If AI changes how we earn, who keeps consumers spending?

For most of modern consumer capitalism, the relationship between work and spending has been relatively straightforward: people earn through employment, and that income gives them the purchasing power to participate in the economy. But what happens when artificial intelligence begins to change that relationship?

AI is expected to transform occupations across the global economy. The IMF estimates that around 40% of global employment is exposed to AI, while the World Economic Forum anticipates substantial job creation alongside displacement.

If employment becomes a less reliable source of income, what happens to the consumer demand on which retail depends?

This does not necessarily mean the end of mass consumption. Instead, it raises a sharper question: who will have the money to keep spending, and how might governments step in to protect that purchasing power?

1. The AI income problem

The first complication is that “AI exposure” should not be confused with mass unemployment. Exposure means overlap between AI capabilities and occupational abilities; it does not mean replacement. The IMF estimates that around 60% of jobs in advanced economies will be exposed to AI, but this does not mean 60% will disappear. For some workers, AI may complement existing skills and increase productivity; for others, particularly in highly exposed occupations, the risk of displacement is greater.

The WEF estimates that 170 million jobs could be created while 92 million could be displaced by 2030. However, these figures describe broader structural labour-market change rather than AI alone. Even where new jobs outnumber those lost, transitions may not be painless: workers losing existing roles cannot automatically move into newly created ones.

AI can also affect tasks without eliminating occupations. This could benefit employees by removing repetitive duties while leaving more time for judgement, decision-making and customer interaction.

The more important question may therefore be who captures the gains from AI. Acemoglu highlights the risk that, if AI mainly automates existing tasks, more of the benefits could flow to capital rather than workers. For retail, this matters because a growing economy does not automatically mean stronger consumer spending. That said, this outcome is not inevitable: productivity gains could also create new industries, lower prices and raise incomes.

2. If wages fall, who supports consumption?

If employment becomes a less reliable route to household income, governments may face growing pressure to rethink how purchasing power is protected. This does not necessarily mean a universal basic income.

Governments could instead expand welfare, offer targeted cash transfers or guaranteed-income schemes, subsidise wages, guarantee jobs, or trial universal payments. The appeal of a universal basic income is its simplicity, but simplicity does not necessarily mean affordability. OECD analysis points to a difficult trade-off: a payment large enough to replace lost earnings in a meaningful way would require substantial public funding, while a cheaper payment may offer only limited support. The design of income support therefore matters, especially who receives it and whether it adds to, or replaces, existing benefits.

The link between government income support and retail spending is not just theoretical. Evidence from US stimulus payments suggests that when households receive money from the state, many spend more, especially those on lower incomes or with fewer savings. However, Covid-19 stimulus research found that households spent around 40% of their payments on average, with much of the rest saved or used to repay debt. Extra income does not always turn directly into retail spending. For retailers, the key question is not only how much support governments provide, but who receives it and whether they feel secure enough to spend it.

3. What happens to the consumer?

If government support becomes a bigger part of household income, shopping habits could change even if overall spending remains steady. A modest but reliable income floor may make households more careful about non-essential purchases. This could benefit retailers focused on essentials and value, including supermarkets, own-brand products, discount shops, budget fashion and second-hand goods.

The bigger question is what happens to middle-income shoppers. If lower-income households receive more government support while higher-income households benefit most from AI-related gains, middle-income consumers may feel more uncertain about their wages without having the same safety net or savings. That could put pressure on categories such as fashion, restaurants, electronics, homeware and holidays, where spending can be delayed or cut back more easily.

The result could be a more divided retail market: value-focused consumers at one end and wealthier consumers at the other, while the traditional middle market comes under pressure. AI disruption would not necessarily weaken demand for luxury goods. If AI-related wealth is concentrated among business owners, investors and highly skilled workers, luxury spending could remain strong. The same changes that reduce spending power for some consumers could increase it for others.

Income is not the only factor; time matters too. If AI allows some people to work fewer hours rather than simply replacing jobs, extra leisure time could create demand for experiences, travel, entertainment, fitness and other forms of discretionary spending. AI’s impact on retail may therefore be less about a simple fall in consumption and more about who has money to spend, what they spend it on, and how confident they feel about spending it.

4. The global divide

There will not be a single global retail response if AI changes how people earn money. The impact will partly depend on how well governments can share the benefits of technological change. European countries, which tend to have stronger welfare systems, may be better able to protect household incomes if work patterns change. In the US, where spending power is more closely tied to employment and redistribution is more politically contested, the outcome could look different. Developing economies may face an even bigger challenge: even if AI causes major disruption, governments may have less money available to fund large-scale income support.

This means the countries most affected by AI may not always be the ones best able to share its benefits. For global retailers, the same technological change could lead to very different consumer markets depending on how strong a country’s safety net is and who benefits most from AI-generated wealth.

5. Conclusion

The future of retail in an AI-driven economy may depend less on how many jobs disappear and more on who benefits from the wealth AI creates. AI could change or replace some jobs, but it could also create new ones, improve productivity and generate new economic value. If governments respond by strengthening income support, they could help protect spending power and keep consumer demand steady. However, the impact would depend on how much support is provided and who receives it.

For retailers, this may not mean a simple fall in demand. Instead, the market could become more divided: more focus on value at one end, continued strength in luxury at the other, and new opportunities in experiences and leisure. If employment becomes a less reliable way for people to benefit from economic growth, retailers will need to understand who still has money to spend and how they choose to spend it.