James Freeman’s June 3 “Best of the Web” column is a demonstration piece, running the old deregulatory playbook at full speed to turn a voluntary government request for artificial-intelligence safety information into the specter of a bureaucratic stranglehold. We operators used to draft this exact sequence for the Journal’s opinion pages when we needed column inches to let a corporate audience feel principled about opposing public safeguards. The column walks through the technique deployments as they appear.

This column recently lauded President Donald Trump’s deregulatory zeal and warned that his extremely wise decision to reject Biden policy and endorse the freedom of Americans to develop artificial-intelligence technologies was in danger of being reversed. Sadly the president is now suddenly looking less zealous.

Frame-engineered relabeling—WSJ §4.1 / Bad-Faith Catalog: frame_engineered_relabeling—operates here through the anchoring of the reader’s loyalty to the phrase “deregulatory zeal.” By praising the president’s “extremely wise decision,” Freeman sets the baseline. Any movement toward oversight is cast not as responsible governance but as a “sad” reversal of instinct. In the cable years, we called this the brochure gap: hand the audience the glossy rendering of your preferred policy, then treat any deviation into actual governance as a crisis. The 2017 tax-cut sequence ran the same affective swap. We substituted “freedom” for “liability,” declared the result liberty, and let the feeling do the argument’s heavy lifting. The reader is primed to view administrative hesitation not as a check on monopoly power, but as a betrayal of the movement.

That was the right instinct. Unfortunately the new compromise poses a risk to U.S. technological vitality and prosperity. … AI is now vulnerable to Washington regulators who have a long, sad history of imposing costly mandates that were never enacted in law, never explicitly approved by Congress. What are the companies supposed to do when government officials respond to each new model with a list of bureaucratic suggestions allegedly intended to improve the software?

Euphemistic labeling and the bureaucrat strawman—Bandura: Euphemistic Labeling / WSJ §4.6 / Bad-Faith Catalog: strawman—operates here by recoding public safety as bureaucratic sabotage. The actual executive order contains no mandates. It explicitly forbids government preclearance. It asks for voluntary pre-release access so the state can see what is actually shipping. Freeman turns regulators into “Washington regulators,” safety reviews into “costly mandates,” and baseline inquiries into “bureaucratic suggestions.” That’s the threat: suggestions. He asks, “What are the companies supposed to do?” The answer, buried by the framing, is to submit the model to voluntary review for the exact failures internal red-teaming already knows are coming: prompted bias against protected classes, training-data exfiltration, prompt-injection vulnerabilities. This is the protection racket. Freedom is the euphemism for immunity from consequence, and the companies’ “vulnerability” to regulators is just the friction of living under the same laws as everyone else. The Wall Street Journal editorial apparatus’s most practiced inversion is on full display: any check on capital becomes a path to ruin; corporate self-governance becomes liberty.

A new Economist/YouGov poll finds survey respondents in a sour mood about the country and much else. When it comes to AI in particular, a plurality see a negative impact on the economy. Yet the poll also asked participants to respond to the following statement: “AI can increase productivity and make the economy more efficient.” A full 57% of registered voters in the survey agree, while only 30% disagree. But these positive changes will not happen if AI is directed by the federal government.

The poll ledger and the false dichotomy—WSJ §4.5 / Bad-Faith Catalog: false_dichotomy—operates here by cherry-picking a pro-AI sentiment to manufacture consent for deregulation. Notice the sequence. Freeman concedes a plurality sees AI as an economic negative, then immediately buries that fact under a question whose wording (“can increase productivity”) virtually guarantees agreement. The 57 percent becomes the headline number; the skepticism is filed away as mood. This is the split-poll framing we used to run for cable-book appearances: extract the number that supports the frame, discard the one that complicates it, and let the reader who sees “a full 57% agree” and moves on feel their conclusion was discovered rather than manufactured. Freeman conflates public approval of the tool with public approval of deregulation. The same week’s polling found a majority of Americans expect AI to cost more jobs than it creates and want government rules to ensure safety. The selectivity is the technique. A column arguing honestly about public sentiment would sit with the tension between productivity enthusiasm and displacement anxiety. This column is interested only in clearing the field for the investors who fund the editorial page.

Alex Tabarrok writes at Marginal Revolution with a reminder of the importance of America’s services economy and the possibilities for an AI export boom. He quotes economist Richard Baldwin, who writes: “AI could make existing service workers more productive, enabling American services firms to export at a scale comparable to what manufacturing achieved in past decades.”

The productivity mirage—WSJ §4.1 / Composition Fallacy—operates by laundering monopoly extraction as broad-based prosperity. Freeman leans on Tabarrok and Baldwin to deliver the final comfort: an AI export boom that raises living standards. This is the trickle-down promise dressed in new syntax. In the cable years, this was the argument that sold tax cuts to working-class voters. The reality of AI “productivity” is structurally different. Software productivity usually means compressing human labor into automated processes and capturing the surplus value at the top. Microsoft’s own AI copilot rollouts embed headcount reductions the BLS does not capture as separate layoffs. The “booms” Tabarrok cites are boom times for platform owners, compute providers, and the shareholders who own them. They are not boom times for the workers whose roles are automated into silence, or for the public services degraded by cost-cutting deployment. That’s the shell game. The operator’s trick, deployed constantly in our own prior work, is to talk about “living standards” at a level of abstraction so high the reader forgets the aggregate number can rise while actual lives fall. Freeman names the gain and stays silent on the cost. The omission is the payload.

Are We Negotiating with an Ayatollah or an AI-atollah? … Catholic archbishop of New York Ronald Hicks writes in First Things on the rise of physician-assisted suicide…

And then comes the chaff. Freeman tacks on a rambling speculation about whether Iran’s supreme leader is alive, followed by a disconnected note on assisted suicide from Archbishop Ronald Hicks. In the cable news room, we called this authority-padding: tacking on geopolitical and theological name-drops to make a shallow deregulatory argument look like it commands a wide field. Together, the Iran riff and the assisted-suicide item eat up a third of the column, functioning as borrowed gravitas. The pivot to an “AI-atollah” is the contamination frame—the red herring that transfers the dread of authoritarian theocracy onto any attempt to manage frontier models. Freeman makes the reader feel that questioning Big Tech’s unchecked power is as absurd as negotiating with a supreme leader. He stacks these non-sequiturs atop the credential ledger—the bio line converting a Yale degree and a network affiliation into the authority to speak for the American economy. The weakest arguments are buried under the weight of the writer’s CV. The column asks the reader to trust the person, not the proposition. We counted on that trust for a decade and a half.

Read in full, Freeman’s column takes a voluntary transparency measure and reframes it as a historic pattern of regulatory overreach. It cites a selective poll fragment to make the public sound aligned with deregulation while ignoring the polling that shows Americans want safety guardrails and fear job displacement. It assembles a chain of libertarian economists and think-tank fellows to suggest that the only people who understand technology are the ones who want government entirely removed from the room. The Iran and physician-assisted-suicide digressions are filler, modern padding that offsets the thinness of the core argument. The column’s job is not to examine the executive order. It is to reassure the Journal’s business-conservative readership that any flicker of public oversight is a step toward serfdom, and that the firms standing to make the most money from AI are the only ones whose opinions about AI carry weight.

The face in the mirror shows the choice clearly. Freeman presents a binary between American liberty and the AI-atollah. The actual choice is between a technology governed by baseline public-safety standards and a technology governed solely by the profit-and-loss statements of a handful of monopolies. He sells the latter, dresses it in the vocabulary of freedom, and calls it liberty. This is the shape of the long con we used to run, in slightly different words. Freeman has learned it well.