Deans of public health, senators, tort lawyers, and foundation heads do not compete for authority by admitting they want power. They compete by invoking languages of child safety, evidence-based governance, and protecting the vulnerable. These are not descriptions of what they do. They are access keys. They open budgets, hearings, grants, verdicts, and new administrative domains. In the youth mental health panic of the 2020s, these vocabularies have become something more: a compression engine that takes messy, multifactorial adolescent suffering and reduces it to one clean, actionable claim. Big Tech’s addictive design is the dominant cause. Once that claim stabilizes, everything else falls into line behind it.
\nThe engine starts with careers. A junior researcher in a respected lab learns fast what survives peer review and what dies. A paper arguing that effects are small, heterogeneous, and mostly downstream of family instability is hard to fund and impossible to translate into policy. A paper that tightens the causal story, cites Jonathan Haidt and Jean Twenge, and frames the problem as a design defect becomes legible to journals, journalists, and Hill staff. Tenure tracks compress reality long before Congress sees it. The selection is not corrupt in any simple sense. It is structural. Over time it produces a filtered pool of people who have convinced themselves, as Robert Trivers would predict, that the simplified story is the true one. That self-deception is load-bearing. It lets the entire regime run on genuine moral energy.
\nFollow the money on a calendar and the mechanism becomes visible. A large jury verdict lands in March. Plaintiff firms scale intake and advertising in April. Foundations reallocate portfolios toward youth digital harm by early summer. Federal agencies issue new grant calls by fall framed around platform risk mitigation. The lag is not incidental. It is the metabolism of the system. Courtrooms convert narrative into cash and legitimacy. Grants and policy convert that legitimacy into durable programs. Each cycle increases the number of people whose careers depend on the same causal story holding.
\nInsurance then turns soft norms into hard constraints, bypassing the democratic process entirely. Once juries attach nine-figure liability to specific platform features, underwriters demand best practices. Those best practices are written by the same expert networks supplying testimony and white papers. Compliance vendors appear. Certifications follow. A mid-size company now needs a youth safety officer, an external audit, and documented friction in its product design. The Delaware court ruling of March 2026, holding that Hartford and Chubb have no duty to defend Meta because addiction allegedly flows from deliberate design choices rather than accidents, stripped tech companies of coverage and handed experts veto power over product roadmaps. No statute required. The liability environment wrote the policy.
\nThe legal system also functions as a content factory. Discovery produces emails and A\/B tests. Internal debates become exhibits. In the hands of a courtroom performer like Mark Lanier, they become a story a jury can act on in an afternoon. The verdict then circulates into media and hearings as if it were a scientific conclusion. A $6 million award in Los Angeles or a $375 million judgment in New Mexico does not merely compensate. It anchors the public story. From there the distance to statute and standard is short. No one in a future trial wants to explain to a board why they retained a feature that just cost a competitor a third of a billion dollars.
\nThe reason this narrative holds against competing explanations is an epistemic asymmetry that nobody designed but everyone benefits from. The variables that plausibly matter most are the least tractable to regulation. Family structure, assortative mating, baseline temperament, neighborhood effects, and the aftershocks of COVID school closures do not map cleanly onto levers a senator can pull. Platform features do. Infinite scroll, autoplay, algorithmic push notifications, age gating. The system gravitates toward what it can act on, then backfills causation to match actionability. The proxy becomes the problem because the proxy can be governed. Raw multifactorial reality arrives at the policy layer already filtered: twin studies showing 40 to 60 percent heritability of anxiety and depression, evidence of family structure collapse, post-lockdown learning loss data, and correlational screen-time spikes all enter the pipeline together. By the time the material reaches a jury or a Senate hearing, the uncertainty has become a confident verdict that addictive features caused the harm.
\nErnest Becker helps explain the emotional adhesive. Elite institutions run on hero systems. The highest-status move inside those institutions is not solving a problem that resists easy solution. It is being seen by other elites as protecting the vulnerable. That pays out in New York Times bylines, testimony slots, Aspen panels, advisory roles, and invitations into rooms where policy is drafted. Quietly saying that this is multifactorial and our leverage is limited is low-status behavior. Saying we know what drives this and we must act is high-status behavior. The moral vocabulary supplies the theater. The career ladder supplies the incentives. Together they produce a system whose participants experience their work as protection and whose critics sound, from inside the system, like apologists for corporations that profit from children’s pain.
