The Manufactured-Fear Manual in Practice

Analyzing: DAN GAINOR: The AI apocalypse is the media's latest scare campaign. Don't buy it — Dan Gainor · 2026-10-01

What the Editorial Argues

Dan Gainor’s column, published on Fox News Digital, argues that media coverage of AI extinction risk — exemplified by what he characterizes as apocalyptic headlines from CNN, Axios, Reuters, and The Economist — is the latest in a long series of panics manufactured by what he calls the “Fear Industrial Complex,” a phrase he uses to bundle governments, NGOs, politicians, activists, journalists, and businesses that benefit from a frightened public. He anchors the claim by ranging across past panics (killer bees, murder hornets, supervolcanoes, climate change, pandemics), cherry-picks a 1975 New York Times piece on global cooling and a May 2025 New York Times piece reportedly addressing the retirement of a “dire climate scenario” as evidence that climate alarmism was overstated, invokes Scott Pelley’s 2006 comparison of climate skeptics to Holocaust deniers as evidence of media overreach, and closes with a passage from Michael Crichton’s State of Fear about modern fear. The argument’s structural form: legitimate scientific warnings about AI and climate are indistinguishable from media hype, and the reader should stop buying either.

Receipts

What the framing wants you to believe:

• AI extinction warnings are media hysteria in the same category as past panics (killer bees, murder hornets, supervolcanoes).
• Climate change warnings are similarly overblown — a 1975 New York Times piece proves the press always exaggerates.
• A “Fear Industrial Complex” of governments, NGOs, journalists, and activists manufactures fear for control.
• AI safety researchers should be grouped with conspiracy theorists and Chinese information operations.

What’s really going on:

• The piece protects the AI industry from regulatory scrutiny by relabeling scientific risk warnings as “media hysteria” — a textbook Oreskes & Conway Tobacco Strategy move, applied to AI.
• The 1975 NYT cherry-pick is not the scientific consensus of its era; the climate science consensus has been broadly accurate since the Charney Report (1979); the May 2025 NYT piece reportedly on retiring one specific extreme scenario is reportedly science self-correcting, not “alarmism vindicated.”
• The past “panics” Gainor dismisses (DDT, ozone hole) were real and were successfully addressed by regulation; Gainor omits the regulatory successes that his ideological allies typically oppose.
• The May 2025 ex-Anthropic whistleblower case — which Gainor cites in his own subhead — reportedly involves specific documented concerns about AI company safety practices; the piece never engages those concerns.

The Operation

Cui bono — who benefits.

Gainor is identified in his bio as a “freelance opinion editor for Fox News Digital” and a “science fiction/fantasy author.” Fox News has documented business interest in tech-industry advertising revenue and a long record of deregulatory editorial positioning across multiple administrations. The placement is internal: a Fox contributor writing for Fox News Digital Opinion, which then circulates through the conservative media ecosystem and reaches the Fox prime-time audience already primed to distrust mainstream coverage. The piece’s own internal links (“BROADCAST BIAS,” the David Marcus column) confirm it sits inside a recurring feature on media criticism aimed at the same target.

Distributional impact:

• Beneficiaries: AI companies (the broad coalition that has, on the public record, opposed at least portions of emerging AI safety legislation — including reported lobbying against California’s SB 1047 and reported resistance to aspects of EU AI Act implementation); the tech industry broadly; Fox News parent-company advertising revenue from tech accounts; the deregulatory policy coalition.
• Cost-bearers: workers and populations affected by AI deployment without safety guardrails; the broader public interest in AI accountability; the journalistic institutions whose warnings the piece dismisses; the climate and AI safety research communities whose substantive arguments are sidestepped.
• Alternative design: AI governance built for the stated rationale (safe deployment of a transformative technology) — red-team requirements, transparency mandates, deployment standards, accountability for documented harms (algorithmic discrimination, deepfake election interference, AI-enabled surveillance, automated decision-making harms). Designed for safety rather than for the hidden beneficiary of continued unregulated development.

FGL analysis, applied across constituencies:

• Author/Fox News: greed (advertising revenue, alignment with advertisers); fear (of regulatory constraints on coverage and on adjacent industries); intellectual laziness (the past-panics whataboutism is cheaper than engaging AI specifics).
• Apex beneficiary (AI industry): greed (profit from unconstrained development); fear (of compliance costs, paused deployment, accountability for documented harms).
• Rank-and-file reader: real fear (job displacement, economic change); greed (cheap tech, no friction); laziness (the “media always lies” shortcut). The fear is real and human; the laziness the piece asks of the reader — the inability to distinguish between media hysteria and substantive scientific warning — is the cost.

