The cycle that began with [YOU] on AI asks a question that sounds like technology policy but is really a question about trust: do you believe the people around you—including the children—are capable of navigating powerful tools without being destroyed by them? Skenazy has been answering that question about children for two decades, and her answer is the same whether the tool is a subway car or a large language model. The capability is there. The question is whether the adults will permit its exercise.
The cycle describes the “silent middle”—the majority who feel both the exhilaration and the terror of AI’s arrival but lack a clean narrative. Skenazy knows this silence. She has spent her career speaking to its parenting equivalent: parents who sense their children need more freedom than the culture allows but fear that granting it makes them negligent. Her contribution to the cycle is the diagnostic: the discomfort parents feel about children and AI is not evidence that the tool is dangerous. It is evidence that worst-first thinking is operating. The cure is not prohibition. It is the deliberate, supported, calibrated extension of trust that she calls scaffolded autonomy.
Her analysis lands hardest on the question of competence formation. Competence cannot precede the opportunity to practice; it is produced by it. A child kept away from AI until some imagined future readiness arrives will reach adulthood without the critical faculty that only encounter with the technology can build. The identical logic that produced the supervised generation—the most protected cohort in American history, and the most anxious one at college entry—is now being applied to the most consequential tools of that cohort’s adult lives. The pattern is recognizable. The invoice will be familiar.
Skenazy was a New York City journalist when the subway column ran. The reaction it generated was disproportionate enough to be diagnostic: what she had treated as a straightforward parenting decision—a nine-year-old with a MetroCard, a map, and twenty dollars—was read by a significant portion of the public as an act of reckless endangerment. Crime in New York had been falling for fifteen years. The probability of a stranger kidnapping in the United States was roughly one in 1.4 million. None of this mattered against the felt certainty that the world was dangerous and children required constant protection.
The overreaction became her subject. She documented the school that banned running at recess, the park where a mother faced child protective services for allowing a ten-year-old to walk there unaccompanied, the summer camp that required helmets for tetherball. Each case was dismissible individually; collectively they mapped a pathology—a culture that had decided childhood was a condition of extreme vulnerability requiring continuous adult intervention, not based on data but on a feeling so pervasive it had become invisible. The Free-Range Kids movement she founded was an attempt to name the pathology and supply its antidote.
Her collaboration with Haidt and Lukianoff on The Coddling of the American Mind gave the pathology its conceptual vocabulary: safetyism, the approach that prioritizes feelings of safety at the cost of intellectual rigor, open debate, and the developmental friction that growth requires. By 2025, the concept had migrated directly into AI policy discourse, where it proved as diagnostically useful as it had been in the campus setting.
Worst-First Thinking. The reflex at the center of Skenazy’s analysis: encountering something new, imagining the worst thing it could do, and treating that worst thing as the most likely thing. Applied to AI and children, worst-first thinking bypasses the actual child and responds to an imagined catastrophe. The cure is not dismissing the fear but examining the evidence—asking whether the worst case is actually the base case, and what is lost through the prevention it demands. Worst-first thinking forecloses questions rather than answering them.
The Overprotection Paradox. The central mechanism: a genuine risk is identified, a protection is implemented, and the protection prevents development of the very capacity that would have allowed the person to manage the risk independently. The result is not safety but fragility. The generation raised under the most protective conditions in American history arrived at college as the most anxious. Safetyism applied to AI produces the same structure: children kept away from the technology never develop the critical faculty that only direct encounter can build.
Scaffolded Autonomy. Skenazy’s operational alternative to both prohibition and permissiveness. Structure without control: the child has access to the tool, the adult is available for conversation, and the trust is extended slightly ahead of the demonstrated capability because the capability can only be demonstrated through the experience the trust makes possible. The kitchen-table conversation—genuine inquiry into what the child is finding, not surveillance disguised as interest—is the practice through which scaffolded autonomy operates at home. Scaffolded autonomy is trust, then verify.
The AI Playground. Old playgrounds built physical calibration through real risk and immediate feedback. The AI playground builds intellectual calibration through a different architecture: challenges where the failure mode is invisible, where fluent and wrong feel identical to fluent and right, and where the adult’s role is not to pad the surface but to make the invisible feedback visible. The question “Does anything seem off?” asked consistently over dozens of interactions builds the faculty that no prohibition can install. The default mode network—the neural substrate of creative synthesis—requires unstructured time; boredom is developmental pressure, and eliminating it eliminates the teaching.
Competence Is Built, Not Given. Albert Bandura’s self-efficacy research established that competence beliefs are produced primarily by mastery experience—direct personal evidence that you have faced a challenge and met it. This means readiness does not precede the opportunity to practice; it is produced by it. The child who arrives at adulthood never having encountered AI’s intellectual challenges arrives without the self-concept of someone who can handle them. Waiting for readiness before granting access is the autonomy paradox in its purest form.