It is expected that AI productivity increases will vastly transform the U.S. economy. Firms are utilizing AI productivity enhancements to automate repetitive tasks, and research and coding functions have already been implemented. The obvious problem is that when machines perform functions once done by humans, what are the humans supposed to do for income? This increase in AI productivity is measurable across various sectors, as supply chains operate more efficiently, data analysis accelerates, and customer service utilizes automated agents to streamline tasks. Manufacturing, once considered a stable sector of the economy, is increasingly using robotics to reduce labor costs. Professional services are also increasingly displacing workers in medical, legal, and other areas of the service economy to improve output (read: profits) per worker.

This is not a new thing. It has been accelerating since the invention of the fax machine and phone answering devices. The use of AI productivity-enhancing technology is becoming increasingly apparent. But as shown, the shift by corporations to focus on worker productivity is ongoing.
 

Recent corporate statements confirm this shift. At a 2025 financial conference, JPMorgan Chase reported that AI adoption doubled productivity gains in certain operations from 3% to 6%, with some roles seeing efficiency increases of 40% to 50%. Other banks said AI allows them to accomplish more work with the same headcount.

In theory, the promise of AI productivity increases is alluring. While firms can produce more with fewer inputs, humans will have more time to pursue education, leisure, and spend time with their families, increasing overall health and happiness. Again, that is theory, and the subject of today’s commentary.
 

Productivity Set To Surge

The strict definition of “productivity” is the output per unit of input. In other words, if output rises, it should correspond to an increase in employee compensation, as economic demand leads to the production of more products. Since 1947, a correlation has existed between economic output and the 3-month average of the annual rate of change in employee compensation.
 

Between 2004 and the pandemic, annual labor productivity growth averaged just 1.5% per year, significantly below the pace required for sustained real wage improvement. Recent gains measured in 2023 showed a temporary uptick; however, whether this marks a trend driven by AI rather than short-term business cycles remains unclear.

Furthermore, emerging research suggests that AI has the potential to deliver significant productivity improvements. A study of generative AI usage found that average workers using tools like ChatGPT completed tasks 40% faster with higher quality, implying substantial productivity enhancements when AI is integrated into work processes. The Federal Reserve Bank of St. Louis estimated that generative AI contributed a roughly 1.1% boost to aggregate productivity, with individual workers saving multiple hours per week on routine tasks. Lastly, a TIME-published analysis of Anthropic research suggests that AI has the potential to double U.S. labor productivity growth, increasing it by approximately 1.8% if widespread adoption occurs.

These projections also align with broader institutional forecasts. The IMF reports that AI could significantly impact nearly 40% of jobs worldwide, presenting both opportunities and risks for income growth and inequality. Yet, productivity gains alone do not automatically lead to wage increases or employment growth.
 

The Problem

The problem arises when productivity increases without a corresponding demand for labor. AI operates without downtime, 24/7, and does not require traditional wages, benefits, or breaks. If AI performs tasks that previously employed millions of workers, the question of how displaced workers earn income becomes central. Corporate leaders acknowledge this challenge. Federal Reserve Chair Jerome Powell has highlighted the unpredictability of AI’s impact, noting that productivity gains may come with labor market disruptions that current policy tools are ill-equipped to manage.

Historical examples show how technological shifts displace workers in the short term. For instance, during the “Industrial Revolution,” artisans lost jobs to mechanized production. Horse‑drawn carriage drivers disappeared with the advent of automobiles. Yes, workers eventually moved into new fields, but the transition involved hardship and community upheaval. Automation in prior eras often created new kinds of jobs, but the pace and breadth of AI disruption could set this wave apart. Instead of merely replacing manual labor, AI now substitutes for tasks across both blue-collar and white-collar jobs. Research by Oxford economists Carl Frey and Michael Osborne highlighted that many occupations have tasks that are susceptible to automation, and could disappear entirely.

Compounding the challenge, since the late 1970s, productivity gains started diverging from typical worker compensation. According to the Economic Policy Institute, productivity growth far outpaced wage growth for the median worker, signaling that gains from technology and economic expansion have accrued disproportionately to capital owners and high‑skill labor. This productivity-pay gap signals that, even before AI’s full impact arrives, workers were not sharing equitably in productivity-driven prosperity.
 

