The promise that artificial intelligence will usher in an era of better jobs and widespread abundance is facing growing scrutiny from economists and labor researchers who say the transition will be far messier than technology leaders suggest. At the center of the debate is a simple question that has proven surprisingly difficult to answer: if AI takes over existing work, what exactly are the new jobs it creates?
That question was put directly to some of the most powerful figures in technology during a recent television segment, where they were asked to name concrete examples of future employment. The responses were notably thin. One executive pointed to wellness centers and spas, an industry that has existed for thousands of years. Another declined to name any job at all, suggesting instead that work itself would become optional within a generation. The exchange highlighted a gap between confident predictions of a golden age and the absence of specific, credible pathways for displaced workers.
Economists who study technological disruption argue that the vagueness is not a dodge but a historical constant. No one in 1940 could have predicted that their grandchildren would work as podcast producers or user experience researchers. Research from MIT has found that roughly 60 percent of jobs Americans hold today are in occupations that did not exist in 1940. The majority of modern work was simply unimaginable to previous generations. That historical pattern, however, does not guarantee that the transition will be smooth or that the new jobs will arrive quickly enough to absorb those displaced.
Daron Acemoglu, who won the Nobel Prize in economics in 2024, has estimated that AI will add only about one percent to GDP over a decade, describing the impact as nontrivial but modest. He has found no evidence that the technology will reduce inequality and projects that some workers will see their real wages fall. Molly Kinder, who spent three years studying AI and work at the Brookings Institution, calls the reassuring narrative soothing but false. She warns of a long, messy middle period of concentrated job loss for which neither government nor industry has a credible plan.
The core problem, according to these researchers, is a mismatch of speeds. The International Monetary Fund has warned that generative AI can spread much faster than previous disruptive technologies. Meanwhile, the mechanisms meant to help workers adapt, such as retraining programs and the emergence of new industries, move at the same slow human pace they always have. The gap between those two speeds is where the damage occurs, and the people caught in that gap are the ones who pay the price.
History offers repeated examples of what happens when workers are left to navigate such transitions alone. The Luddites of the early nineteenth century are often remembered as opponents of technology, but they were not afraid of machines. They wrote a plan to survive them and were met with soldiers instead. A handloom weaver in 1820 could not have named a single job his trade would eventually become, just as a factory worker in 1940 could not have described a web developer or a logistics coordinator. The new jobs have always been invisible until they arrive.
What concerns economists is not the eventual destination but the journey. The abundance that technology leaders promise might materialize on a long enough timeline. The timeline itself is the problem. When work changes under people's feet, the question is who is actually responsible for carrying them across the gap. The answer, according to the research, is that no one is coming to retrain workers. Not companies, not governments, and not the billionaires building the intelligence that is reshaping entire industries.
For workers in fields like design, product development, and engineering, the practical implication is that adaptation will be largely self-directed. The fear running through these industries is not irrational. It reflects a real recognition that the transition is already underway and that the safety net is thin. The most useful response, researchers suggest, is not to wait for a rescue but to begin building the skills and flexibility that the next economy will require, even if the exact shape of that economy remains impossible to predict.