I still remember the quiet shock I felt when the latest cost-of-living adjustment for Social Security came through. Nearly 72 million Americans will see a 3.6 percent raise next January. On the surface it looks like good news for retirees. Dig a little deeper and the picture turns unsettling fast. Benefits have already climbed 26 percent since 2021, the steepest five-year jump since the early 1980s. And the forces driving those increases show no sign of easing. Energy prices keep climbing, producer costs remain elevated, and something far more disruptive is quietly rearranging the entire funding base of the program: artificial intelligence.
The Quiet Math That Could Break the System
Social Security does not run on magic. It runs almost entirely on a 12.4 percent payroll tax split between workers and employers. That single revenue stream supplies more than nine out of every ten dollars the program receives. The old trust-fund surplus and the interest it once earned are nearly gone. As long as wages grow faster than prices, the system can limp along. Sustained inflation above 4 percent, however, starts to crack the foundation. History already showed us how this plays out.
Between 1970 and 1981 wages more than doubled while consumer prices rose 136 percent and benefits tripled. The trust funds ran deficits every year after the first automatic COLA arrived in 1975. Congress eventually had to raise taxes. Even then the program came within months of being unable to mail checks. Today the starting point looks worse. The generation that once paid the bills is now collecting them, and the workers who were supposed to replace them face a new competitor that never sleeps, never asks for a raise, and improves every month.
White-Collar Jobs Sit Directly in the Crosshairs
Look at the occupations most exposed to current AI systems: managers, business and finance staff, programmers, engineers, scientists, lawyers, writers, designers, sales roles, and office support. Together these jobs account for roughly 42 percent of American employment and 52 percent of all wages. That is about $5.7 trillion in annual pay, with an average salary around $86,500. These same workers also carry a disproportionate share of the federal tax burden. The top tenth of earners pay roughly 70 percent of all income taxes, and the bulk of that money comes from white-collar paychecks.
When companies face rising input costs, they look for ways to protect margins. Labor is often the largest controllable expense. AI tools that can draft reports, analyze data, write code, answer customer queries, or generate marketing copy become attractive almost overnight. Outplacement data already shows AI cited as the leading reason for announced layoffs for several consecutive months. More than 100,000 positions have been openly linked to AI replacement in the first half of the year alone, double the previous full-year total. Those numbers will grow.
I have spoken with mid-level managers who already use AI daily to handle tasks that once required junior staff. The efficiency gains are real. The headcount reductions follow. Entry-level roles feel the pressure first, but the technology climbs the ladder quickly. Some industry leaders have publicly estimated that half of all entry-level white-collar positions could disappear within five years, pushing unemployment into double digits. Whether the exact figure lands at 15 million jobs by the end of 2028 or somewhat higher, the direction is clear.
Inflation Keeps the Cost Side Climbing
While the revenue base faces pressure from fewer high-wage paychecks, the expense side is being pushed higher by persistent inflation. Producer prices have run well above the Federal Reserve’s 2 percent target for years. Energy costs, especially diesel and crude, have surged. Even after stripping out food and energy, core producer prices remain elevated. Those higher costs take months to work through supply chains and eventually appear in the consumer prices that determine the next Social Security COLA.
By the time the 2028 adjustment is calculated, annual increases of 6 percent or more look entirely plausible. Each extra percentage point of COLA adds roughly $16 billion a year to the program’s permanent cost base, and that amount compounds. The printing of money that funds higher benefits feeds the very inflation that drives the next round of increases. It is a feedback loop with no easy exit.
The real threat is not simply higher prices. It is the combination of higher benefit costs and a shrinking base of taxable wages at the exact moment when the system can least afford either.
A Funding Gap That Arrives Faster Than Expected
Put the pieces together. AI-driven job losses of even 10 to 15 million white-collar positions would reduce payroll-tax collections by a meaningful percentage. At the same time COLAs of 6 or 7 percent would drive spending sharply higher. The arithmetic points toward the program exhausting its remaining reserves around 2029 under these conditions. That timeline is years earlier than most official projections that assume stable employment and moderate inflation.
Can the rest of the federal budget cover the shortfall? The broader picture is not encouraging. The annual deficit already sits near $2 trillion even with relatively full employment. Rising interest costs on the existing debt, higher unemployment, and continued inflation would push that figure much higher. More than a third of the national debt rolls over within a year and must be refinanced at whatever rates the market demands. Higher yields raise the interest bill. Lower yields require more money creation, which risks more inflation. Both paths carry heavy costs.
In my view the most likely political response will involve a combination of higher payroll taxes, a later full-retirement age, and some form of means testing. Raising the retirement age toward 75 would be deeply unpopular, yet the alternative of benefit cuts or dramatic tax increases looks equally painful. The political fallout would be substantial either way.