The pace of technological change means millions of Americans face an uncertain labor market. Young workers entering the workforce find fewer traditional hiring pathways and rising expectations around digital and AI‑related skills. Older workers frequently lack the time or resources to retrain in rapidly shifting skill environments. Across age groups, employers deploying AI experience reduced labor costs and increased productivity, which simultaneously puts pressure on wages and job security.
 

The reality is stark. The economy may grow, but how the gains are distributed will determine whether everyday Americans thrive or struggle. Without structural policy interventions, technological displacement risks widening income inequality and weakening labor market attachment. The promise of more leisure, education, and family time from productivity gains remains theoretical. If workers lack stable incomes, employment opportunities, or bridging support, the rest won’t matter.

But, this is where the “cries for UBI” become most vocal.
 

The Wrong Solution

Legendary investor Howard Marks has described AI’s impact on employment as “terrifying.” He emphasized that work provides purpose and identity beyond mere income. Notably, he stated that “…financial support alone will not replace the psychological and social benefits of employment.” That is a crucially important statement, which we now have the data to support. Universal Basic Income (UBI) is the default proposal to offset the impacts of increased AI productivity. The logic sounds simple enough: “If AI displaces workers, send checks to households to replace lost wages and economic stability returns.”

The problem is that the evidence does not support this conclusion.

Following the pandemic-driven shutdown of the economy, we sent checks to households, which was a form of Universal Basic Income. Many articles espoused the benefits of such an operation, but the results were far less appealing. Surging inflation eroded the benefits of the stimulus and left Americans far worse off than they would have been otherwise. However, other real-time tests have also yielded less than promising outcomes.

We previously discussed one of the UBI experiments, which found predictable results. Short-term relief did not translate into higher employment, improved skills, or long-term income growth. Cash transfers temporarily increased consumption but did not raise productivity, increase labor force participation, or improve economic mobility.

“Participants in the study generally did not use the extra time to seek new or better jobs—even though younger participants were slightly more likely to pursue additional education. There was no clear indication that the participants in the study were more likely to take the risk of starting a new business, although Vivalt points out that there was a significant uptick in “precursors” to entrepreneurialism. Instead, the largest increases were in categories that the researchers termed social and solo leisure activities.”

The Argument magazine also reviewed multiple studies on guaranteed income and reached a similar conclusion. While recipients reported lower stress and higher short-term satisfaction, these gains faded quickly. Employment outcomes showed little improvement, job search intensity declined in several cases, and participation in education and retraining did not rise significantly.

In other words, giving people money without purpose helped much less than promised.

The core flaw in UBI is structural, as it treats income as the problem. Employment is the real issue. Yes, work provides wages, but it also offers skill development, social structure, and a sense of purpose, along with long-term stability. A simple check replaces none of those, and unfortunately, as 2020 shows, when producers realize that checks are being sent, they raise prices to capitalize on it. In other words, an artificial increase in incomes will quickly be absorbed by higher prices (inflation), effectively rendering the UBI useless.

Here is the most critical point.

“An economy cannot function on transfers alone; production must precede consumption. UBI reverses this order.“

Cost also matters. A national UBI program large enough to offset AI-driven displacement would require trillions of dollars annually. Funding such a program would either require higher taxes, debt expansion, or both. While each option will reduce future growth, higher taxes reduce investment incentives, while increased debt raises interest costs and crowds out private capital. Neither path supports long-term prosperity.

UBI also weakens the labor signal. Wages communicate where labor is needed, and training follows opportunity. UBI dulls this signal by separating income from work, and, over time, workforce attachment erodes, skills decay, and reentry into employment becomes increasingly complex. This dynamic showed up repeatedly in pilot programs.

Most importantly, UBI avoids the hard work of reform. It sidesteps education reform, workforce retraining, mobility assistance, and pro-growth labor policy. It accepts displacement as inevitable and permanent. History shows this approach fails, and past technological shifts succeeded because workers moved into new roles. In other words, policy supported adaptation, not withdrawal.

AI productivity gains will demand active solutions, not government gifts. Skill development, apprenticeships, employer-based training, wage insurance, and mobility support. These tools address displacement directly, while UBI does not.

Defaulting to UBI is an admission of policy failure and signals surrender to the disruption rather than managing it. The United States grew prosperous by expanding opportunity, not replacing work with checks. That lesson remains relevant today as AI continues to reshape the economy.

More By This Author:

A Weak Yen, The BoJ, And The Carry Trade
Luxury Slump Bankrupts Saks Fifth Avenue
The Silver Surge: Micro Bubble Or Reasonable Valuation?