They're Coming For You

Do you ever wonder who’s quietly building a dossier on your posts, purchases, movements, and even your face—and how that dossier could be used when you say something unpopular? In They’re Coming for You, former House Oversight chair Jason Chaffetz argues that America has entered a dangerous era where political power is fused with vast, opaque data pipelines that can be used to reward compliance and punish dissent. He contends that the combination of government agencies, activist nonprofits (NGOs), and tech/finance corporations—what he dubs a modern Cyclops—has learned to do together what each entity is limited from doing alone: surveil, sort, censor, and sanction everyday people at scale.

The book’s central claim is stark: unfettered data access has become the new currency of political power in the United States, turning privacy trade-offs into levers that shape elections, careers, and speech. The stakes, Chaffetz warns, are generational. If you don’t understand how your data flows—from loyalty apps and Ring cameras to Section 702 intercepts and facial-recognition databases—you can’t grasp how easily it can be repurposed to debank you, downrank your news, deny you work, or erase your voice.

The Core Architecture of Power

Chaffetz lays out a repeatable pattern: government accesses or acquires data (sometimes by purchase through brokers rather than by warrant), NGOs translate partisan priorities into "research" and pressure campaigns, and corporations operationalize the results through product policies, moderation, or account closures. This is the story behind Executive Order 14019 (Promoting Access to Voting), where agencies were tasked to register voters while working with “nonpartisan” third parties—later revealed, via unredacted DOJ notes obtained by watchdogs, to be overwhelmingly progressive groups such as the Leadership Conference, the Brennan Center, the ACLU, Black Voters Matter, and others. The goal, he argues, was to mobilize specific Democratic-leaning cohorts using government resources and data channels that campaigns themselves can’t legally tap (compare this to Seamus Bruner’s Controligarchs on elite-funded infrastructure).

He then shows the same template in finance. Treasury’s Financial Crimes Enforcement Network (FinCEN) asked banks and platforms like PayPal to hunt for “suspicious” political indicators—terms like “MAGA,” bulk cash withdrawals, travel to rallies, or even purchases of religious texts—using merchant codes and keyword scrapes. Government can’t surveil Americans’ constitutionally protected activity directly, but it can ask industry to do it and share the flags. That, Chaffetz writes, is how debanking metastasized from Operation Choke Point (Obama era) to Choke Point 2.0 (alleged under Biden)—squeezing gun stores, crypto startups, and even apolitical Christian charities like the Timothy Two Project and Indigenous Advance for “reputational risk.”

From Surveillance to Speech Control

The second arm of the Cyclops is narrative control. Chaffetz documents the birth of a "censorship-laundering" ecosystem: Obama-era moves (elections labeled “critical infrastructure”), the rise of CISA inside DHS, and the emergence of public–private censorship hubs like the Election Integrity Partnership and the Cyber Threat Intelligence League (CTIL). With the 2020–2022 “Twitter Files” as receipts, he shows how federal actors escalated content-removal/visibility filtering through tech portals, while a ratings cartel—NewsGuard and the State Department–funded Global Disinformation Index (GDI)—steered ad money and discovery away from right-of-center outlets (Federalist, Daily Wire, New York Post) and toward establishment media.

Chaffetz’s point isn’t to claim conservatives never make mistakes; it’s to show how the process punishes one side’s mistakes while shielding the other’s. He cites YouTube seeking White House feedback on moderation during COVID, Facebook admitting it downgraded true side-effect content at the Surgeon General’s push, and Amazon quietly adding “Do Not Promote” labels to disfavored vaccine books after administration pressure. In parallel, Google’s search/autocomplete “errors” and Go Vote reminders skewed toward left-leaning users in swing states (per Robert Epstein’s SEME research). The result was a systemic thumb on the scale of what you see and what sponsors support.

Why Facial Recognition and Data Brokers Matter to You

Much of your daily life emits data you barely notice—car telematics, location pings, loyalty swipes, DMV photos, cloud backups, and the faces you upload. Chaffetz shows how agencies short-circuit Fourth Amendment limits by buying commercial data they’d otherwise need a warrant to obtain (Sen. Ron Wyden’s disclosures about the NSA purchasing internet metadata are a prime example). He also revisits congressional fights over Stingrays (cell-site simulators), FISA 702, and the SEC’s Consolidated Audit Trail (CAT), warning that “solve-crime” pretexts can normalize bulk capture that later gets repurposed for political ends.

