James Freeman deploys AI-productivity hype to shield entitlement benefit cuts from revenue scrutiny — and, in the finest tradition of the page I helped build, the move is executed so cleanly you could miss the trick entirely. In the June 9 “Best of the Web” column on the Wall Street Journal editorial page, Freeman wraps a call for Social Security and Medicare benefit reductions inside a Tyler Cowen–framed AI-growth miracle story, and treats the one as the cure for the other. The column runs approximately six hundred words and deploys five distinct rhetorical techniques across its paragraphs. This column walks through them as they appear. I know these techniques because I wrote sentences like this. I know them because they are the playbook. I know them because they work.
Media folk, perhaps fearful of the impact of artificial intelligence on their livelihoods, tend to cover potential economic disruption from AI as a threat. Today brings a reminder that the really scary possibility is that AI does not boost U.S. productivity at all, retirement programs go bust, and our kids and grandkids never crawl out from under the gargantuan boulder of federal debt.
Nick Lichtenberg reports for Fortune on economist Tyler Cowen, who explains why a technological revolution that makes various services more plentiful and cheap may be uncomfortable for some but is exactly what the U.S. needs. It’s about affordability for consumers—and also for taxpayers. — paragraphs 1–3
The opener executes a classic ad hominem (circumstantial) — Bad-Faith Catalog: ad_hominem — and it’s doing it inside a multiple-audience-targeting structure (WSJ §4.3) so smooth you can hear the morning-meeting planning in it. The “media folk” dismissal is aimed at the business-class reader who’s been trained since the Ailes-Limbaugh fusion to distrust any newsroom’s word on anything. The pivot to “our kids and grandkids” yanks the emotional register toward the concerned-parent reader who’s never heard of Tyler Cowen but recognizes debt-anxiety from the nightly-news segment the editorial board spent two decades building. We used to call this the “broaden-the-coalition” move in the greenroom: you hit the base with the tribal cue, then you throw a bridge to the persuadable middle with the family-heritage framing. The reader gets the dismissal and the worry at once. The dismissal does the gatekeeping; the worry does the moving.
Today’s reminder from the federal government is that the kids and grandkids won’t be the only ones in a very tough situation. Population is growing slowly. Birthrates are low and immigration is being curtailed. U.S. public retirement systems have far fewer workers per retiree than when Washington created the legalized Ponzi scheme of the entitlement state. — paragraph 5
Frame-engineered relabeling — WSJ §A.4.1 / Bad-Faith Catalog: frame_engineered_relabeling — operates here through the substitution of “Ponzi scheme” for “social-insurance program.” Social Security is a statutory insurance mechanism funded by a dedicated payroll tax; its solvency mechanics are documented annually in the trustees’ report, which is precisely the document Freeman is about to cite. A Ponzi scheme is a criminal-fraud construct in which earlier participants are paid from new entrants’ capital with no productive activity between them, and the label carries a criminal undertone that “social-insurance program” does not. I built sentences of this kind. We knew the word “Ponzi” would make the reader recoil before the policy argument even began, and that recoil would do the argumentative work so the argument itself didn’t need to happen. Once the program is labeled a fraud, benefit cuts cease to be austerity and become the exposure of a con. The relabeling is the con — it replaces a policy discussion with a moral one, and the moral frame is designed to produce a conclusion the evidence would not reach on its own.
The 2025 Social Security Trustees Report records a worker-to-beneficiary ratio of 2.7 to 1 in 2024, down from 5.1 to 1 in 1960 — the demographic squeeze Freeman invokes without specifying. And the “study-shows” ledger (WSJ §4.5) is at work too: a Tyler Cowen lecture is treated as a neutral expert finding, never mind that Cowen is among the most prominent advocates of AI-driven labor-market disruption as a feature, not a bug, and the column never surfaces that advocacy.
Social Security trustees report on the combined fund of Social Security’s retirement and disability insurance programs…
The Medicare report is no less sobering, with the Hospital Insurance trust fund, otherwise known as Medicare Part A, scheduled for depletion in 2033, when revenues are expected to cover 89% of incurred program costs.
Barring significant reforms of these programs—or faster growth to pay for them—the only options in our very near future are huge tax hikes and benefit cuts. — paragraphs 6–7 (paragraph gap elided at original)
The solvency numbers are worth stating plainly because the rhetorical architecture depends on them feeling intractable. According to the 2025 Social Security and Medicare Trustees Reports, the combined OASDI trust funds are projected to be depleted in 2034, with the Hospital Insurance trust fund running out in 2033 — at which point continuing revenues would cover approximately 80% and 89% of scheduled benefits, respectively. These are actuarial gaps, not a binary insolvency. They can be closed through measured revenue adjustments without any AI-driven growth at all. But Freeman’s column treats them as a catastrophe that only an AI boom can avert — which is the permission structure the piece needs to convert a solvency problem into a growth-deregulation demand.
