by nearly 40% by 2040, but income gains could be 14% in rich countries vs. 8% in poorer ones, unless digital infrastructure gaps are closed.
• IMF Managing Director Kristalina Georgieva has argued that most countries lack the regulatory and ethical foundation for AI, and low-income countries especially lag on infrastructure, skills, and innovation.

In short: AI could become a powerful engine of convergence or a new driver of divergence, depending on whether access to data, compute, and skills is broadened or hoarded.

4. Three Ways AI Changes Inequality

To understand the justice question, it helps to break AI’s impact into three channels.

Table 1. AI and Inequality: Three Key Channels

The net effect on justice depends on which channel dominates and how policy responds.

5. Scholars, Researchers, and Critics: Key Debates

5.1 Mainstream Policy View: Regulate, Reskill, Protect

Major institutions (IMF, ILO, OECD, World Bank) converge on a broad agenda:

1. Skills and lifelong learning – Massive investment in digital and cognitive skills so workers can complement AI rather than be replaced by it.
2. Stronger social protection – Unemployment insurance, income support, and active labour-market policies to manage displacement.
3. AI regulation – Rules on transparency, bias, worker rights, and data protection, such as the EU’s AI Act and emerging OECD–GPAI guidelines.

The ILO’s “Mind the AI Divide” initiative explicitly calls for a global perspective on the future of work, emphasizing that governance must “put people at the center” of digital transformation.

5.2 Critical Perspectives: AI Colonialism, Surveillance, and Power

More critical scholars and activists raise deeper structural concerns:

• “AI colonialism”: Data and compute are concentrated in a few companies and countries, while the Global South often provides raw data and outsourced labour with limited control or benefit.
• Surveillance and “bossware”: AI-enabled monitoring of workers can intensify workloads and harm mental health. A UK report on “bossware” found increased stress and deteriorating wellbeing where intrusive monitoring was deployed.
• Gender and racial bias: Algorithmic hiring and performance tools can encode biases, locking marginalized groups out of high-quality jobs while automating discrimination.

These critics argue that justice requires not only social cushions, but power shifts: worker participation in AI design, stronger unions, data rights, and global rules that curb excessive concentration.

6. Policy for a Fair AI Future: Four Pillars

6.1 Skills and Education as a Global Public Good

• Make digital literacy and basic AI concepts part of core education worldwide, not a luxury subject.
• Invest in TVET and short-cycle training for workers in exposed professions, especially in developing countries.
• Support international initiatives on AI and skills through UNESCO, ILO, and regional bodies so curricula don’t diverge wildly.

World Bank and ILO data show that workers with stronger skills are more likely to use AI as a complement, not a competitor.

6.2 Social Protection and Just Transition

• Expand unemployment insurance, cash transfers, and public works to support those displaced by automation.
• Design “just transition” funds for sectors at high risk (e.g., back-office services, certain logistics roles) similar to those used in climate policy.
• Encourage private-sector contributions to reskilling funds when AI deployments significantly reduce headcount.

The IMF and ILO repeatedly stress that income-support systems and active labour-market policies are essential to prevent AI shocks from turning into long-term poverty traps.

6.3 Labour Rights, Regulation, and Worker Voice

• Update labour law to cover algorithmic management, platform work, and remote monitoring.
• Guarantee transparency and human oversight in AI-driven hiring, firing, and performance evaluation.
• Protect the right to organize and bargain collectively over AI deployment, workload, and data use.

The OECD notes that, without clear rules, AI can make work more intense and less humane, even when productivity rises.

6.4 Global Governance and the AI Divide

• Use WTO, IMF, World Bank, and UN forums to address cross-border data flows, taxation of digital giants, and access to compute and chips.
• Support open and regional AI infrastructure (e.g., African or Latin American cloud and compute initiatives) to reduce dependence on a handful of providers.
• Create global funding mechanisms for AI capacity-building in low- and middle-income countries, modeled on global health or climate funds.

Without such measures, WTO modelling suggests that AI could boost trade but leave poorer countries with smaller gains, deepening global divides.

7. The Future of Work Is a Political Choice

AI is not destiny. The best available evidence shows that:

• Many jobs will change; relatively few will vanish overnight.
• Inequality will rise or fall depending on who controls AI, who gets the skills, and how we protect people during transitions.
• The biggest risks are not only job losses, but power imbalances—between firms and workers, and between rich and poor countries.

If governments, employers, unions, and citizens treat AI as a shared project—with clear rules, investments in people, and genuine global cooperation—then AI could help build more inclusive, safer, and less monotonous work. If not, it risks becoming another wave of technocratic disruption that widens gaps and erodes trust.

The core question is therefore less technical than ethical:
Do we want AI to maximize short-term efficiency, or to expand human dignity and opportunity?

The answer will be written not in code, but in policy.

Suggested Further Readings:

International Labour Organization. (2024). Mind the AI divide: Shaping a global perspective on the future of work. ILO.
https://www.ilo.org/publications/major-publications/mind-ai-divide-shaping-global-perspective-future-work

International Monetary Fund. (2024). Gen-AI: Artificial intelligence and the future of work (Staff Discussion Note). IMF.
https://www.imf.org/en/publications/staff-discussion-notes/issues/2024/01/14/gen-ai-artificial-intelligence-and-the-future-of-work-542379

OECD. (2023). OECD employment outlook 2023: Artificial intelligence and the labour market. OECD Publishing.
https://doi.org/10.1787/08785bba-en

World Bank. (2024). Generative AI and jobs in Latin America and the Caribbean. World Bank.
https://www.worldbank.org/en/topic/poverty/publication/generative-ai-and-jobs-in-lac

World Trade Organization. (2025). AI, trade and inequality: Seizing opportunities, managing risks (flagship report). WTO.