Adrien Matray, in collaboration with Johan Hombert, has produced rigorous economic research that challenges the prevailing optimism surrounding artificial intelligence. Rather than treating AI as an inevitable technological breakthrough, their work examines the underlying economic forces shaping its impact, including labor markets, institutional incentives, and the long-term consequences for workers.
Their comprehensive study of the late 1990s technological revolution offers more than historical insight—it provides a predictive framework for understanding the potential long-term consequences of our current AI enthusiasm. Just as the dot-com era reshaped economic expectations, the AI sector is now experiencing a similar transformative moment that demands nuanced economic analysis.
During the Information and Communication Technology (ICT) boom, the technology sector wasn’t just a financial phenomenon—it was a comprehensive economic ecosystem that fundamentally restructured workforce dynamics. One-third of skilled workers redirected their careers toward this seemingly limitless frontier, drawn by promises of revolutionary technologies and unprecedented economic opportunities.
The initial narrative was seductive: massive capital investments, exponential growth projections, and a pervasive sense that traditional career paths were becoming obsolete. Today’s AI sector follows an almost identical trajectory, with venture capital firms and tech giants investing billions, creating an environment of intense technological experimentation.
However, Hombert and Matray’s research reveals a more sophisticated economic reality. The professionals who joined innovative sectors during peak excitement didn’t just experience a momentary career detour—they faced significant long-term professional consequences. Fifteen years after entering the ICT sector at its height, these workers experienced a substantial 7% wage penalty compared to their peers.
“Periods of rapid technological expansion tend to attract highly skilled workers precisely when the underlying skills are most likely to become obsolete,” said Adrien Matray. “The long-run costs do not come from failure, but from specialization in technologies whose economic value depreciates faster than careers can adjust.”
The mechanism driving this wage depreciation is rooted in the rapid obsolescence of specialized technological skills. In the late 1990s, professionals who built their expertise on specific technological vintages found their specialized knowledge quickly devalued. Web developers specializing in early HTML implementations, IT consultants focused on on-premise CRM systems—these professionals discovered that technological specialization could become a professional liability.
For the current AI sector, this presents a profound challenge. The landscape is characterized by intense capital flows directing resources toward firms pushing technological boundaries. While this might seem like an ideal environment for innovation, it also generates significant risks of skill depreciation.
Consider the current AI landscape: Generative models are evolving at an unprecedented rate. Today’s state-of-the-art neural network architecture could become obsolete within months, not years. Professionals who tie their entire career trajectory to a specific technological approach risk finding themselves professionally marooned.
The most critical economic question isn’t whether AI will be transformative, but how professionals can build resilient, adaptable skill sets in an environment of constant technological flux. The researchers’ work suggests that the most successful professionals won’t be those who chase the latest technological wave, but those who develop transferable skills and maintain cognitive flexibility.
This doesn’t diminish AI’s potential. Instead, it’s a sophisticated call for a more measured approach to technological innovation—one that recognizes the complex interplay between technological advancement, human capital, and long-term economic value.
Companies and educational institutions must redesign professional development frameworks to prioritize meta-learning skills: the ability to learn, unlearn, and rapidly adapt to technological shifts. The future belongs to professionals who can navigate technological change with strategic adaptability.
Investors and policymakers should view the AI boom through a more nuanced lens. Technological enthusiasm must be tempered with historical perspective and a commitment to sustainable human capital development. The sophistication of our algorithms won’t determine the true value of the AI revolution, but rather our ability to create economic structures that support continuous professional evolution.
For individual professionals, the strategy is clear: cultivate a portfolio of skills that transcends any single technological paradigm. Critical thinking, interdisciplinary understanding, ethical reasoning, and the capacity for rapid skill acquisition will be more valuable than deep specialization in any particular AI approach.
The economic ecosystem surrounding AI isn’t just about developing sophisticated algorithms; it’s about creating professional frameworks that allow human talent to remain dynamic and adaptable. We must shift from viewing technological skills as fixed assets to understanding them as fluid, continuously evolving capabilities.
As we stand at this technological crossroads, the lesson from Hombert and Matray’s research is unambiguous: innovation is most powerful when it prioritizes human potential over technological hype. The AI revolution isn’t just about what machines can do, but about how humans can strategically position themselves within an increasingly complex technological ecosystem.
The future belongs not to those who predict technological change most accurately, but to those who remain most adaptable in the face of constant transformation. Our economic success will be determined not by the algorithms we create, but by our ability to remain intellectually nimble, ethically grounded, and perpetually curious.