The facial-recognition chapter drives this home. Police-success stories using Clearview AI (40+ billion images scraped from the open web) are real—and so are the abuses when unregulated. Harvard students showed how a $300 pair of Ray-Ban Meta glasses plus a little code can dox a stranger in seconds using PimEyes and an LLM. Meanwhile, TSA expands airport face-match, the NFL rolls out Wicket access control at 32 stadiums, and the UK’s ULEZ camera backlash hints at what happens when surveillance aids behavior controls. If China is the dystopian endpoint—700 million cameras, gait recognition, and a programmable e-currency—Chaffetz’s question is: how far down this road do we slide before we blink?

What This Summary Covers—and Why It Matters

You’ll see how EO 14019 and agency–NGO partnerships rewire government for electioneering; how FinCEN emails and SPLC–ISD lists fuel data-driven debanking; how the FBI’s security-clearance process sidelined dissenting agents; how warrantless pathways—from Hemisphere to Section 702 purchases—build shadow dossiers; and how CISA, GARM, and ratings cartels funnel speech toward preapproved narratives. You’ll also see Chaffetz’s remedies: strengthen Inspectors General and FOIA teeth, pass the Fourth Amendment Is Not for Sale Act, end agency GOTV, protect financial civil rights at the state level, revive Schedule F and relocate agencies, and recruit private-sector innovators (Oracle, Apple, Starlink) to harden privacy/cyber infrastructure.

Big takeaway

“People who want power need data. It’s become far too easy to get.” The book is less about left vs. right than about process: when rules, oversight, and warrants are replaced by back channels and purchase orders, whoever holds the data can tilt the field—no matter who you voted for.

If you’re a parent, professional, donor, or just a citizen who values privacy and pluralism, this matters right now. The same pipelines that sort ad offers and boarding lines can sort livelihoods and liberties. Chaffetz’s message is not to abandon technology—it’s to reassert guardrails, transparency, and equal protection before the dossier economy becomes a de facto social credit system.

Chaffetz argues that Executive Order 14019 (Promoting Access to Voting) turned hundreds of federal entities into a sprawling, election-adjacent mobilization machine. On paper, the order was about removing barriers to voting and partnering with “approved, nonpartisan third-party organizations.” In practice, he says, it invited progressive-aligned nonprofits into federal agencies to target historically Democratic-leaning constituencies—leveraging taxpayer-funded channels and datasets that campaigns can’t legally access.

The Secret “Listening Session”

The smoking gun is a July 2021 White House “Listening Session.” After watchdogs sued under FOIA, unredacted DOJ notes surfaced showing invitees were overwhelmingly progressive: Wade Henderson’s Leadership Conference (a 240-group left-leaning coalition), the Brennan Center, ACLU, SPLC, Open Society Policy Center, AFL-CIO, NEA, Black Voters Matter, and others. There were “zero Republicans, independents, or politically conservative individuals,” Heritage’s Oversight Project concluded.

What did they discuss? According to the notes, groups pitched agency-specific targeting ideas. Demos suggested HUD use its in-house data to register tenants in public housing. The ACLU wanted Head Start sites as registration hubs. MALDEF floated limiting the enforcement of noncitizen voting bans to postelection prosecutions instead of up-front checks, making contests harder to unwind later. These are not academic hypotheticals; they became blueprints.

SBA, Education, HUD, Interior: Case Studies

• Small Business Administration (SBA): In 2024, the House Small Business Committee found that 22 of 25 SBA outreach events (Jan–Apr) took place in counties with the highest concentration of DNC-target demographics. Associate Administrator Jennifer Kim—whose background is civic engagement rather than business—repeatedly appeared in Michigan, where Rep. Lisa McClain displayed a map suggesting saturation in deep-blue corridors around Detroit and Flint. Administrator Isabel Guzman’s travel schedule similarly clustered in battlegrounds. When pressed in a July hearing, Kim called the effort “nonpartisan,” but could not explain the geographic asymmetry.

• Education Department: Five months before the 2024 election, Secretary Miguel Cardona blasted out emails and social posts to 40 million borrowers bearing the subject line “Biden-Harris Administration Takes Next Step Toward Additional Debt Relief…”—despite the Supreme Court’s 2023 smackdown of the plan. Since Democrats are more likely to hold student debt (YouGov: +10 points vs. Republicans, and more behind on payoff), Chaffetz argues the messaging functioned like microtargeted GOTV via federal lists.

