The Jevons Paradox: Why Artificial Intelligence Will Not Destroy Jobs

When in 2016 Geoffrey Hinton, one of the fathers of artificial intelligence, declared that within five years radiologists would become obsolete, it seemed an inevitable prediction. Yet, nearly ten years later, the demand for radiologists has not only not dropped, but has reached all-time highs. How is this possible? The answer lies in an economic principle formulated over 150 years ago: the Jevons paradox.

What the Jevons Paradox Is

In the mid-19th century, the British economist William Stanley Jevons observed a counterintuitive phenomenon: technological improvements that made the use of coal more efficient did not reduce its overall consumption, but rather increased it. When a resource becomes cheaper and easier to use, the demand for that resource can explode, revealing previously unmet latent needs. The principle is simple: greater efficiency can generate greater demand, not less. Let’s return to radiologists. When AI systems capable of detecting and classifying hundreds of diseases faster and more accurately than humans were launched, many predicted the end of the profession. Instead, the opposite happened. AI tools have made diagnostic scans faster and cheaper. More scans are performed, and more scans mean a greater need for complex diagnoses and treatment planning by qualified professionals. Radiologists have not disappeared: their role has evolved towards higher value-added activities, while AI manages routine analysis.

The Urban Example: The Paradox of New Roads

The same phenomenon is observed in urban planning. When administrations build new roads or ring roads to relieve city traffic, common intuition would suggest a reduction in traffic. But the reality is different. New road infrastructure makes it more convenient and faster to travel by car. This encourages: People who previously used public transport to switch to cars Residents to take longer trips for work or leisure New real estate developments in previously poorly accessible areas Increased urbanization of peripheral areas Instead of reducing traffic, new roads often contribute to further urbanization and, paradoxically, to even greater congestion in the medium to long term. The latent demand for car mobility, made possible by the new infrastructure, exceeds the additional capacity created.

What This Means for Artificial Intelligence and the Future of Work

When containerization in the 1960s made maritime transport 90% cheaper, some port workers initially lost their jobs. But global trade exploded, creating empires in logistics, distribution, and warehousing. When cloud computing made IT infrastructure ten times cheaper, server administrators did not disappear: they transformed into DevOps engineers and cloud architects, managing infrastructures on scales previously unthinkable. With AI, we should expect similar dynamics. When the cost of drafting legal documents, analyzing medical data, or writing code drastically decreases, the demand for these services will not collapse: it will almost certainly explode, revealing previously unexpressed or latent needs.

Conclusion: Neither Utopia Nor Dystopia

This does not mean that jobs will remain unchanged. Many roles that currently require manual intervention will become roles of supervising “AI agent teams.” Artificial intelligence will first transform repetitive jobs, with little context and that tolerate errors: customer service, data entry, routine paperwork. But even these roles, rather than disappearing, will evolve. A call center operator might become a supervisor of AI agents, focusing on complex cases and customer relationship management. An administrative assistant might move from data entry to coordinating exceptions in business processes. This often involves making boring and repetitive jobs much more interesting and stimulating. The debate on AI and work is polarized between pessimists who predict mass unemployment and optimists who dismiss everything as media exaggeration. Economic history suggests a more realistic middle ground: profound transformation, not destruction. The Jevons paradox reminds us that efficiency does not eliminate demand: it often liberates it, multiplies it, transforms it.

Recommended further reading:

Alcott, B., Jevons’ paradox, Ecological Economics, vol. 54, 2005, pp. 9-21. Autor, D., Why Are There Still So Many Jobs? The History and Future of Workplace Automation, Journal of Economic Perspectives, vol. 29, n. 3, 2015, pp. 3-30. Brynjolfsson, E., McAfee, A., The Second Machine Age: Work, Progress, and Prosperity in a Time of Brilliant Technologies, W.W. Norton & Company, 2014. Y Combinator, What Everyone Gets Wrong About AI and Jobs, YouTube, 2024.