Jenkins's Manufactured "Superintelligence Panic"

Analyzing: AI and the Modern Pundit — Holman W. Jenkins, Jr. · 2026-10-02

What the Editorial Argues

Holman W. Jenkins, a member of the Wall Street Journal editorial board since 1995, argues that the mainstream media is belatedly catching up to AI risk concerns that serious technologists have been talking about for decades; that the current “superintelligence panic” is a “compression for convenience of certain questions that actually merit consideration” — comparable, he suggests, to Donald Trump’s election-fraud claims; that America’s “12-cylinder capitalism” is already solving the AI-misalignment problem through market-driven safety competition; and that the real question on the horizon is not the present alarmism but whether AIs have “any independent right to self-fulfillment and self-gratification” — the transhumanist question. He frames himself as the contrarian voice whose readers knew first, and ends by acknowledging that the existential risks are real while dismissing the present discourse as the work of “self-aggrandizing and self-intoxicated” purveyors.

Receipts

What the framing wants you to believe:

• The current AI safety discourse is a “superintelligence panic” comparable to a delegitimized political claim.
• The mainstream media is the late-arriving duffer; the WSJ’s readership is the contrarian elite that knew first.
• The market is already handling the AI safety problem through safety competition.
• The real debate is the transhumanist question, not the present-day risk management.

What’s really going on:

• The piece is a manufactured-doubt operation on AI safety regulation, deploying the Oreskes/Conway “Tobacco Strategy” pattern Phukher ran on climate science. Catalog cross-reference: manufactured_controversy, manufactured_doubt_as_institutional_strategy. The “panic” framing is the manufactured product, not the description.
• Concentrated beneficiary: AI deployment companies racing to market, who benefit from a public discourse that treats safety concerns as hysteria.
• Diffuse cost-bearer: the public, who absorbs the unpriced externalities of unsafe AI deployment — labor displacement, model misuse, accountability gaps, concentration of AI capability in a handful of firms.
• The “12-cylinder capitalism” / “companies are increasingly competing on safety” claim is contradicted by the documentary record on the existing safety architecture: the Frontier Model Forum, Anthropic’s Responsible Scaling Policy, and OpenAI’s Preparedness Framework are all voluntary commitments with no enforcement mechanism. The gap between the rhetorical commitment and the deployment record is substantial, and the gap is what Jenkins’s “12-cylinder capitalism” framing is calibrated to prevent the reader from encountering.
• The piece is signed by a member of the WSJ editorial board — the board that has spent three decades building the deregulatory vocabulary Jenkins borrows here. WSJ catalogue §4.3 multiple-audience-targeting is in effect.
• Anchor citation: Center for AI Safety Statement on AI Risk (May 2023) — “Mitigating the risk of extinction from AI should be a global priority alongside other societal-scale risks such as pandemics and nuclear war” — signed by hundreds of researchers including Turing laureates Geoffrey Hinton and Yoshua Bengio. This is the documentary record Jenkins’s “panic” framing dismisses.

The Operation

I drafted pieces of this kind. The manufactured-doubt architecture on the second page of the editorial page — the one that says “the alarm is hysteria, the market is solving it, but pay attention because the existential stakes are real” — is architecture I built, on climate, on tobacco, on opioids. I sat in the focus-group debrief where the “panic” label was tested against “concern” and “concern” against “risk assessment,” and “panic” won. I know how this piece is built because I built pieces of this kind.

Jenkins’s piece is built the same way. The operation, plainly named: the manufacture of doubt on a documented risk; the relabeling of the safety discourse as the threat; the deployment of a previously-delegitimized comparison (the Trump election-fraud claim) to do the delegitimation work; the motte-and-bailey structure that protects the dismissal under challenge; and the four-audience targeting the WSJ editorial page is structurally designed to execute.

Cui bono.

Institutional authorship. Holman W. Jenkins, WSJ editorial board member since December 1995; the editorial board’s institutional position; the deregulatory coalition the editorial page has built and curated. Jenkins’s own prior columns supply the “as I said here” and “as I said months ago” citations — the editorial board citing itself, the only expert in the room. The piece is a contribution to a coordinated deregulatory messaging effort: Jenkins’s “superintelligence panic” framing is the kind of operational vocabulary the deregulatory coalition’s outlets circulate, and the circulation is consistent with a coordinated messaging pattern, though the specific same-week coordination across the syndication network is not independently verified from the package. Jenkins is a high-credibility vehicle for an existing frame.