• HUD: The agency encouraged 3,000+ public housing agencies to act as voter-registration sites, reversing long-standing guidance from 1996. Letters asked locals to consider accepting registrations and installing ballot drop boxes “if state law allows.” Given public-housing demographics, the net effect was predictably blue-leaning.

• Interior & Indian Affairs: FOIA emails show coordination with Demos and Indian Health Services to target Native American voters (crucial in Arizona). One now-walked-back plan would have sent registration materials home with Bureau of Indian Education schoolchildren. Spreadsheets tracked precinct-level results from 2016 and 2020—odd for a “nonpartisan” federal registration program.

Why FOIA and “Presidential Privilege” Matter

When the Foundation for Government Accountability and the Heritage Foundation asked to see all agency plans (as the law allows), DOJ ignored deadlines and then claimed “presidential communications privilege.” In court, it argued disclosure might cause “public confusion.” Chaffetz’s broader point: sunlight is the only meaningful check when agencies are told to plan political-adjacent work with outside groups, but the plans themselves are sealed until after the election.

Why this is new

Unlike “Zuckerbucks” in 2020—private money deployed through election offices—EO 14019 harnessed federal agencies, federal data, and federal infrastructure, then laundered the partisanship through NGO partners designated as "nonpartisan."

You don’t have to be a Republican to find this troubling. If government can selectively mobilize public housing, prisons, borrower lists, and federal workforces for partisan advantage, you’ve changed the playing field regardless of who the candidates are. Chaffetz’s prescription: ban federal agencies from electioneering, force preelection disclosure of any registration plan, and add penalties when agencies bury records behind dubious privileges. (For a complementary take on NGO-government fusion, see Mollie Hemingway’s coverage of EO 14019 and Bruner’s Controligarchs.)

What happens when your bank, processor, or crowdfunding site suddenly closes your account—and won’t tell you why? Chaffetz traces how “reputational risk” morphed from a prudential concept into a political weapon. His timeline runs from the Obama-era Operation Choke Point (2013–2015), which coerced banks to sever ties with legal but disfavored industries (payday lenders, gun dealers), to what venture capitalist Marc Andreessen dubbed “Operation Choke Point 2.0” under Biden, where crypto founders and cultural dissidents say they lost access en masse. Because the Bank Secrecy Act forbids banks from revealing Suspicious Activity Reports (SARs), you’re often left with a curt letter and no defense.

The FinCEN Playbook

Ten days after January 6, 2021, Treasury’s FinCEN emailed regulated institutions—including PayPal and Western Union—seeking help to find “violent extremists.” Attached was “Bankrolling Bigotry,” a report from the London-based Institute for Strategic Dialogue (funded in part by Open Society and others) that proposed a model to defund “hate.” The report’s hate-map leaned heavily on SPLC/ADL lists that lump mainstream conservative nonprofits—Family Research Council, Alliance Defending Freedom, the American College of Pediatricians, NumbersUSA—together with genuine bigots.

The Weaponization Committee later obtained evidence that FinCEN urged banks to flag transactions with search terms like “MAGA,” “America First,” and “Trump,” frequent ATM withdrawals, travel to rallies, and purchases of religious texts. This wasn’t probable-cause warranting; it was pattern-matching political belonging. From there, NGOs and media campaigns converted flagged activity into “risk” justifications for account terminations—completing the public–private loop.

Real People, Real Losses

• Christian charities: Timothy Two Project International (pastoral training in 65 countries) and Indigenous Advance (serving widows and orphans in Uganda) were summarily debanked by Bank of America. Letters cited vague “risk-tolerance” changes. Public pressure later revealed a broader pattern: Moms for Liberty found PayPal frozen; Lt. Gen. Michael Flynn had Chase cards canceled for “reputational risk” (later reversed); the Idaho Constitution Party’s account was closed by U.S. Bank without explanation.

• Selective zeal: House Oversight chair James Comer learned Treasury had over 170 SARs on Biden family members (money laundering, human trafficking, tax fraud allegations)—but no known debanking there. In contrast, former First Lady Melania Trump says her bank account was closed in 2021; her son Barron’s application was denied.

The IRS as a Data-Driven Lever

Chaffetz reminds readers that the IRS has already been caught targeting conservatives (2013 TIGTA report). Now add an $80 billion funding surge (Inflation Reduction Act), planned AI-driven audit expansions, and a Supreme Court ruling (Polselli) confirming broad bank-record access without notice. Whistleblowers allege DOJ interference in Hunter Biden’s tax probes; meanwhile, the IRS opened an investigation into the nonprofit that had exposed IRS bias (American Accountability Foundation). The concern isn’t audits per se; it’s whether who gets audited will be steered by political equity goals (America First Legal is suing to expose any race-based audit algorithm changes via Census imputation).

