AI’s Wage Boost: A Looming Reversal

The relentless march of artificial intelligence has largely been met with a chorus of optimism, a vision of technological progress inevitably leading to enhanced productivity, novel industries, and an overall uplift in human prosperity. 1

For many, the prevailing narrative sketches a future where AI acts as a powerful co-pilot, amplifying human capabilities and driving wages ever higher. 2

Yet, a disquieting counter-narrative emerges from the halls of academia, specifically from Ioana Marinescu, an associate professor at the University of Pennsylvania’s School of Social Policy & Practice.

Her assessment is not merely a caution; it is a stark warning that the AI-fueled wage boost many have come to expect might not only be nearing its apex but could soon reverse course, ushering in an era of wage suppression.

Marinescu, whose insights are detailed in a new Brookings Institution paper, suggests that the economic landscape is rapidly approaching a critical juncture, a state she terms “intelligence saturation.” 3

Her meticulously developed model indicates that once approximately 37% of cognitive tasks are automated, the upward trajectory of wages could begin to waver and then decline.

This isn’t a distant, theoretical possibility; Marinescu’s research contends that we are already significantly into this process, with over 14% of “intelligence” tasks having already been automated.

This figure, she explains, is derived from the substantial decline in routine cognitive jobs since the late 1970s and 80s, painting a picture that places us far from the starting line and unsettlingly closer to a pivotal turning point.

The professor’s framework introduces an “AI pay curve.” 4

In its nascent stages, AI functions as a powerful enhancer, making workers more efficient and productive, which naturally tends to elevate wages.

This is the collaborative sweet spot, where human intellect and machine processing create synergy.

However, this upward slope is not infinite.

After an initial surge, wages are projected to flatten, and then, as automation permeates deeper into cognitive domains, they could ultimately decline.

The critical shift occurs when AI transitions from augmenting human labor to outright displacing it.

It’s a nuanced but profoundly important distinction, marking the passage from shared prosperity to potential economic vulnerability.

So, what are the harbingers of this impending peak?

Marinescu pinpoints two crucial indicators.

Firstly, a discernible slowdown in wage growth across sectors heavily reliant on intelligence. 5

The robust salary increases that have become characteristic of tech-driven economies might begin to plateau or even recede.

Secondly, and perhaps more tellingly, a significant migration of employment away from traditional intelligence-sector jobs towards more physical, embodied roles.

One can envision a future where the demand for software developers, translators, or marketing writers—professions once considered bastions of intellectual endeavor—starts to diminish, not because the work itself vanishes, but because AI assumes an ever-larger share.

Should these job losses not be adequately compensated by the emergence of new, high-value intelligence roles, the overall proportion of cognitive employment could shrink, laying the groundwork for a broader wage downturn.

The model operates on the logical premise that automation will first target the “easiest” intelligence tasks – those involving pattern recognition, data processing, translation, or basic content generation, where AI can quickly match or surpass human capabilities.

Marinescu explicitly mentions “translators or basic marketing copywriting” as professions likely to experience this early decline.

This thought is particularly disquieting for countless individuals whose livelihoods depend on these very skills.

The pace of this transformation, she adds, hinges on the degree of substitutability between physical and intelligence sectors, and even within the intelligence sector itself.

The more readily AI can replicate a human task, the swifter the potential wage erosion.

If intelligence and physical labor become increasingly interchangeable in the eyes of the market, the inherent value of human labor in both domains could be diminished.

This is far more than a mere academic exercise; it represents a critical forecast that demands our immediate and serious attention.

The prevailing narrative of AI as an unalloyed good for the labor market requires a rigorous re-evaluation.

If Marinescu’s projections prove accurate, we could be facing a future where the very technologies celebrated for their potential to elevate humanity instead exert a downward pressure on the earnings of a substantial portion of the workforce.

The implications for economic inequality, social cohesion, and the fundamental structure of work are immense.

What happens when a significant segment of the population finds their intellectual capital devalued, compelling them into lower-paying physical jobs, or, worse, into prolonged unemployment?

The societal contract, long predicated on the idea of progress leading to widespread prosperity, would face an unprecedented test.

The single most vital metric, Marinescu emphasizes, is the evolving ratio of workers in physical versus intelligence-sector jobs.

This serves as a potent barometer for the health of our economy in the AI age.

Are we moving towards a future where human intelligence is increasingly marginalized, or are we creatively forging new domains where uniquely human cognitive skills remain indispensable and highly valued?

The answer will determine whether the AI pay curve continues its gentle ascent, flattens into a plateau of stagnation, or descends into a valley of wage deflation.

This is not a challenge to defer until tomorrow; it is a complex problem that demands foresight, proactive policy-making, and a willingness to confront uncomfortable truths today, before the peak is undeniably behind us.

The future of work, and indeed, the future of human prosperity, hangs precariously in the balance.