Why This Time Feels Different From the 1970s and 1980s
The last major rescue of Social Security occurred in 1983. At that moment a large generation of young workers was entering the labor force and could absorb higher taxes. Today those same workers are the beneficiaries. The younger cohorts replacing them face slower wage growth in many white-collar fields and direct competition from software that improves at a rapid pace. Demographic support that once existed has largely disappeared.
Monetary conditions also differ. The volume of money created in recent years far exceeds earlier episodes. Asset prices, housing costs, and government debt levels start from much higher baselines. The room for policy error is smaller. Suppressing demand through higher interest rates cannot fix supply constraints in energy or shipping. Creating more money to cover deficits risks accelerating the COLA spiral.
Practical Steps Worth Considering Now
None of this is inevitable in every detail, but the direction of travel is hard to ignore. Individuals still have time to adjust. Diversifying income sources beyond a single employer makes sense. Building skills that complement rather than compete with current AI tools can improve resilience. For those closer to retirement, stress-testing personal plans against higher inflation and potentially lower future benefits becomes essential.
Some assets have historically performed well during periods of elevated inflation and monetary stress. Companies with pricing power, hard assets, and certain commodity-related businesses often appear on that list. The exact vehicles matter less than the principle of not relying solely on the assumption that government benefits will remain unchanged in real terms.
- Review your expected Social Security benefit under different COLA and tax scenarios
- Consider how much of your future income depends on white-collar employment trends
- Evaluate holdings that have historically held value when paper currencies face pressure
- Maintain flexibility in career and geographic choices while the labor market is still fluid
- Watch the actual pace of AI adoption in your own industry rather than relying only on headlines
I have found that people who treat these shifts as distant problems often end up reacting later under more pressure. Those who begin adjusting earlier usually preserve more options. The combination of rapid technological change and structural fiscal challenges does not wait for convenient timing.
The Broader Economic Feedback Loop
Job losses in high-wage sectors reduce consumer spending power among people who traditionally support discretionary industries. That can slow growth even as inflation remains sticky in essential categories. Lower growth then pressures government revenues further while automatic stabilizers and higher benefit costs expand outlays. The resulting deficits feed back into the monetary system. It is not a simple linear story.
Energy markets illustrate the point. Higher fuel and electricity costs raise operating expenses for nearly every business. Companies respond by cutting labor where possible. AI provides a ready tool. The same inflation that justifies larger COLAs simultaneously encourages the job reductions that shrink the tax base needed to pay those COLAs. The loop is self-reinforcing until something breaks the cycle, either through policy change, technological limits, or market discipline.
Perhaps the most interesting aspect is how quickly the conversation has shifted. A few years ago discussions of AI focused mainly on productivity gains and new opportunities. Today the displacement numbers are appearing in layoff announcements and executive interviews. The speed of capability improvement continues to surprise even many of the people building the systems. Underestimating that speed has already proven costly for some firms and workers.
What Policymakers Face and What Individuals Can Control
Lawmakers will eventually confront choices that involve higher taxes, reduced benefits, later retirement ages, or some mix of all three. Each option carries political pain. Delaying the decisions only raises the eventual adjustment costs. In the meantime individuals retain more agency than they sometimes realize. Career choices, savings rates, investment allocation, and skill development remain under personal control.
I do not claim to know the precise year the pressure becomes acute. Official projections have been revised before and will be revised again. What seems clear is that the combination of AI-driven changes in the wage base and persistent inflation creates a more fragile environment for a program that already faced long-term demographic challenges. Preparing for a range of outcomes is simply prudent.
The coming years will test assumptions that many households and policymakers have treated as fixed. Social Security has survived previous crises through a mix of higher taxes and benefit adjustments. The difference this time lies in the speed of technological change and the limited demographic cushion. Whether the system adapts smoothly or experiences sharper disruption will depend on decisions still ahead. For anyone counting on those future checks, understanding the pressures is the first step toward navigating them.
Looking ahead, the interaction between automation and public finance deserves far more attention than it currently receives in everyday discussion. The tools arriving now are not incremental improvements. They represent a structural shift in how certain kinds of work get done. Ignoring that shift while simultaneously assuming benefit formulas and tax bases will behave as they have in the past is a risky combination. Better to examine the numbers clearly and adjust plans while options remain open.
In the end the story is less about any single technology and more about the gap between rising obligations and the economic base needed to support them. AI simply accelerates one side of that equation at a moment when the other side already faces strain. Recognizing the dynamic early gives individuals and families a better chance to protect their own financial security regardless of how the broader political process unfolds.