State-Level Guardrails

Because Congress moves slowly, Chaffetz spotlights state solutions. Florida now bars banks from discriminating based on political or religious beliefs, gun ownership, or ESG noncompliance. Tennessee passed bans on “social-credit” scoring. Texas, Louisiana, and West Virginia have taken steps to penalize banks that boycott fossil fuels or firearms. Shareholders are also filing resolutions demanding civil-rights audits of debanking practices (e.g., at JPMorgan Chase), using the same reputational logic the left leveraged.

The principle at stake

Financial access is the onramp to work, speech, and association. If NGOs and agencies can algorithmically brand you a risk, debanking becomes a silent blacklist—a punishment without process.

Chaffetz’s fixes: codify viewpoint nondiscrimination for core banking services (like public accommodations), overhaul FOIA enforcement so Treasury and DOJ can’t sit on records until after elections, and simplify the tax code so audit discretion can’t be gamed. You may disagree with a ministry or a gun shop; the question is whether their access to the financial grid should depend on their politics—or on clear, evenly applied laws.

Chaffetz contends that the federal workforce has increasingly functioned as a political monoculture, where dissenting views are chilled, sidelined, or purged—especially in law enforcement and national security. He doesn’t argue conservatives deserve special treatment; he argues equal treatment is vanishing, replaced by informal litmus tests and process slow-rolling.

Whistleblowers Without a Lifeline

Consider FBI agents Steve Friend and Garret O’Boyle. Friend raised protected disclosures about January 6 case-handling (use-of-force concerns; inflation of DVE stats) and was suspended without pay, initially denied permission to seek outside work, and blocked from his own training records. O’Boyle was approved for a cross-country transfer he says the bureau knew about—only to have his clearance suspended on Day One in the new office; his family’s winter clothes were stuck in storage he couldn’t access without $10,000. The FBI’s HR head Jennifer Leigh Moore told Congress the timing was a coincidence and not punitive.

The DOJ Inspector General later issued a damning memo: at least 106 FBI agents sat in clearance-suspension limbo for an average of 18 months, with no appeal path—conditions that “create the risk that the security process could be misused…to encourage an employee to resign.” Then came the leaked colleague questionnaire used in clearance reviews, with prompts like “Vocalize support for President Trump?” and “Vocalize objection to COVID-19 vaccination.” When pressed, Director Wray called the wording “completely inappropriate” and blamed a contractor—but whistleblower lawyer Tristan Leavitt obtained completed forms and agents say they were asked those questions.

From Extremism Purges to Vaccine Purges

After January 6, defense leaders launched a Countering Extremism Working Group led by Bishop Garrison, with advice from SPLC, ACLU, and ADL. Conservative veterans like Pete Hegseth warned the target looked more like “wrongthink” than violence indicators. The working group fizzled after pushback (as chronicled by Revolver News and Media Matters’ counter-narrative), but the vaccine mandate accomplished a similar purge by other means: more than 8,300 service members were separated before Congress ended the rule, with just 400 religious exemptions approved of 37,000 requests. Because skepticism skewed right in 2021 (Gallup, NYT/NPR/Atlantic coverage), the mandate’s effect tilted ideologically even if not explicitly designed for that.

Religious Tests, Soft and Hard

Religious affiliation is drifting into disqualifier territory. During judicial confirmations, then-Senator Kamala Harris questioned whether a nominee’s membership in the Catholic Knights of Columbus was “extremist.” In Idaho, a National Guard officer was stripped of command after a complaint that his private posts opposing puberty blockers for minors proved he “hates the LGBTQ community”—a case now in litigation with Liberty Counsel alleging a “No Christians as Commanders” policy. Justice Alito has warned about juror disqualifications based on orthodox belief in same-sex marriage cases.

Rebalancing vs. Shrinking the State

Chaffetz backs reviving “Schedule F”—an October 2020 executive order that would have reclassified policy-determining civil servants as at-will. Biden rescinded it and OPM went further in 2023, entrenching protections and adding vague “fitness” standards. Critics like James Sherk note that career staff often slow-walk or rewrite drafts to counter elected policy (e.g., the DOJ Civil Rights Division allegedly refusing to pursue anti-coercion cases for pro-life nurses). Chaffetz supports relocating agencies out of Washington to diversify the workforce and embed regulators closer to the people living under the rules. Long-term, he argues, the only durable fix is smaller scope: move powers back to the states and reduce the incentive to weaponize the center.