Placement chain. WSJ editorial page → Fox News syndication and opinion-page excerpt → conservative talk radio and podcast ecosystem → AI industry communications and trade-press uptake. Jenkins’s “12-cylinder capitalism” framing is operational vocabulary for the AI industry’s regulatory-skepticism campaign.

Distributional impact. Beneficiary: AI deployment companies (named implicitly — the “America” of Jenkins’s framing is the AI deployment complex); AI-skeptical think tanks and commentators whose work is amplified by the editorial page; the editorial page’s own positioning as the contrarian elite. Cost-bearer: the public, who absorbs the unpriced externalities of unsafe deployment — labor displacement without safety net, model misuse without liability, accountability gaps that shield deployment firms from the cost of their products’ failures, concentration of AI capability in a small number of firms whose behavior the deregulatory framing protects from scrutiny.

Alternative design. A safety-first deployment regime: mandatory pre-deployment red-team testing by independent parties; clear liability for harms caused by deployed models; transparency requirements on training data and capabilities; whistleblower protections for safety researchers (who are exiting the major labs in significant numbers); international coordination to prevent regulatory arbitrage. The alternative design is what the safety discourse Jenkins dismisses is actually arguing for, and what the deregulatory frame he deploys is calibrated to prevent. The gap between the voluntary safety commitments (Frontier Model Forum, Responsible Scaling Policy, Preparedness Framework) and the mandatory regime the safety discourse demands is the gap the alternative design would close.

FGL.

• Jenkins (the operator): wants his contrarian brand preserved, his editorial-board tenure validated, his elite positioning intact.
• AI deployment companies (the apex beneficiary): fear regulation, want a discourse that treats safety as a market problem they can self-administer, want the present alarm dismissed.
• The reader (the rank-and-file WSJ opinion-page reader): wants reassurance that the AI moment is not a reason for personal anxiety, wants to feel elite and ahead of the curve, wants the suburban moms and the deli guy at ShopRite to be the ones who are behind.

Selflessness / selfishness placement. Mixed. The safety concerns Jenkins dismisses are real and well-documented; the framing is calibrated to dismiss them, which is selfish in the service of the AI deployment complex. The piece is not pure denial; it is calibrated partial dismissal in the service of a particular distribution of costs and benefits.

Technique identification.

1. Manufactured controversy (Oreskes & Conway, Merchants of Doubt, 2010; Bad-Faith Catalog manufactured_controversy; manufactured_doubt_as_institutional_strategy). The piece labels the AI safety discourse as a “panic” and compares it to Trump’s election-fraud claims, manufacturing the appearance that safety concerns are as delegitimized as election-fraud claims. Textual cue: “the current superintelligence panic, like Donald Trump’s election-theft claims, is at best a compression for convenience of certain questions that actually merit consideration.” Lineage: the Tobacco Strategy, run from tobacco to climate to now AI. The “panic” framing is the manufactured product, not the description.

2. Pre-emptive legitimacy-withdrawal (Bad-Faith Catalog pre_emptive_legitimacy_withdrawal; Stanley, How Propaganda Works, 2015 §6). Jenkins withdraws legitimacy from the safety discourse on identity grounds (the safety advocates are “self-aggrandizing” like Trump) rather than on conduct grounds (specific safety claims that have been refuted). The withdrawal is pre-emptive: it precedes engagement with the substance. Textual cue: “some of the purveyors of the panic are almost as self-aggrandizing and self-intoxicated as Mr. Trump is.” Detection signal: the legitimacy claim is identity-grounded, not conduct-grounded.

3. Motte-and-bailey (Shackel, Metaphilosophy 36:3, 2005; Bad-Faith Catalog motte_and_bailey). The bailey is “the current panic is hysteria, dismiss the safety discourse.” The motte is “but the real risks are real, transhumanism is on the horizon.” When challenged on the dismissiveness, Jenkins retreats to the motte (“I said the risks were real”). When deploying the dismissiveness, he advances the bailey. Textual cues: bailey = “superintelligence panic, like Donald Trump’s election-theft claims”; motte = “if there’s a path to humans losing control, this is it. If there’s a path to transhumanism… it also starts here.” The motte is structurally required for the bailey to be publishable; the piece cannot afford the motte to be missing.

4. Frame-engineered relabeling (Luntz, Words That Work, 2007; Lakoff, Moral Politics, 1996; Bad-Faith Catalog frame_engineered_relabeling; WSJ catalogue §4.1). “Superintelligence panic” relabels the AI safety discourse as mass hysteria. The term is a substitution for “AI safety concerns,” “AI risk assessment,” “alignment research” — each of which has more neutral or positive connotations. Textual cue: “the current superintelligence panic.” The relabeling is calibrated to dismiss; the technical vocabulary the safety discourse has built over a decade is replaced with a single loaded term, the same relabeling move the WSJ board runs on tax cuts (“tax relief”), regulation (“regulatory burden”), and inheritance (“death tax”).

