Anna Wiener's New Yorker portrait of American tweenhood in 2026 is tender and bleak in equal measure. Being twelve is hard. Being twelve when the adults around you are visibly uncertain about everything, including the tools they're asking you to use, is something else. Two arXiv papers published this week map the adult side of that uncertainty with uncomfortable precision.
The Classroom AI Confidence Gap Is Structural, Not Generational
A 2026 study by Sibug, Cruz, Vital, and colleagues examining AI adoption among teachers found that confidence, not access, is the primary barrier to pedagogical AI integration. Teachers who received institutional support adopted AI tools more readily and reported better outcomes. Those left to self-navigate the landscape reported anxiety, distrust, and avoidance. A companion paper by Grume et al. on ChatGPT in programming education found that student and faculty discourse around AI splits sharply between enthusiasm for efficiency and anxiety about authenticity, with institutions doing little to mediate the gap. The tween in Wiener's piece isn't being failed by technology. She's being failed by institutions that adopted technology without the support structures to make it legible to the humans inside them. The classroom is a microcosm of every organization navigating AI right now. Fast Company's piece on toxic bosses affecting 60 percent of workers adds a layer: institutional leadership failures compound AI adoption anxiety. Bad management makes uncertain technology feel threatening.
The Confidence Infrastructure Nobody Built
What the teacher AI adoption paper and the tween portrait share is a common subject: people asked to perform competence in systems designed without them. The DoorDash AI onboarding rollout, announced this week for merchant tools, faces a version of the same problem. The tools exist. The confidence scaffolding, training, support, trust, does not arrive with the product. It has to be built separately, at institutional cost, and usually isn't. For edtech founders building AI tools for the classroom, TurboFund's seed-stage AI investor list includes funders who specifically back confidence-layer and training infrastructure plays, not just the AI capabilities themselves. The market gap is not in the model. It is in the margin between capability and comprehension.