NEW YORK, June 26 (IPS) – Artificial intelligence (AI) promises remarkable gains in productivity, science, medicine and education. But it is also poised to wipe out millions of jobs, hollow out the middle class, and drain the tax revenues that pay for hospitals, schools and pensions. The process has already begun, and the time to act is running out.
The International Monetary Fund (IMF) estimates that AI will affect almost 40% of jobs worldwide. In advanced economies, around 60% of jobs are exposed and as many as one in three (33%) human jobs are at high risk of being replaced by AI. In emerging markets, about 40% are exposed, with roughly one in four (24%) at high displacement risk; and in low-income countries, an estimated 26%, with close to one in five (18%) human jobs lost to AI.
Job losses shrink the middle class
The most exposed jobs include many occupations long seen as the backbone of middle-class stability: clerical work, customer service, translation, journalism, legal support, financial analysis, marketing content, and even parts of software and data work. These jobs support middle-class incomes, consumer demand and, ultimately, tax-paying households, yet many are among those the IMF finds most exposed to AI.
New jobs will appear but, according to the IMF, far more are likely to vanish. The effects spread beyond the workers who lose their jobs. Wages fall, insecure work multiplies, and bargaining power collapses once employers can credibly threaten to swap workers for AI. More income flows to those who own the technology and to a handful of dominant firms, while the share reaching ordinary employees and workers shrinks.
Middle-class households are the economy’s main consumers. If their incomes fall, shops and small businesses sell less, investment slows, and closures rise. The economy can then slip into a low-growth trap of weak demand, low wages and chronic underemployment.
Falling tax revenues weaken the welfare state
The pressure then moves to public finances. Much of governments’ funding depends on the middle class: income taxes, consumption taxes and social security contributions. If wage income falls and stable employment shrinks, public revenues shrink with it. At the same time, more people need unemployment support, retraining, healthcare and income assistance. Governments then face the fiscal vise of lower revenue and higher need, a risk highlighted in the IMF’s 2026 analysis of AI, labor markets and public policy.
Public pension systems rely on pay-as-you-go financing, where current workers fund retirees. In health, healthy people finance those who are sick. If the pool of contributors shrinks, sustainability collapses; then governments tend to cut benefits, raise charges or shift more costs onto households, as explained in the UNRISD article AI and the Future of the Social Contract.
Public services and democracy come under strain
History suggests what often comes next: austerity policies. Governments under pressure raise consumption taxes, increase user fees, tighten eligibility rules and cut public spending. When revenues weaken, education, health, care services and social protection are often treated as budget lines to be “rationalized,” even though they are human rights and indispensable public services that hold societies together. The result is a two-tier world: quality private services for the wealthy few and failing public provision for everyone else.
Economic insecurity erodes democratic trust. If people feel that work no longer provides stability, that public institutions no longer protect them, and that the gains from technology flow upward to a small elite, resentment grows. Polarization intensifies. Scapegoating becomes easier, as does the appeal of surveillance, manipulation and more authoritarian forms of control, especially when AI itself can be used to shape information and public debate.
The future is ours to shape
None of this is inevitable. As Nobel laureates Acemoglu and Johnson argue, the impact of AI depends far less on the technology than on the political and economic choices we make about how to use it. Governments can tax the windfall profits and concentrated power AI creates. With these funds, they can protect demand and guarantee income security through the transition. Governments can and should expand public services and social security as fundamental human rights. States should also give workers and citizens a real say in how AI is deployed, and regulate AI to strengthen democracy, prevent disinformation and surveillance from eroding civic trust before it is damaged beyond repair.
AI is already transforming society. The decisive question is whether democracies can ensure that its enormous gains are shared widely enough to foster prosperity for all, preserving the social contract on which stable, dignified societies depend. That choice is still ours, but not for much longer.
Isabel Ortiz, Director, Global Social Justice, was Director at the International Labor Organization (ILO) and UNICEF, and a senior official at the UN and the Asian Development Bank.
Bill Shoulder is an AI software engineer and a researcher, with a background in artificial intelligence and international project management.
IPS UN Bureau
© Inter Press Service (20260626164156) — All Rights Reserved. Original source: Inter Press Service