Bottom line

A merit civil service is supposed to insulate public administration from politics—not make it one party’s reserve army. Chaffetz’s reforms aim to protect real whistleblowing, end process-as-punishment, and restore the elected president’s ability to execute duly enacted policy.

If you work inside government, this chapter reads less like a partisan tract and more like a plea for consistent rules. If orthodoxy tests (political or medical) can cost you your badge, you don’t have a neutral civil service—you have a loyalty program.

Chaffetz catalogs a decade of end-runs around traditional Fourth Amendment safeguards. Rather than argue that all surveillance is illegitimate—he acknowledges crime-fighting benefits—he shows how agencies increasingly swap warrants for purchase orders, bulk “programs,” and nondisclosure gags that sidestep adversarial scrutiny. The worry isn’t capability—it’s oversight and scope creep.

Buy It Instead of Warrant It

Senator Ron Wyden revealed in 2024 that the NSA buys Americans’ internet browsing records from brokers—no warrant—based on the theory that “commercially available” data falls outside constitutional protections. The DOJ’s Hemisphere Project (rebranded “Data Analytical Services”) embedded AT&T analysts with drug units to give investigators access to decades of call records; funding was laundered through obscure grants to dodge public privacy reviews. The House even passed the bipartisan Fourth Amendment Is Not for Sale Act to close the broker loophole, but the bill stalled after the administration opposed it as “unworkable.”

Big New Pipes: SEC’s CAT and FISA 702

The SEC’s Consolidated Audit Trail (CAT) aspires to be the largest repository of securities data ever assembled, capturing every order, quote, and trade—down to retail investor IDs—for regulators (and thousands of personnel across 23 agencies). Proponents say it will expose naked shorting; William Barr warns it’s a hack magnet and an “evisceration” of privacy. Meanwhile, April 2024’s two-year extension of Section 702 (which Chaffetz notes Trump opposed with “KILL FISA” posts) broadened compelled compliance to data centers, giving any future administration more reach to incidentally sweep in Americans while targeting foreigners (Liza Goitein of the Brennan Center called the expansion a “gift” to presidents who might spy on enemies).

Twitter, Stingrays, and Mail Covers

Special Counsel Jack Smith obtained a warrant plus gag order compelling Twitter to hand over everything from Trump’s direct messages and blocked users to IP logs and search history—and data on every account that interacted with him during a set period (88.9 million followers by late 2020). Twitter balked; Judge Florence Pan held the platform in contempt and fined it $350,000; the Supreme Court declined review. Chaffetz frames this as precedent for dragnetting suspicionless Americans’ data whenever prosecutors chase a high-profile target.

He revisits cell-site simulators (Stingrays), which mimic towers to scoop up location and metadata from all nearby phones. After Chaffetz’s 2016 Oversight probe revealed DOJ/DHS had spent $95 million on the tech and routinely used NDAs to hide it from courts, agencies adopted warrant policies; several states (CA, WA, VA, UT) passed laws. Yet new models are stealthier, multi-band, and Wi-Fi-capable; without a federal statute, policies can be reversed at any time. Even the old USPS “mail cover” system lets law enforcement record exterior mail data with a simple request—no judge.

The Data-Broker Galaxy

A booming $389B–$561B global market (by 2029) traffics in location, purchase, and identity data, which agencies can buy. The CDC purchased cell-phone mobility data during COVID to monitor lockdown compliance; Chase is monetizing transaction history via its new media network; Facebook let Netflix peek at users’ DMs, and snooped on Snapchat traffic to map competitor usage. Apps from Grindr to Muslim Pro have sold sensitive location or status data, risking lives and livelihoods when reidentified (e.g., a Catholic staffer outed and fired after a media group obtained app-derived location pings).

Chaffetz’s standard

If the government would need a warrant to compel a company to hand it data, it should need one to buy the same data. Convenience is not a constitutional value.

For you, the action item is double-edged: push for laws like Fourth Amendment Is Not for Sale, and practice low-tech hygiene—limit app permissions, opt out of brokers, and prefer services with local (on-device) processing over cloud retention (Apple’s privacy stance is one model).