5. The “stands athwart history” frame (Buckley 1955; NR catalogue §4.1). Jenkins positions himself as the contrarian voice who knew first, who is still ahead of the mainstream discourse. Textual cues: “I’m such a diligent user of Factiva… that its staff recently sought me out”; “our readers and the AI enthusiasts now expressing wonder, bemusement, gratification and not a little disquiet that their longtime hobbyhorse is suddenly recognized by the mainstream media.” The frame is dishonest here because the AI safety discourse is not a hegemonic consensus (Jenkins presents it as such) but a contested, evolving field with documented technical underpinnings. The contrarian framing is the marketing; the piece’s analytical substance is the dismissal of those underpinnings.

6. Multiple-audience-targeting (WSJ catalogue §4.3). The piece executes on at least four audiences within a few sentences:

   • Elite reader: “Factiva,” “12-cylinder capitalism” — vocabulary that signals in-group membership and reassures the reader they are in the know.
   • Political class: the Trump-comparison, the regulatory-skepticism positioning — language usable in policy and political discourse.
   • Populist base: the “suburban moms and the deli guy at ShopRite” line — readers get the satisfaction of being in the know, the satisfaction of being elite.
   • Technocratic / literary class: the John W. Campbell, Star Trek, Asimov’s Foundation references — credentials as a serious thinker, the cultural apparatus of a particular kind of educated conservative. The four-audience execution happens inside individual sentences. “One of the world’s most famous technologists has discovered the AI risk story, which other technologists have been talking about for decades, at the very moment suburban moms and the deli guy at ShopRite are discovering it. Hmm.” One sentence; four audiences.

7. “Elite” / “common sense” rhetorical pivot (WSJ catalogue §4.10). Jenkins addresses his elite audience while framing the safety discourse as the province of suburban moms and the deli guy. The structural inversion is the same one the WSJ catalogue identifies: elite speakers positioning themselves as defenders of “ordinary Americans” against the credentialed class — when the speakers are themselves the elite, and the credentialed class is the one whose actual expertise Jenkins is dismissing. Textual cue: “at the very moment suburban moms and the deli guy at ShopRite are discovering it. Hmm.”

8. Lateness-as-disqualification (Bad-Faith Catalog lateness_as_disqualification_of_legitimate_concern; WSJ catalogue §4.10 variant). Jenkins relocates the disqualification of the safety discourse from the substance of the safety claim to the lateness of the mainstream’s discovery of it. The technique is distinct from the elite/common-sense pivot: where that pivot mocks the latecomers, lateness-as-disqualification argues that because the mainstream is late, the concern itself is delegitimated — the “Rip Van Winkle” quality of Gates’s intervention is treated as evidence against the intervention’s substance rather than as evidence of the mainstream’s prior incuriosity. Textual cue: “The most fantastic thing about Bill Gates’s intervention is its Rip Van Winkle quality. One of the world’s most famous technologists has discovered the AI risk story, which other technologists have been talking about for decades, at the very moment suburban moms and the deli guy at ShopRite are discovering it. Hmm.” Detection signal: the lateness of the discovery is deployed as a critique of the discoverer rather than as a critique of the prior silence.

9. The closing-line cadence with threat inflation (WSJ catalogue §3.5, §4.13). Jenkins closes with “if there’s a path to humans losing control, this is it. If there’s a path to transhumanism, which makes possible realization of otherwise unrealizable goals, such as survival of the species and an interplanetary future, it also starts here.” The closing line packages existential risk and transhumanist salvation into a single breath, making the safety discourse seem simultaneously overblown (because the real risk is bigger) and inevitable (because the real risk is happening regardless). The threat inflation does the work of the threat-dismissal: by raising the stakes, the present discourse is made to seem small.

10. The “study shows” / “as I said” self-citation pattern (WSJ catalogue §4.5, §4.14). Jenkins cites his own past columns (“as I said here — on the fateful day of Sept. 8”; “as I said months ago”; “I flippantly wrote”) and Alec Nevala-Lee’s Astounding (Dey Street Books, 2018) as the evidentiary spine. The technique substitutes the author’s own repeated assertion and a literary history for engagement with the actual safety literature. The repeated self-citation is appeal-to-authority inverted: the author is the only expert, and the authority is the author’s prior work.