\nThere is also competition inside the elite that sharpens rather than complicates the story. Tort lawyers prefer a frame that expands liability. Regulators prefer a frame that expands rulemaking. Public health prefers a frame that expands surveillance and intervention programs. Each faction has a jurisdictional sweet spot. The Haidt narrative satisfies all three simultaneously: courtroom-friendly, regulator-friendly, and media-friendly. That triple utility explains why it dominates even as longitudinal evidence, including Candice Odgers’ 2024 analysis in Nature and multiple studies showing small or null effects once confounders are controlled, continues to accumulate against the strong causal version of the claim.
\nInternational policy creates a ratchet. The UK’s Online Safety Act and the EU’s Digital Services Act become proof of concept. U.S. advocates cite them to lower the rhetorical cost of domestic action. Multinationals adopt global compliance frameworks, which then normalize the standards at home. A design norm written in Brussels shows up in a product sprint in San Francisco. Age-verification vendors like Yoti, which has raised over $116 million, position themselves as the passport offices of the internet, charging between ten cents and two dollars per check and building a multi-billion dollar friction tax on digital interaction. The compliance market produces its own constituency of investors, consultants, and credentialed gatekeepers who need the regulatory regime to justify their existence.
\nDissent does not disappear. It gets filtered. Researchers who emphasize heritability, family context, and heterogeneous effects publish in lower-prestige venues, shift topics, or exit the field. What remains looks like consensus. That consensus gets cited in hearings as settled science. Journals, funding bodies, and institutional review boards all face the same selection pressure: research that does not advance the actionable narrative is labeled unproductive or risky. The range of reasonable disagreement narrows, which then gets presented as confirmation that the science is clear.
\nJuries close the loop. They are not asked to weigh multifactorial causation under controlled conditions. They are asked to choose between narratives under moral time pressure. Plaintiffs offer a vivid, singular cause: they engineered addiction to profit from children. Defendants offer a probabilistic, complex account that sounds, in a courtroom, like evasion. The legal forum structurally favors compression. Once a jury finds negligence in a specific design feature, that feature becomes effectively illegal across the industry without any new statute. The courtroom has replaced the laboratory as the arena where claims about harm get certified.
\nStephen Turner’s account of Democracy 3.0 names the political logic. We have shifted from citizens electing representatives to deliberate on values toward a system where complex social realities get delegated to expert bodies claiming superior information. When the political is pushed into the realm of expertise, the cost of dissent rises sharply. A citizen who disagrees is no longer a person with different values. She is uninformed or, worse, in the pocket of Big Tech. The expert class enjoys this arrangement because it insulates their authority from democratic accountability. More laws, more regulation, more need for expert guidance: each expansion of jurisdiction is presented as protection for the vulnerable rather than as a transfer of power upward.
\nThe feedback loop runs cleanly. Lawyers win verdicts. Insurers pull coverage. Experts write the safety standards required for new coverage. Tech companies adopt those standards to stay in business. Citizens are nowhere in this sequence except as the moral justification for each step. The political question of how a society should raise its children has been successfully converted into a technical insurance problem managed by a small caste of people who benefit from its continued complexity.
\nWhat would falsification look like. It has to be concrete. If large-scale interventions targeting platform features do not move pre-registered mental health outcomes within two years, the causal story should weaken. If they do not, but funding and regulation continue to expand, the system has decoupled from feedback entirely. If researchers who correctly predict null effects continue to lose grants and promotion despite predictive accuracy, selection has overtaken truth-seeking. If insurers and compliance regimes keep ratcheting requirements while adolescent outcomes remain flat, the liability tail is wagging the policy dog.
\nNone of this requires bad motives. Many participants care about teenagers with genuine intensity. That is what makes the system powerful and difficult to dislodge. The danger is that the compression infrastructure becomes self-sustaining, smoothing away the parts of reality that do not fit the levers the expert class can pull. The gap between what helps adolescents and what can be regulated widens quietly, then all at once.
\nReality will arbitrate. Not in hearings or on panels, but in cohorts. The teens who live with the actual causes of their suffering never read the expert briefs. They simply experience the distance between what the menu promised and what the intervention delivered. The vocabulary will remain intact. The careers will advance. The jurisdiction will expand. The only question that finally matters is whether the map matches the territory. The system will not answer that question honestly on its own. It will be answered by outcomes that no amount of compression can hide.<\/p>\n","protected":false},"excerpt":{"rendered":"
Deans of public health, senators, tort lawyers, and foundation heads do not compete for authority by admitting they want power. They compete by invoking languages of child safety, evidence-based governance, and protecting the vulnerable. These are not descriptions of what … Continue reading