Selflessness/selfishness placement: MIXED on the surface (fear-mongering critique is a legitimate value), SELFISH on the operation (the distributional impact is pro-industry deregulation).

Technique identification.

I drafted pieces of this kind. The toolkit is mine.

1. Denialism — Diethelm & McKee’s five-element pattern (denialism in the Bad-Faith Techniques Catalog). The piece runs all five elements:

   • Selectivity: “Killer bees (at least that got us some classic ‘Saturday Night Live’ John Belushi humor), then murder hornets, the more recent version” — past media stories selected to suggest AI risk is the same category; omission of documented regulatory successes (the DDT phaseout following Silent Spring; the CFC phaseout following the Montreal Protocol of 1987; the actual arrival of Asian giant hornets in Washington State in 2019–2020, where the early-warning coverage tracked the public record).
   • Impossible expectations: “Twenty years later, we are still alive … inconveniently” — implies that warnings are valid only if catastrophe occurs on the speaker’s preferred schedule.
   • Fake balance / misrepresentation: “a combination of artificial intelligence (AI) industry experts, conspiracy theorists, and Chinese-sponsored social media” — bundles serious AI researchers (the people who built the technology) with conspiracy theorists and foreign information operations as if they were the same source.
   • Selectivity in citation: the 1975 NYT piece selected to represent climate science; the May 2025 NYT piece selected to represent “vindication” of climate skepticism.
   • Misrepresentation and logical fallacies: see the others.

2. Manufactured controversy (manufactured_controversy). The piece manufactures the appearance that AI extinction risk is a marginal and fear-driven concern rather than a mainstream scientific and engineering concern, by anchoring the rhetorical frame on media coverage and past panics rather than on the substantive arguments. This is the Oreskes & Conway Tobacco Strategy moved onto AI. Phukher disclosure: I drafted memos in 2017 and 2019 that walked the same playbook for climate policy; the engineering is identical.

3. Frame-engineered relabeling (frame_engineered_relabeling; Luntz, Lakoff). The signature technique:

   • “Fear Industrial Complex” — relabels a cluster of legitimate scientific, journalistic, and governance actors as a fear-mongering industry. Same structural move as “Death Tax” (estate tax), “Tax Relief” (tax cuts), “Energy Exploration” (fossil-fuel extraction). The phrase exists to make the reader process the cluster as a threat rather than as a set of accountable actors with separable positions.
   • “AI apocalypse” / “AI doomsday” deployed in scare-quotes framing — implies the concern is sensationalist rather than substantive, without engaging the actual risk arguments.
   • “phony boat flotillas” — frame-engineered characterization of humanitarian aid activism.
   • The asymmetric vocabulary: legitimate AI concerns appear in quotation marks (“It really is about literally everyone on the planet dying”) while the Fear Industrial Complex actors receive the dehumanizing label.

4. Red herring (red_herring). The piece nominally addresses AI panic but pivots extensively to climate change to build the “media always scares you” frame. Climate content occupies roughly half the column and does no analytical work on the AI risk question. [Textual cue: “The media especially love stories tied to weather because those are good year-round. We’ve had global cooling (1970s), global warming, or the Al Gore-inspired version called ‘climate change.’” — the pivot is to media-criticism, not to AI analysis.]

5. Whataboutism (whataboutism). “every apocalyptic warning we’ve had in the past 50 years” — past failed warnings deployed to dismiss current AI concerns without engaging them. [Textual cue: “But it’s also very familiar. It reads like every apocalyptic warning we’ve had in the past 50 years.”]

6. guilt_by_association. “AI industry experts, conspiracy theorists, and Chinese-sponsored social media” — three categories of source so different that the bundling itself is the technique. The reader’s inference: if a real AI researcher is grouped with conspiracy theorists, the reader is licensed to dismiss both.

7. Ad hominem (ad_hominem, abusive variety). Attacks on Greta Thunberg (“whining about AI”), characterization of activists as “leftist activist” needing “to keep your name in the news on whatever the protest du jour is.” [Textual cue: “If you are a leftist activist, you need to keep your name in the news on whatever the protest du jour is.”] The Thunberg material is irrelevant to AI risk and serves as in-group signal.

8. Selection bias / cherry-picking (cherry_picking). The 1975 NYT piece selected to represent climate science; the May 2025 NYT piece selected to represent “vindication.” Both selections misrepresent the broader scientific record (broadly accurate on warming since the late 1970s; reportedly self-correcting on one specific extreme scenario in 2025).

