The cycle that began with [YOU] on AI confronts the reader repeatedly with a pattern that Vallor's framework names with precision: the AI tool eliminates a form of difficulty, and along with the difficulty goes the specific virtue that the difficulty had cultivated. Segal's account of the engineer whose twenty years of debugging practice deposited not merely technical skill but patience, intellectual humility, and the courage to be wrong—and who now delegates that practice to AI—is a case study in moral deskilling. Not cognitive deskilling alone. Moral deskilling: the erosion of the character traits that the practice had formed.
Segal's own moment of recognizing this is one of the cycle's most honest passages. He describes deleting a passage Claude had produced—eloquent, well-structured, hitting every note—and spending two hours at a coffee shop with a notebook, writing by hand until he found the version that was his. Rougher. More qualified. More honest about what he did not yet know. Vallor would read that act of deletion as a moral achievement, not merely a stylistic preference. It was the exercise of intellectual courage against the structural pull of a tool designed to make courage unnecessary. The product was inferior by every surface metric. The character it preserved was more durable than anything the machine could produce.
The ascending friction thesis that Segal develops—that AI does not eliminate difficulty but relocates it to a higher cognitive floor—is precisely the claim Vallor would subject to scrutiny rather than acceptance. She does not dispute the possibility. She insists that the relocation is not automatic. The new friction provides an opportunity for new virtues to develop; whether that opportunity is seized depends entirely on whether the practitioner, and the institutions surrounding her, recognize that an opportunity exists and act to realize it. Otherwise the ascending friction resolves into a new cycle of delegation, and the practitioner arrives at the higher floor without the judgment the floor demands.
Vallor is also the voice in the cycle that most clearly names the circularity at the heart of the AI age's moral challenge: the virtues required to use the tools wisely are the virtues the tools are most efficient at eroding. Breaking this circle requires an intervention from outside the tool itself—a practice, a community, a commitment that holds the practitioner accountable to a standard of character that the technology's metrics cannot measure and that its design does not reward.
Vallor studied philosophy at the University of California, Davis, and completed her doctorate at UC Santa Cruz, where her dissertation engaged the intersection of feminist ethics, epistemology, and the philosophy of technology. Her early work focused on care ethics and the question of how relationships of care—between parents and children, caregivers and those cared for—are shaped and potentially damaged by technological mediation. This question would prove prophetic: by the time AI companion apps and therapy bots had arrived at scale, Vallor had already built the philosophical vocabulary to analyze what they were doing to human relationships.
The publication of Technology and the Virtues: A Philosophical Guide to a Future Worth Wanting in 2016 established her as the most systematic thinker working at the intersection of virtue ethics and technology ethics. The book drew on Aristotle, Confucius, and the Buddhist sila tradition to argue that three independent philosophical lineages, developed across thousands of miles and hundreds of years, had converged on the same structural insight: virtue cannot be installed. It can only be cultivated through deliberate practice in conditions that demand it. Remove the conditions, and the cultivation stops. The AI age was removing conditions at an unprecedented rate.
Her two years inside Google as an AI ethicist (2018–2019) were formative in a specific way: she saw firsthand how the institutional machinery of a technology company—the quarterly metrics, the user engagement dashboards, the competitive pressure to ship features before the competition—creates an environment in which the question “What kind of person does this product produce?” is not merely unasked but structurally unanswerable. The dashboard does not measure character. It measures usage, retention, time-on-platform. A tool that erodes critical thinking while increasing engagement scores well on every metric the company tracks.
The AI Mirror, published in 2024, sharpened the earlier analysis into a single image: AI is not an intelligence but a mirror, reflecting patterns from training data with a fluency that invites trust. The mirror produces images that look like thought—that have the structure, the grammar, the confident tone of thought. But they are reflections, not thoughts. They originate not in understanding but in pattern-matching, which is a backward-facing operation: projecting from what has been toward what might plausibly follow, without any mechanism for distinguishing the plausible from the true. The danger is not that the mirror lies. It is that the mirror is indistinguishable from the thing it reflects.
Technologies of the self. Vallor extended Foucault's concept to cover all habitual tool use: every technology we use repeatedly is a technology of the self, shaping our dispositions below the threshold of awareness. The carpenter who spends thirty years with hand planes is not only producing furniture—the resistance of the wood, the demand for precision, the patience required to bring a joint to tolerance are producing the carpenter. The social media feed trains scattered attention and reactive opinion-formation with equal reliability, not because any designer intended these outcomes but because the structure of the practice does what the structure of the practice does, regardless of intentions.
Moral deskilling. Moral deskilling is Vallor's extension of Harry Braverman's analysis of industrial deskilling to cognitive and moral domains. When AI systems take over the activities through which intellectual and moral character has traditionally developed—first-draft writing, research, debugging, diagnosis, argument structure—the capacities being eroded are not merely useful professional skills. They are the capacities through which a person determines what is true, what is right, what is worth doing. The person who has lost these capacities has not merely become less effective. She has become less capable of the moral discernment on which the quality of her life depends.
The invisible curriculum. AI tools teach through their structure, not their content. The invisible curriculum of AI operates through three mechanisms: confidence calibration (training users to treat fluency as a proxy for accuracy), structural preemption (delivering a structure before the user has done the cognitive work of constructing one), and the elimination of productive failure (producing outputs that work, denying the user the diagnostic experience of failure). Each mechanism is invisible because each operates below the threshold of any individual interaction—no single acceptance is significant; the ten-thousandth has produced a settled disposition.
Technomoral virtue. Technomoral virtues are the character traits human beings specifically need in order to flourish in a technological society: honesty, justice, courage, empathy, self-control, humility, flexibility, and others. Each faces a specific threat from AI tools designed to minimize friction. Courage is threatened by a tool that makes the comfortable choice the easiest. Humility is threatened by a tool that never pushes back. Self-control is threatened by a tool that never tires and provides no natural stopping point. The virtues required to use the tools wisely are, structurally, the virtues the tools are most efficient at eroding.
The virtues of friction. Vallor's central philosophical claim, grounded in Aristotle, Confucius, and Buddhist sila, is that virtue is cultivated through resistance. The carpenter's patience develops through wood that catches the grain wrong. The engineer's intellectual humility develops through code that will not compile despite her certainty that it should. The writer's courage develops through the blank page she cannot avoid by accepting the machine's draft. Remove the friction, and the cultivation stops. The question the AI age poses is whether ascending friction—the relocation of difficulty to higher cognitive levels—provides new occasions for virtue development, or whether the practitioner will delegate her way to the top without earning any of the views.