Facial recognition is no longer a lab toy; it’s everywhere—from your phone unlock to stadium doors to street cameras—and the dataset is your life online. Chaffetz weaves three threads: how easy doxxing has become, how rapidly live vision networks are scaling, and how governments (ours and others) can chain feeds together to reidentify you across contexts.

Doxxing with Smart Glasses

Harvard students AnhPhu Nguyen and Caine Ardayfio demonstrated a weekend project that should shake you: they patched Ray-Ban Meta smart glasses into PimEyes and a large language model (ChatGPT), then streamed to Instagram while their server scraped open-web images and assembled profiles. In seconds, they could greet strangers on the subway by name and profession and display addresses and phone numbers. The point wasn’t to popularize a tool—they withheld the code—but to prove that a curious student with $300 hardware can extract a person’s identity on sight. They even published instructions to remove yourself from major people-finders and advised credit freezes.

Clearview AI: Heroic or Horrific?

Clearview AI scraped 40+ billion open-web images (LinkedIn, Venmo avatars, Facebook, Google Images) to create what Time called one of the world’s most influential tools. Police have used it to catch predators and exonerate innocents; bans followed in Canada, the EU, and Australia; U.S. civil suits restricted its private-sector sales. CEO Hoan Ton-That points to reduced bias compared to smaller celebrity-trained systems because of the dataset’s breadth. Chaffetz’s takeaway isn’t to ban police usage; it’s to legislate clear, audited rules—warrants where appropriate, accuracy benchmarks, retention limits, and transparency about when and how it’s used.

Live Cameras as Control Surfaces

• Airports & Stadiums: TSA is rolling out face-match at checkpoints (claiming instant deletion of images), and the NFL adopted Wicket biometric access for 32 stadiums. In Las Vegas, police unions asked who would access those faceprints and whether activists could weaponize the feeds against officers.

• Cities: The 50 largest U.S. cities host over 537,000 cameras—11 per 1,000 residents—with DC at 55 per 1,000. Many traffic cameras are now AI-enabled, able to peer into vehicles. In London, the expansion of ULEZ low-emission surveillance triggered “Blade Runners” who disabled or removed hundreds of license-plate readers; critics fear mission creep into live face-tracking and behavior scoring.

• China’s Cautionary Tale: The CCP combines 600–700 million cameras with face and gait recognition, a digital yuan that can be turned off, and “ethnicity classifiers” allegedly used to track Uighurs. A video of toilet-time timers at a tourist site went viral—dark comedy on the edge of feature creep. Chaffetz’s refrain: technology does not come with a conscience; policy must provide one.

ReID: Stitching You Across Feeds

The ODNI’s IARPA program “Video LINCS” aims to autonomously identify and track a person or vehicle across multiple video sources—even after clothing changes—by learning unique signature patterns (reidentification, or reID). Imagine forensic power after a terror attack. Now imagine the same power enforcing lockdowns, climate zones, or protest penalties. Which use becomes normal depends on law first, not “trust us.”

Congressional oversight matters

In 2017, Chaffetz’s Oversight hearing forced the FBI to admit it had access (via MOUs with 18 states) to millions of DMV photos for face searches without publishing required privacy impact assessments. His admonition—“You’re required by law to put out a privacy statement, and you didn’t”—is the minimum standard we must revive across agencies and vendors.

For you, the move is twofold: demand laws (or city ordinances) that require warrants, audits, accuracy testing, and notice when face-matches are used against you; and reduce your public faceprint by removing images from major people-search sites. Tech won’t slow down; policy must catch up.

Chaffetz’s censorship chapter braids three systems: government “switchboards” that route takedown/visibility requests into platforms; NGO-led ad boycotts that starve outlets; and search/social algorithms that quietly tilt discovery. Together they create what he calls a “censorship-laundering” pipeline—government preferences actioned through private levers to dodge First Amendment constraints.

From Critical Infrastructure to CISA

In 2017, DHS designated elections “critical infrastructure,” expanding federal authority to monitor cyber threats (and, critics say, discussions around elections). CISA, created in 2018, morphed from cyber defense to narrative steward. During 2020, the Election Integrity Partnership served as a CISA-adjacent hub that flagged hundreds of millions of tweets for “misinformation.” The Twitter Files later exposed portal-based takedown workflows; Bari Weiss documented internal “visibility filtering” labels like “Do Not Amplify” for right-leaning accounts. Matt Taibbi published email threads where FBI, DHS, and even the White House sent content spreadsheets.

