AI pay gains could peak soon, but a Penn professor sees a way out

The rapid advancement of artificial intelligence (AI) has led to significant increases in productivity and wages, but experts warn that this growth might not last indefinitely. Ioana Marinescu, an associate professor at the University of Pennsylvania's School of Social Policy & Practice, along with her colleague Konrad Kording, has presented a thought-provoking model in a recent Brookings Institution paper that outlines a potential trajectory for wages in the face of increasing automation. According to their research, the evolution of wages may resemble a hump-shaped curve: initially rising, leveling off, and ultimately declining as machines take over more cognitive tasks previously performed by humans. Despite this concerning forecast, Marinescu believes that a wage downturn is not a foregone conclusion. She highlights that over 14% of 'intelligence' tasks are already automated, a stark contrast to the 49% of routine cognitive jobs that existed in the late 20th century, which had dwindled to 35% by 2018. The Brookings simulation indicates that wage declines could start when around 37% of intelligence tasks become automated—a milestone that may be reached sooner than anticipated if the pace of AI adoption accelerates. While signs of this shift are not yet widespread, Marinescu points to evidence suggesting that younger, less experienced workers in AI-affected fields may already be facing job displacement. A study from Stanford shows that employment for early-career workers aged 22 to 25 in roles like software development and customer service has dropped by 13% since generative AI emerged, contrasting with the stable or increasing employment of older, less exposed workers. Marinescu warns that the critical indicator of a looming crisis would be a significant decline in the proportion of intelligence jobs across the economy, signaling a shift toward more physical labor. She emphasizes the importance of managing the transition between physical and intelligence sectors, suggesting that they should work in tandem, much like labor and capital. For continued wage growth, both sectors need to complement each other, as unchecked automation could hinder long-term gains. The Brookings paper advocates for a more measured approach to automation, recommending increased investment in physical capital. They propose taxing virtual labor to encourage companies to maintain a balanced co-existence of machine and human production, echoing similar sentiments made by Sen. Bernie Sanders regarding a 'robot tax'. Ultimately, the future of wages hinges on how substitutable AI becomes for human labor. If AI can effectively replace most cognitive tasks, wage growth may stagnate. Conversely, if humans and machines can enhance each other's productivity, the economy may continue to thrive without sidelining workers.