And then the false dichotomy — false dichotomy, Bad-Faith Catalog: false_dichotomy — deployed in the sentence that offers “huge tax hikes and benefit cuts” as “the only options.” The trustees’ report identifies revenue-side mechanisms that have been available since the program’s inception: raising or eliminating the Social Security taxable-wage cap, which currently stops collecting FICA on income above $184,500; subjecting all income to FICA; lowering healthcare costs through direct price regulation; using general revenue to shore up the trust funds. The column frames the binary as punitive “hikes” versus corrective “cuts” while omitting the structurally available option of broadening the contribution base — a fix that the trustees’ own modeling shows extends solvency by over two decades without a single rate increase. Options narrowing is the oldest operator move on the page: you don’t lose the argument on the merits when the merits aren’t on the ballot. The frame is the scam — it manufactures a binary and then asks the reader to choose between two bad choices, both of which serve the same donor-class outcome.
Now imagine a world where we opt into this fiscal calamity by unilaterally halting the development of a productivity-enhancing, growth-enabling technology that is now dominated by the United States. — paragraph 8
Strawman — Bad-Faith Catalog: strawman (selectional) — constructed in the proposition that anyone would “opt into this fiscal calamity by unilaterally halting the development” of AI. No serious policymaker, labor organization, or academic has proposed unilaterally halting AI development. The actual debate is about pacing, safety standards, labor-transition policy, and the FICA-revenue impact of displacement — a variable the column must never mention, because mentioning it would reveal what the piece is actually doing. AI is already projected to displace millions of workers whose payroll taxes fund the trust funds; the most immediate risk is that AI worsens the shortfall by shrinking the taxable wage base under high-adoption scenarios, as projected by the Congressional Budget Office. Mentioning that would force the column to confront the quiet truth at its center: the very machines Freeman frames as the rescue are the engines that will hollow out the tax base the rescue is supposed to fund. The strawman — invent the extreme, defeat it, claim the victory applies to the moderate — clears the field so his borrowed Cowen argument proceeds unchallenged. That’s the play. You don’t need to refute the actual argument when you can invent the one you know how to beat.
Anyone tempted to believe that the solution to the high cost of highly regulated U.S. health care is an even larger role for government should consider the ongoing tragedy of patient care in the U.K.’s National Health Service. — final paragraphs
The blue-state-failure frame — WSJ §A.4.9, cross-applied here to a foreign-government system as scare-tactic — deploys the NHS as a horror story: “ongoing tragedy,” invoked not as a policy subject but as an emotional anchor designed to foreclose any conversation about a public role in U.S. healthcare. The U.K. health system operates under successive austerity programs imposed since 2010, with real-terms funding reductions and staff budgets set by political choice, not by the inherent logic of public healthcare. The column treats the policy outcome of deliberate starvation as the structural destiny of the model. The closing line is the classic threat-inflation closer — WSJ §A.4.13 — converting a national health service with per-capita expenditures that run at roughly half the U.S. rate into an eighty-year “ghastly experiment” that must be abandoned, with not a single outcome statistic in sight.
So here is what Freeman’s column actually does when you put the excerpts together and stop letting the paragraph-by-paragraph sleight of hand distract from the architecture. He takes the real, solvable shortfall in the Social Security and Medicare trust funds — a shortfall driven largely by an aging population and the underexposed problem of healthcare cost growth — and turns it into a single-choice demand: either we let AI rip with no guardrails, or we bankrupt the old and burden the young with debt. The names he will not say are the names of the people whose payroll taxes fill the trust funds. The variable he will not name is the displacement track: the job loss that would reduce FICA contributions and accelerate the very insolvency Freeman claims AI will solve. The machines replace the workers; the workers stop paying the tax; the trust fund empties faster, not slower. This is not an analysis of how AI interacts with entitlement solvency. It is a prayer dressed in a productivity chart, and the prayer is that nobody asks who the machines pay taxes for.
The column’s true product is permission: permission for the owner-class reader to demand AI deregulation not out of self-interest but out of a solemn duty to “save Social Security.” That is the move I used to write for the page. The reader gets the policy they wanted all along, but now they get to feel virtuous about it. Freeman needs his audience to feel that cutting benefits is a rescue operation — that the scissors are the lifeboat — and AI productivity is the narrative life raft he throws into the water: if the machines work fast enough, the math fixes itself, nobody has to choose between the two bad options the column let them see, and the policy the page always wanted — benefit cuts — can proceed without the political cost of saying it out loud. The arc is the arc: ad hominem disqualification, Ponzi-frame relabeling, the false dichotomy that vanishes the revenue-side remedy, the strawman that invents an opponent in order to beat one, the NHS boogeyman deployed to shame the reader away from any public-option thought — each move serving to bury the variable the page cannot afford to name, which is what happens to the payroll tax base when the machines replace the payers. It is not economics. It is a permission structure, and the permission is to feel good about what the page has always wanted.
— Phukher Tarlson