The AI pay boost could peak sooner than you think, professor says

According to a recent analysis by Ioana Marinescu, an associate professor at the University of Pennsylvania, the acceleration of AI could soon lead to a peak in wage growth for many workers. In her study published through the Brookings Institution, Marinescu highlights the concept of 'intelligence saturation,' a phenomenon where the initial boost in productivity from AI tools may give way to stagnation or even declines in wages. Currently, over 14% of cognitive tasks have been automated, a trend that has been steadily increasing since the 1980s. This shift raises concerns about the future of wage growth, particularly as automation encroaches upon jobs traditionally held by humans. Marinescu's findings suggest that once approximately 37% of cognitive tasks are automated, the potential for wage growth may reverse, leading to a decrease rather than an increase in earnings. Marinescu notes that the early indicators of this shift could manifest as a slowdown in wage growth and a transition of employment from cognitive-heavy roles to those requiring more physical labor. Her research indicates a significant drop in routine cognitive jobs, from 49% in the late 1970s to just 35% in 2018, which she interprets as a clear signal of rising automation in intelligence-based work. A critical aspect of this dynamic is the assumption that automation will initially target the simplest tasks within the intelligence sector, such as basic translation or marketing writing, which can be easily performed by AI. If these roles are phased out without the emergence of new high-value positions, the overall number of jobs in the intelligence sector could decline, potentially leading to a broader wage downturn. The speed of this transition is closely connected to how intertwined physical labor and cognitive tasks are within the economy. As AI continues to replace more types of human work, Marinescu warns that wage losses could accelerate. She emphasizes the importance of monitoring the balance between physical and intelligence sector employment, as any significant shift could indicate that we are approaching a peak in the AI pay curve.