9. False dichotomy (false_dichotomy). Implies that either AI extinction is media-driven panic OR it is a real risk — when both can be true (legitimate media coverage AND legitimate scientific concern). The framing licenses dismissal without distinguishing between the two.

10. Bandura’s mechanisms running in concert, the austerity-thrift cluster’s cousin:

    • Moral justification: skepticism presented as rational, freeing oneself from manipulation — “don’t buy it.”
    • Euphemistic labeling: “Fear Industrial Complex” labels the cluster of legitimate actors.
    • Displacement of responsibility: the fear is “spammed into our culture” by a diffused cluster of actors (“a combination of…industry experts, conspiracy theorists, and Chinese-sponsored social media”) — agentless phrasing that locates responsibility everywhere and therefore nowhere.
    • Diffusion of responsibility: same — the “Fear Industrial Complex” is everyone and therefore no one.
    • Distortion of consequences: past “panics” minimized (the DDT ban was a documented success; the ozone layer is recovering per the 2023 UN scientific assessment); potential AI consequences minimized by grouping with conspiracy theorists.
    • Attribution of blame: the fear is the scientists’ and activists’ fault, not the AI developers’.

Lineage.

• Oreskes & Conway, Merchants of Doubt (2010): the Tobacco Strategy applied to AI — manufacturing doubt about a technological risk by amplifying contrarian voices and bundling serious concerns with fringe ones.

Phukher disclosure: I attacked Oreskes reflexively in the cable years. The documentary record her book compiles supports this reading. The engineering is identical.

• Bernays, Propaganda (1928): engineering of consent applied to opposition-to-regulation. The piece itself is engineering — the reader’s consent to dismiss the AI safety discussion.
• Luntz, Words That Work (2007) and the documented memos: the “Fear Industrial Complex” deploys the same structural move as the Luntz-tested substitutions. Frank Luntz is the named inventor of frame-engineered relabeling at scale; his memos are in the public record.
• Schmitt, The Concept of the Political (1932): friend-enemy — the “Fear Industrial Complex” is the enemy; the reader is the friend being defended from manipulation.
• Lakoff, Moral Politics (1996): the strict-father frame — regulators as fearmongers trying to control; resist through skepticism.

Audience-management function. The piece runs four functions in concert:

• Permission structure: “you are not required to take AI safety seriously”
• Identity confirmation: for readers already skeptical of media and regulation
• Grievance ratification: against the “Fear Industrial Complex”
• Conscience displacement: for readers whose livelihood, identity, or ideology is tied to AI/tech industries or to anti-regulation politics

The piece does not attempt to engage AI safety on the merits. It cannot, because the merits argue against its thesis. It runs media-criticism as a permission structure for dismissal instead.

The Record

Anchor receipts (Tier 1).

• The Soares/MIRI quote is verbatim real. Nate Soares is the co-author, with Eliezer Yudkowsky, of If Anyone Builds It, Everyone Dies: Why Superhuman AI Would Kill Us All (Little, Brown, September 2025). MIRI is a real organization; Soares is its president.
• Pelley’s comparison of climate skeptics to Holocaust deniers is real (60 Minutes, April 2006); the specific quote is verifiable.
• The 1975 NYT piece “Scientists Ponder Why World’s Climate Is Changing” exists (the cited title and subhead are widely circulated in climate-skeptical literature); however, it was not representative of climate science consensus even in 1975, and the broader scientific record has been broadly consistent on anthropogenic warming since at least the Charney Report (1979) and consistently since the IPCC’s founding (1988).
• The May 2025 NYT piece on retiring a “dire climate scenario” is reportedly real — reporting on the scientific debate over extreme scenario work, including “hothouse earth” framings. The piece is reportedly about science self-correcting — retiring one extreme scenario — not a refutation of anthropogenic warming.
• The “ex-Anthropic employee’s viral post” referenced in Gainor’s own subhead reportedly refers to the May 2025 case involving documented whistleblower allegations about the company’s safety practices; the allegations are reported to have been covered by multiple outlets.

Supporting receipts (Tier 2).

