AI economic threats: jobs, concentration, costs, and instability

Five channels of economic risk

First, task automation can alter hiring, wages, and the value of particular skills. Second, firms with control over computing capacity, distribution, or valuable data may gain bargaining power. Third, businesses can become dependent on a provider whose prices or terms later change. Fourth, infrastructure projects can create financing and resource pressures. Fifth, unreliable automated decisions can impose losses on customers, workers, or the public.

The ILO's 2025 exposure study is evidence about tasks and occupations, not a timetable for unemployment. The IEA's 2025 Energy and AI report examines rising electricity needs and infrastructure constraints. These sources address different parts of the problem and should not be combined into a single dramatic prediction. International Labour Organization: Generative AI and Jobs — A Refined Global Index of Occupational Exposure International Energy Agency: Energy and AI — Executive Summary

Why useful technology can still disappoint economically

A product may help users and still cost more to operate than those users will pay. A business may save time on one task while adding review work elsewhere. A market can overestimate how quickly adoption, revenue, or infrastructure utilization will grow. None of these possibilities proves that AI is useless; they are reasons to inspect the economics separately from the demonstrations.

For a small business, consider the cost of changing providers. Can you export your documents and workflows? Could you keep serving customers during an outage? Do you understand which charges grow with use? A cheap pilot can become expensive when volume, integrations, and supervision increase.

Read a forecast with a scenario table

Write down the mechanism, the affected group, the time period, and the assumption that matters most. For example, a forecast of lower service prices depends on savings reaching customers rather than being retained by suppliers. A forecast of mass job displacement depends on adoption and organizational changes as well as technical capability.

Analyze this economic forecast. Separate observed data from assumptions about adoption, prices, demand, and employment. Give a plausible upside case, a downside case, and the evidence that would help distinguish them. Do not assign precise probabilities without a defensible basis.

What to monitor

Useful indicators include actual workflow adoption, quality-adjusted costs, changes in hiring, customer switching options, and the terms offered by major providers. Notice whether a claim describes a local business, an industry, or the whole economy. Effects at one scale do not automatically apply at another.

The practical response is to build flexibility: retain usable records, compare alternatives, and avoid making irreversible operational commitments based on a single forecast. This page is an economic framework, not a recommendation to buy or sell any investment.