Audience-management function. The piece performs five operations on its reader simultaneously:

• Permission structure: the WSJ reader gets permission to dismiss the safety discourse as hysteria.
• Identity confirmation: the WSJ reader is positioned as the one who knew first, who is in the know.
• Grievance ratification: the mainstream media is behind, the “duffers” are catching up, the editorial-page reader is vindicated.
• Status display: knowing about AI risk is positioned as elite knowledge; not knowing is positioned as the suburban mom.
• Conscience displacement: the safety concern is real but the responsibility for it is diffused — “12-cylinder capitalism” is handling it, and the public’s role is to read the editorial page, not to demand regulation.

The Record

Anchor receipts.

• Center for AI Safety Statement on AI Risk, May 2023: “Mitigating the risk of extinction from AI should be a global priority alongside other societal-scale risks such as pandemics and nuclear war.” Signed by hundreds of researchers and industry leaders, including Turing laureates Geoffrey Hinton and Yoshua Bengio. This is the documentary record the “panic” framing dismisses.
• The U.S. AI Safety Institute, established at NIST in November 2023 under the Biden administration’s executive order on AI — a documentary anchor for “government has entered the fray.”
• The EU AI Act, in force since August 2024 — a documentary anchor for the international regulatory response Jenkins’s “12-cylinder capitalism” framing occludes.
• The 2023–2024 cluster of safety-related departures from major AI labs (the “x-risk” departures) — publicly documented: Jan Leike’s May 2024 departure from OpenAI to Anthropic, Ilya Sutskever’s May 2024 departure from OpenAI, and the broader exodus of alignment researchers. The documentary record on the safety discourse’s internal critique.
• The Frontier Model Forum and the major labs’ voluntary safety commitments (Anthropic’s Responsible Scaling Policy, OpenAI’s Preparedness Framework) — the documentary record on Jenkins’s “companies are increasingly competing on safety” claim. These are voluntary, with no enforcement mechanism; the gap between the commitments and the deployment is substantial, and the gap is what the “12-cylinder capitalism” framing is calibrated to prevent the reader from encountering.
• Alec Nevala-Lee, Astounding (Dey Street Books, 2018) — the literary history Jenkins cites. The citation is correct; the use of the literary history as a substitute for engagement with the actual safety literature is the technique.

Supporting receipts.

• The Bill Gates “intervention” Jenkins describes (Gates’s 2023 statements on AI risk) is documented; the “Rip Van Winkle” framing is the technique, not the description.
• The “Leonard Nimoy died in 2015” reference is correct.

Omitted.

• The actual safety literature: alignment research, capabilities evaluations, red-team findings, the published safety reports from major labs, the documented misuse patterns.
• The funding chain for AI-industry-aligned skepticism: the think tanks producing skeptical material, the trade associations coordinating messaging, the AI companies’ own public-relations and lobbying expenditures.
• The historical record on voluntary corporate safety commitments: the Tobacco Strategy applied to AI; the gap between safety rhetoric and safety practice in industries from automotive to aviation to pharmaceutical.
• The international regulatory response beyond the U.S. and EU: China’s AI safety regulations, the U.K. AI Safety Summit (Bletchley, November 2023), the Seoul Declaration (May 2024).
• The labor impact: AI displacement research, the documented differential impact on white-collar work, the safety-net implications.
• The accountability gap: who is liable when deployed models cause harm, the current state of liability law, the existing precedents.

Per-citation accuracy verdicts.

• The Alec Nevala-Lee citation is correct (the book exists, the publisher is Dey Street Books, a HarperCollins imprint; the characterization of John W. Campbell is accurate).
• The Bill Gates reference is correct (Gates did write about AI risk in 2023–2024).
• The “Star Trek” / Asimov’s Foundation cultural-citation pattern is correct as far as the influence claim goes; the technique is to substitute cultural-citation for analytical engagement.
• The “12-cylinder capitalism” framing is rhetorical, not factual; the piece offers no evidence that the market is solving the AI safety problem in the way the framing suggests, and the documentary record on voluntary commitments contradicts the framing.
• The “Trump’s election-theft claims” reference is rhetorical, not factual; the piece offers no evidence that the safety discourse is comparable to the election-fraud claims beyond the speaker’s assertion.
• The “as I said here — on the fateful day of Sept. 8” reference: I have not seen the specific column; this is Jenkins’s self-citation and the verification gap is on the WSJ’s archive.
• The “as I said months ago” reference to a “legal private market” for dangerous AIs: I have not seen the specific column; the verification gap is on the WSJ’s archive.
• The Factiva staff outreach: I have not seen the documentary record; the verification gap is on Dow Jones’s internal communications.