Rating the News, Throttling the Funds

NewsGuard (for-profit) and GDI (State Department–funded) graded outlets for “reliability.” Conservatives like the Federalist and Daily Wire scored low or landed on “riskiest sites” lists; establishment outlets with major corrections (e.g., Washington Post’s Steele-dossier walk-backs) still netted near-perfect ratings. Advertisers then fed these scores into “brand safety” tools organized by GARM—the Global Alliance for Responsible Media—whose internal communications, revealed by Congress, showed coordination to demonetize specific personalities (e.g., Joe Rogan) and platforms (X under Elon Musk). After X sued in 2024, GARM shuttered; DOJ’s incoming antitrust chief called the pattern “quite troubling.”

Platform “Errors” That Break One Way

A strange series of “mistakes” recurred in 2024: Google’s autocomplete returned “assassination attempt on Truman/the pope” but not on Trump; Facebook applied a doctored-image fact-check to all versions of the iconic post-assassination-attempt Trump photo; Google displayed “Where can I vote for Harris?” polling-place maps but not for Trump—later calling it an algorithm error. Meanwhile, Dr. Robert Epstein’s monitoring of “Go Vote” prompts found systematic over-targeting of left-leaning users in swing states. A Princeton study suggests pre-Musk Twitter’s moderation measurably reduced Republican vote share in 2016 and 2020 by shaping persuadable users’ exposure.

NSF-Funded Fact-Policing

House investigators found the National Science Foundation funded “convergence accelerators” to automate misinformation “interventions,” including MIT projects that profiled conservative media habits (e.g., preference for primary sources like the Bible and Constitution) to build counter-messaging. The goal: predict and preempt “inaccurate” narratives at scale. Chaffetz calls this what it is: federally backed speech sorting, often aimed at one side’s heuristics.

A hopeful wrinkle

Two events disrupted the system: Elon Musk opened X to real debate (and Community Notes), and House Republicans used subpoena power to expose the laundering. The result: more voters saw counter-narratives in 2024—and the censors lost the election anyway.

Whether you lean left or right, a society where a small, opaque consortium decides which facts you “need” is brittle. The practical fix isn’t to force platforms to carry everything; it’s to restore transparency (about government requests, algorithms, and ratings inputs), ensure equal-terms ad access, and block federal funding of domestic speech policing. (For deeper background, see Shellenberger/Taibbi’s CTIL Files and Jonathan Turley’s analysis of GARM.)

Chaffetz closes in on the culture front: how entertainment and K–12 programs aim not merely to inform kids but to format them—embedding political priors under the rubric of media literacy, social-emotional learning (SEL), and civics. The common thread is data capture plus activism cues, often underwritten by the same progressive foundations and government agencies he profiles elsewhere.

Animation with an Agenda

Animator Andrew Young (BYU, DreamWorks/PDI) describes his breaking point on Mr. Peabody & Sherman (2014). After a team screening, filmmakers explained the villain was modeled on “religious zealots” hostile to nontraditional families—lines had been lifted from the Heritage Foundation, the Catholic Church, and even the LDS Church’s Proclamation on the Family. A cut scene had the villain shouting “The family is the fundamental unit of society!” while hurling axes. Young asked that parents at least be told the film’s message; a supervisor allegedly replied that the problem was they weren’t being “deceptive enough.” Young quit, asked his name off the credits, and later worked at Project Veritas and the National Center on Sexual Exploitation. His larger claim: across hits from Frozen to The Lego Movie, fathers are buffoons unless gay, capitalism is suspect, and tradition/religion are marked for ridicule. (Even Disney CEO Bob Iger later admitted, “We have to entertain first…not about messages.”)

Media Literacy or Message Literacy?

Groups like the National Association for Media Literacy Education (NAMLE), bankrolled by the State Department and big tech (Facebook, YouTube, TikTok, Twitter), partner with Media Literacy Now to pass state mandates. NAMLE’s conferences feature sessions tying Black Lives Matter and climate activism or K–2 pronoun pedagogy; Newsela (in 90% of U.S. schools) rewrites establishment news to grade level and says its content is reviewed against SPLC’s Learning for Justice (“Identity, Diversity, Justice, Action”). In practice, Chaffetz argues, “media literacy” often means “here is the right lens to read with,” not “here is how to check claims across priors.” DHS-funded “harmony” games like Harmony Square or the Wilson Center’s proposed superhero disinfo game try to “inoculate” students by having them role-play disinformation spreaders—training a reflex rather than a habit of open inquiry.

