Adapted from “The Changing Economics of Knowledge Production,” by Laura Veldkamp from Columbia Business School and the National Bureau of Economic Research, and Simona Abis from the University of Colorado Boulder.
Why the research was done: As new products and technologies have emerged, there has been mounting speculation about the potential of big data and AI to disrupt labor markets. Until recently, empirical data on AI’s influence has been missing from this dialogue. But with this research, CBS Professor Laura Veldkamp and her co-author aim to quantify and assess the true magnitude of AI’s impact on the economy.
“The AI economy is surrounded by hyperbole,” says Laura Veldkamp, the Leon G. Cooperman Professor of Finance at CBS. “People say, ‘It’ll end labor as we know it,’ or ‘Data is the new oil’ — all these catchphrases. But there’s not much measurement of how exactly it’s transforming the economy.”
By focusing on how AI and big data are reshaping labor markets, this study seeks to fill in the gaps and offer a grounded perspective on the real consequences — and opportunities — presented by technological advancement.
How it was done: With no public repository of information about how firms value their data, the researchers had to reach beyond traditional economic analysis to assess the financial sector’s engagement with AI and big data. “Accounting rules generally don’t require — or even encourage — firms to report the value of their data on their balance sheets,” Veldkamp says. “But we realized that firms with a lot of data need to hire a lot of people to work with that data … and we can follow the trail of breadcrumbs.”
Taking an innovative approach, the researchers examined detailed employment trends, including the demand for roles that require data analysis and AI skills. From there, they inferred the underlying value and strategic importance that firms place on data. The researchers also applied theoretical models to align empirical findings with economic theories of profit maximization, which allowed them to deduce how much data is needed for the observed hiring to be profitable.
What the researchers found: Based on their findings, the researchers predicted a 5 percent decline in the labor share of income due to AI and big data technologies. This suggests a fundamental shift in the composition of production inputs, with data becoming a more significant component. The findings also indicate a consequential shift in income distribution, suggesting an intensification of economic disparities as data ownership becomes increasingly lucrative.
This transition comes close to mirroring the historical shifts brought about by the Industrial Revolution, which saw 5 to 15 percent declines in the labor share of income. But interestingly, in the context of AI, the loss of labor share does not equate to a loss of human jobs. “There’s been a lot of talk of AI replacing labor. That didn’t happen in finance. In fact, there was more hiring. Firms that are adopting AI are hiring people with AI skills,” says Veldkamp. “At least in this context, AI wasn’t replacing the people — they just got more work done.”
As such, the findings highlighted an urgent need for workers to acquire new skills pertinent to a data-driven economy. This applies to current employees who will need to upskill or reskill, as well as those on the cusp of entering the workforce, including business school students. For instance, learning to engage with tools like Python and TensorFlow can add tens of thousands of dollars to a worker’s annual income potential.
Why it matters: While AI’s impact is considerable, it’s not without historical precedent. This research suggests that the integration of such technologies is not merely a displacement tool but a productivity enhancer. It can potentially lead to a greater scale of operations and a shift in the labor-capital income share.
This innovative methodology lays the groundwork for broader research in other sectors, providing a model for evaluating the economic contributions of data. “One reason we focused on the financial sector is because it was an early adopter of AI,” explains Veldkamp. “Of course, the finance industry is not a perfect example, but it is the leading indicator — the canary in the coal mine. If things are going really wrong, we might expect to see it here first.”
Furthermore, this research provides empirical evidence that can guide businesses and policymakers in navigating the data-driven revolution in knowledge production. “AI is likely to be transformative for many, perhaps all, sectors of the economy,” says Veldkamp. “I’m pretty optimistic that we will find useful employment for lots of people. But there will be transition costs.”