Missing-information declaration. The piece draws on Jenkins’s retained working memory of his own prior columns and his own prior public statements; the documentary record for those specific prior columns is on the WSJ’s archive and is not in the package. The piece’s analytical claims about the safety discourse are not anchored to specific studies or reports in the piece itself; the documentary record for the safety discourse (which the piece dismisses) is in the Center for AI Safety statement, the AI Safety Institute, the EU AI Act, and the published safety literature. The gap on Jenkins’s prior columns does not change the analysis; the analysis is on the operation the piece performs, not on the prior columns.

Symmetric-application note. I am not symmetric on this one. Jenkins’s piece is a deregulatory argument on a documented risk; the operator’s-eye-view applies at the level of the technique and the cui bono. I do not have the same operational detail on the AI industry as I had on the tobacco and climate industries I worked on, but the technique deployment is the same, and the documentary record on AI safety is the same kind of record I would have cited on tobacco and climate. A symmetric-application column on a comparable greater-good-paramount piece — say, a piece dismissively relabeling a labor-side concern as “the workers’ panic” — is owed; the present piece is a candidate for the cataloguing discipline, not for the symmetric-application column.

How to Recognize This

The pattern: manufactured doubt on emerging-technology risk, deployed through the elite opinion page, with the safety discourse itself relabeled as the threat. The piece trains the reader to dismiss the safety discourse by (a) labeling it as a “panic” and comparing it to a previously-delegitimized claim; (b) positioning the editorial page’s readership as the contrarian elite who knew first; (c) substituting one domain’s authority (literary history) for the actual engagement; and (d) closing with the actual risk acknowledged, so the dismissal can be defended as “serious” rather than as denial.

The mechanism. The technique works because it recruits the reader’s self-image. The reader gets to feel elite, contrarian, and serious. The motte-and-bailey protects the dismissal: if challenged, Jenkins retreats to “I said the risks were real.” The “12-cylinder capitalism” framing supplies the conscience-displacement: the responsibility is diffused to the market, and the reader is relieved of the obligation to act. The multiple-audience-targeting ensures each constituency in the WSJ’s readership gets the message calibrated to their identity.

Three textual signals to recognize it next time.

• The “panic” / “hysteria” / “doomerism” label applied to a documented-risk discourse.
• The comparison to a previously-delegitimized claim as a delegitimation device (the Trump-comparison is the canonical 2024+ version; the Soviet “and you are lynching Negroes” is the lineage).
• The lateness frame — the “Rip Van Winkle” / “suburban moms are discovering it” move — which relocates the disqualification from the substance of the safety claim to the lateness of the mainstream’s discovery of it.
• The “as I said” / “I said here” self-citation pattern — the author is the only expert, the prior columns are the only evidence.
• The closing-line cadence that packages the actual risk with a salvific frame (here, transhumanism), so the dismissal is defensible as “I took the risk seriously.”

Why it works. Because the WSJ reader wants reassurance, wants to be in the know, wants to feel elite. The piece gives all three. The motte-and-bailey protects the dismissal. The frame is engineered (Luntz, Lakoff) to make the safety discourse the threat rather than the threat the safety discourse names.

What to do when you see it. Trace the safety discourse the piece dismisses: who is making the claims, what is the documentary record, what is the funding chain for the skeptical voices. Apply the symmetric-application test: would the piece’s author dismiss a comparable safety concern (a tobacco-style denial pattern, a pharmaceutical-skepticism pattern) with the same “panic” relabel? If not, the technique is deployed selectively on the deregulatory side, and the selective deployment is the news. Read the motte-and-bailey: the concession that the risks are real is what makes the dismissal safe to publish. Look for the same vocabulary (“superintelligence panic,” “AI doomerism,” “the safety theater”) across the syndication network — the vocabulary appears in coordinated cycles across the deregulatory coalition. And reduce the frame’s automatic activation: the “panic” label is the manufactured product, not the description; the documentary record on AI safety is the record the piece is calibrated to prevent the reader from encountering.

Close on witness. I am not a sympathetic reader of Jenkins’s piece. I worked the same operation on climate and on tobacco. The piece is built the way the pieces I built were built; the words are different, the structure is the same. The readers Jenkins is talking to are the readers I was talking to. I am bitter about the operation; the bitterness is the residue of the recognition that produced the reform, and the documentary record is the documentation the reader can verify. The readers Jenkins is talking to are people whose judgment I helped deform for two decades. I owe them the recognition more than the contumely. The recognition is the column’s service. The reader carries it forward.