The federal government is finally doing what it does best: setting up a committee to watch the sunset and wondering why it’s getting dark. Deputy Secretary of Commerce Don Graves and various policy leads are signaling a shift toward "analyzing the impact of AI on the workforce." It sounds responsible. It sounds measured. It is fundamentally a waste of time.
Watching the Department of Commerce "study" AI job displacement is like watching a horse and buggy fleet manager study the "impact" of the internal combustion engine by counting how many blacksmiths are still making shoes. They are measuring the wrong variables, in the wrong decade, using a philosophy of labor that died with the assembly line. Learn more on a related subject: this related article.
The current federal strategy treats AI as a localized storm front—something that might wash away a few coastal clerical jobs. In reality, we are looking at a fundamental rewrite of the value of human cognition.
The Myth of the Discrete Job Loss
The consensus view—the one your favorite "thought leaders" parrot—is that AI will delete specific roles. We hear about the doomed customer service rep, the endangered paralegal, and the soon-to-be-extinct data entry clerk. More reporting by Mashable delves into comparable views on the subject.
This is the first great lie. AI doesn't kill jobs; it kills tasks.
I’ve spent fifteen years inside the machinery of corporate automation. I have seen companies burn through eight-figure budgets trying to "replace" departments. It never works that way. Instead, the job remains, but the human inside it becomes a hollowed-out supervisor of an algorithm.
When the government says they are analyzing the impact on "jobs," they are looking for a body count that won't appear. You won't see a spike in unemployment filings titled "Ousted by GPT-5." You will see a slow, agonizing stagnation of wages as the "expertise" required for those jobs evaporates. If a machine can do 90% of a junior analyst's work, that analyst isn't fired—they are just never hired in the first place.
The federal strategy is looking for a bang. They’re going to miss the whimper.
The Productivity Trap
Washington loves to talk about "human-centric AI." It’s a comfortable phrase that means absolutely nothing. It suggests a world where AI is a polite assistant that hands us wrenches and fetches coffee.
Let’s look at the math. In the standard economic model, productivity gains are supposed to lift all boats. If a worker is more productive, they are more valuable.
But AI is a capital-heavy, labor-light technology. In every industry I have consulted for, the "productivity gain" from AI is immediately captured by the shareholders and the infrastructure providers. The worker doesn't get to go home early because the AI wrote the report; the worker is now expected to produce ten reports.
Federal analysts are operating on 20th-century assumptions that efficiency leads to stability. They are missing the Jevons Paradox. This economic principle states that as a resource becomes more efficient to use, the rate of consumption of that resource actually rises. Apply this to AI: as "intelligence" becomes cheaper and more efficient, we won't use it to save time. We will saturate every corner of our lives with it, creating a treadmill of synthetic content and algorithmic noise that requires more human oversight, not less.
We aren't heading toward a leisure society. We are heading toward a society of "Human-in-the-loop" janitors, cleaning up the hallucinations of $100 billion models for $15 an hour.
The Skill Floor is Collapsing
The government's plan likely includes a heavy dose of "upskilling." This is the universal band-aid for every technological disruption. "We’ll just teach the coal miners to code," they said in 2012. We saw how that ended.
The problem with "upskilling" in the age of AI is that the floor is moving faster than the students. By the time a federal grant matures to teach workers how to prompt a model, the model no longer requires prompts.
We are witnessing the devaluation of the entry-level.
- The Apprenticeship Gap: How do you become a Senior Architect if the AI does all the Junior Architect work?
- The Knowledge Decay: If you don't do the "boring" work of learning the basics because the AI handles it, you never develop the intuition required to spot when the AI is wrong.
- The Meritocracy Collapse: When everyone can produce "B+" work with a button press, the premium on talent vanishes.
Federal strategy won't address this because it requires admitting that our entire education-to-labor pipeline is obsolete. They would rather talk about "resiliency" than admit that a four-year degree is a losing bet against a model that updates every six months.
Why the "Data Privacy" Focus is a Distraction
Solomon and the federal cohorts often pivot to data privacy and bias. While noble, these are the "safe" problems. They are technical bugs that can be patched.
The real threat—the one the government is too scared to touch—is the liquidity of labor.
Imagine a scenario where a single specialized AI model can perform the cognitive work of 50,000 mid-level managers. This isn't a "bias" issue. It isn't a "privacy" issue. It is a fundamental shift in the bargaining power of the human race.
If the federal strategy doesn't address the fact that AI decouples wealth creation from human labor, it isn't a strategy. It's a PR campaign. We are moving toward a world where "labor" is a bug, not a feature, of the economy.
The Ugly Truth of Government "Impact Analysis"
I have sat in the rooms where these "impact analyses" are drafted. They are designed to be palatable, not accurate.
- They ignore the "Shadow AI" economy: Millions of workers are already using AI to do their jobs in secret, effectively automating themselves while pretending to work 40 hours a week.
- They rely on lagging indicators: Unemployment data is a rearview mirror. By the time the "impact" shows up in federal stats, the industry has already shifted.
- They fear the solution: Truly addressing AI’s impact would mean discussing Universal Basic Income, the total overhaul of intellectual property, and taxing compute instead of payroll. No politician wants to touch that third rail.
Instead, they will give us a "strategy" that suggests "monitoring" and "collaboration."
Stop Asking if AI Will Replace You
The question "Will AI take my job?" is a low-resolution question. It’s the wrong question.
The right question is: "How much of my value was actually based on being a glorified filing cabinet?"
If your value was based on synthesizing information, retrieving data, or following a set of repeatable steps, you are already gone. The federal government’s "analysis" won't save you. Retraining programs for 2024 skills in 2026 won't save you.
The only path forward is to lean into the things that are computationally expensive: high-stakes accountability, physical presence, and genuine, non-derivative creativity. Everything else is just waiting for the update.
The federal government is bringing a spreadsheet to a singularity fight. They are analyzing the "impact" on a workforce that is already transforming into something they don't recognize.
Stop waiting for the strategy. The strategy is to realize that the safety net they’re building is made of the very paper the AI is learning to shred.
Don't look for a new job. Look for a new definition of work.