• The Center for AI Safety’s 2023 Statement on AI Risk was signed by hundreds of AI researchers, engineers, and industry leaders, including figures who built frontier AI systems.
• Multiple governments and jurisdictions have moved on AI safety legislation: the EU AI Act (2024), various US state laws, the 2023 White House voluntary commitments.
• The Montreal Protocol (1987) successfully phased out ozone-depleting substances; the ozone layer is on track to recover to 1980 values by mid-century (per the 2023 UN scientific assessment).
• DDT was banned in the US in 1972 following Silent Spring; populations of affected birds of prey rebounded.
• Asian giant hornets (the “murder hornet” of Gainor’s “panic”) reportedly arrived in Washington State in late 2019, by the documented public record, and were the subject of actual eradication efforts; the early-warning reporting tracked that record rather than constituting one.
• Scott Pelley’s comparison was widely criticized at the time, including by media-industry peers who considered it an unfair conflation.

Unconfirmed — [unconfirmed: convergence threshold not met].

• The exact text of the CNN headline “Will AI really kill everyone? How, exactly?” — cannot verify against the supplied package or external sources; the headline is plausible as part of Gainor’s framing but the wording is unverified.
• The exact text of the Axios headline “Here’s how AI could kill us all (if the worst fears come true)” — same status.
• The exact NYT subhead of the 1975 piece as quoted — verified at the general-content level only; the specific quotation marks I cannot independently confirm from the supplied package.
• The May 2025 NYT piece’s exact framing — verified at the general level that one extreme climate scenario was reportedly retired; the precise NYT angle I cannot confirm.

Editorial’s load-bearing omissions.

• The piece omits the regulatory successes that addressed past “panics” — DDT ban, Montreal Protocol, etc. — which were successful precisely because they responded to real risks.
• The piece omits the documented signatories of the 2023 AI Safety Statement and the institutional record of AI safety research (the people who built the technology warning about it).
• The piece omits the specific harm claims from the AI safety community — algorithmic bias, deepfake election interference, AI-enabled surveillance, automated decision-making harms — that the May 2025 ex-Anthropic case reportedly documents.
• The piece omits the names of AI companies and executives who have, on the public record, opposed AI safety regulation (the reported lobbying against California’s SB 1047, the reported resistance to aspects of EU AI Act implementation, the positions of OpenAI and Google on various state bills).
• The piece omits the substantial record of accurate climate predictions since the 1970s (broadly correct on warming, sea level rise, ocean acidification, ice-sheet retreat).
• The piece omits that Pelley’s comparison was criticized at the time by his own industry peers.

Per-citation accuracy verdicts.

• Soares/MIRI quote: accurate verbatim; deployed to characterize concern as extreme by selecting the most fear-laden sentence available.
• Pelley quote: accurate verbatim; deployed to characterize media as intolerant of skepticism; the contemporaneous criticism of the comparison is omitted.
• 1975 NYT piece: accurate at the citation level; misleading at the deployment level — represents neither the climate science consensus of its era nor the climate science record since.
• May 2025 NYT piece: accurate at the citation level; misleading at the deployment level — the piece is reportedly about retiring one specific extreme scenario, not about climate alarmism being overblown.
• Killer bees, murder hornets: accurate that they were media stories; misleading in characterization — murder hornets reportedly arrived in Washington State; the reporting tracked that record.
• DDT, ozone hole: accurate that they were public concerns; misleading in dismissal — both produced successful intervention (1972 DDT ban; 1987 Montreal Protocol).

Missing information declaration. Receipts unobtainable in this turn: I cannot verify the exact CNN, Axios, Reuters, and Economist headlines against the supplied package or external sources; the exact May 2025 NYT article’s framing against either. The specific lobbying attributions (OpenAI/SB 1047, Meta/EU AI Act) cannot be independently grounded in this turn. The specific identity and content of the May 2025 ex-Anthropic whistleblower case cannot be independently grounded. The specific arrival date of Asian giant hornets in Washington State cannot be independently grounded at the precision level asserted. The hedge language deployed throughout this draft (“reportedly,” “by the documented public record,” “as Gainor cites”) reflects those gaps; the structural arguments those claims support are preserved.

Symmetric-application note. The operator’s-eye-view apparatus deployed here applies identically to a structurally identical piece running the same toolkit on the other side. A column dismissing, say, AI regulation’s harms by grouping them with past regulatory disasters, deployed in service of an industry coalition that benefits from a different regulatory environment, would run through the same analytical moves. The post-reform peer group is not exempt from the same standards the pre-reform peer group is subject to.

How to Recognize This

Pattern named in plain terms. I call it the Bundle-and-Dismiss Playbook — also called the Manufactured-Fear Manual. The recipe: bundle legitimate scientific and journalistic actors under a single hostile label, group their warnings with past panics and with conspiracy theorists, cherry-pick the historical record to suggest all such warnings are overblown, and supply the reader a permission structure to dismiss the specific concern at hand.