SEL: From Skills to Ideology

CASEL coined SEL to teach self-regulation and empathy; in 2020 it introduced “Transformative SEL,” infusing intersectionality and “transformative citizenship” defined as actions to advance social justice goals “even if inconsistent with the law.” Vendors like Panorama Education (seeded by the Chan Zuckerberg Initiative; founded by AG Garland’s son-in-law) ask students how often they “confront issues of race/ethnicity,” whether their school helps them “speak out against racism,” and who they trust on race-related topics—while owning the data long-term. Chaffetz warns this looks like the beginnings of a behavioral credit score for children, with little parental control.

Civics as Activism

Action civics programs (e.g., Generation Citizen) backed by Ford, Soros-affiliated foundations, and Bezos-linked initiatives have students earn credit by building coalitions and protesting policy, with teacher “community partner” choices nudging outcomes. The Biden administration’s grant criteria privilege “culturally responsive” and “systemic bias” frames, aligning with the 1619 Project revivalism. Even COVID relief funds referenced the Abolitionist Teaching Network’s “antiracist therapy for white educators.” Chaffetz doesn’t deny that America has hard history; he argues that curriculum should teach how to self-govern, not just how to agitate.

Thomas Sowell’s warning

“The problem isn’t that Johnny can’t think. The problem is that Johnny doesn’t know what thinking is; he confuses it with feeling.” Chaffetz urges schools to re-anchor in knowledge, method, and pluralism—while parents demand visibility into content and data flows.

The practical takeaway: ask your district how Newsela, Panorama, or “media literacy” tools are implemented; request data-retention and sharing policies; and insist on ideological viewpoint parity. If the pitch is “neutral,” the vendor list and reading list should reflect it.

Chaffetz’s final move is constructive: you can’t uninvent AI, cameras, or data brokers—but you can restore process, transparency, and consequences. His remedies aim to make the gears visible, slow down abuses, and redistribute power outward to states and citizens.

Accountability Infrastructure

• Empower Inspectors General: Give IGs testimonial subpoena authority (so they can compel non-federal witnesses) and let DOJ’s IG probe department lawyers. Add enforcement teeth so agencies can’t ignore IG recommendations.

• Fix FOIA and Subpoena Evasion: Impose penalties for chronic delays and bad-faith redactions; designate an external enforcer when DOJ conflicts out (so Congress isn’t asking the executive to police itself). Publish centralized logs of government–platform takedown requests.

Legal Guardrails on Data

• Pass the Fourth Amendment Is Not for Sale Act: Close the broker loophole—if government needs a warrant to compel, it needs one to buy.

• Limit Agency GOTV: Prohibit federal registration partnerships unless states invite specific sites under the NVRA; require preelection publication and audits of any agency voter-outreach plan.

• Protect Financial Civil Rights: Enact state and federal bans on debanking based on lawful political/religious views or “social credit.” Codify transparency for account closures.

Rebalancing the Bureaucracy

• Revive Schedule F and Relocate Agencies: Make policy-advocating roles at-will; move agencies to the states to diversify hiring pools and align regulators with lived effects of rules.

• Narrow the Mission: Shrink federal scope per the Tenth Amendment; when fewer levers exist, fewer can be captured. (This echoes Chevron’s rollback by the Court and James Sherk’s call for presidential firing authority over insubordinates.)

Cyber and Private-Sector Solutions

• Security by Design: Support alt-architectures that minimize cloud exposure (Oracle’s TikTok “Project Texas” model; Apple’s on-device processing posture; Starlink’s decentralized backhaul). Explore data-virtualization and content-signature models that verify integrity without centralizing raw data.

• Deepfake Mitigation: Fund watermarking and provenance standards; criminalize malicious synthetic defamation; require social platforms to label and rapidly adjudicate.

• Kids First: Encourage state experimentation with age-verification that preserves anonymity (tokenized proofs) and voluntary “right-to-be-forgotten”-style features for minors. Demand K–12 vendor contracts disclose data retention and third-party sharing.

The mindset shift

Don’t seek a freeze on tech; seek countervailing power: state laws, IGs with teeth, FOIA with deadlines, and market solutions that compete on privacy by default. As Chaffetz puts it, “sunlight remains the best disinfectant.”

Finally, bring “the nerds” into Congress. Chaffetz’s SOPA-era quip still applies: lawmakers too often legislate the internet like surgeons without medical training. Standing up a bipartisan technical review board for high-impact bills (surveillance, AI, platforms) would help ensure liberty and security are engineered—not wished—into law.