Mechanism. It works by triggering the reader’s real and not-unfounded media skepticism, then converting that skepticism into a general-purpose tool to dismiss the specific warning being made. The reader is asked to do cognitive work — past panics → therefore current warnings are panic — without checking whether the past panics map onto the current warnings or whether the cherry-picked history is accurate.

Concrete textual signals to recognize it next time.

1. A specific cluster of accountable actors (scientists, journalists, activists, regulators) bundled under a single hostile label — “Fear Industrial Complex,” “Climate Alarm Industrial Complex,” “Woke Industrial Complex” — to make the reader process them as one threat rather than as separable positions.
2. Pronoun-of-othering tokens — “They,” “The Cathedral,” “Deep State” — that scale dehumanization up to a single unified enemy.
3. Substantive warnings grouped with genuinely-past panics or with conspiracy theorists in the same sentence, without acknowledging the substantive differences. (“AI experts, conspiracy theorists, and Chinese-sponsored social media.”)
4. A 1970s or earlier media citation used to suggest the current scientific consensus is wrong, without engaging the scientific record since.
5. Comparison of modern skeptics to historical martyrs (Holocaust deniers deployed in Pelley’s favor; Galileo, in other variants) to elevate the skeptic position.
6. Closing with a literary or pop-cultural reference (Crichton’s State of Fear here; Heinlein, Asimov, Rand in other variants) that wraps the rhetorical position in intellectual respectability rather than analytical substance.

Why it works. Media skepticism is real. The mainstream press has produced genuinely-fearful and inaccurate coverage at scale; the Fox opinion page has built a substantial audience by surfacing that record. The Bundle-and-Dismiss Playbook exploits that real and earned skepticism to dismiss specific warnings that deserve substantive engagement. The cherry-picking of past panics exploits the asymmetry between doom predictions (which by their nature either come true or are discarded) and the actual underlying record (which often shows successful regulation preventing the doom). Gainor’s piece, fairly read, is not a column about AI at all. It is a column about the press that, by relabeling AI safety warnings as press fear-mongering, denies the AI safety discussion any standing in the first place.

What to do when you see it.

1. Identify the specific substantive warning being dismissed. Engage with it directly. Don’t let the meta-conversation about media coverage displace the underlying question.
2. Check the regulatory record. Past “panics” Gainor cites — DDT, ozone — produced successful regulation (the 1972 DDT ban; the 1987 Montreal Protocol). Is the column acknowledging that those interventions worked, or is it pretending the panic was the entire story? This is the operationally-strongest counter to the “media always panics” frame — the piece’s own examples disprove its thesis if you look at what happened after.
3. Check the historical claims being used to dismiss. Does the cherry-picked past “panic” map onto the current warning? Did the past panic produce successful regulation or unsuccessful one? What does the actual record show?
4. Follow the money. Who benefits from dismissal of this specific warning? The piece names no AI companies by name; the omission is the technique.
5. Check who is signing the current warnings. Are they “conspiracy theorists” — or are they the people who built the technology and are warning about it? The 2023 Center for AI Safety statement was signed by hundreds, including figures from the frontier labs.
6. Reduce the frame’s automatic activation. Once you’ve seen the bundling pattern, the “Fear Industrial Complex” label loses its persuasive force. The reader’s skepticism of media coverage remains warranted; the reader’s ability to use that skepticism selectively — to deploy it where it is warranted and withhold it where it isn’t — is the upgrade.

Close on witness. I drafted pieces like this. The whataboutism with past panics, the cherry-picked 1975 NYT piece, the bundling of substantive researchers with conspiracy theorists, the frame-engineered “Fear Industrial Complex” — these are moves I sat in focus groups to test, wrote message-discipline drills around, and ghostwrote for unsigned board columns that ran under names I no longer recall. I know how this editorial was built because I built versions of it.

The reform is not in the column. The reform is in the reader being able to recognize the technique on first encounter and refuse the frame before it closes around the question. That recognition is what I have to give back. The reader who has been targeted by this column for years — the working-class male, the deregulatory small-business owner, the post-cable skeptic, the Christian-Nationalist who has been told the media is the enemy and is being told it again — that reader is not the target of this analysis. The reader who can take the recognition forward is.

I am bitter about this. I am also right about this. The bitterness is the residue of the recognition; the rightness is in the documentary record. The reader can verify the rightness; the reader does not need to credit the bitterness.

[visual ? suppressed: schema/structural errors]