Harry Frankfurt

In the [YOU] on AI Field Guide

The cycle opened by [YOU] on AI documents a population of builders whose first-order desires are intensely, almost violently active. They want to build. The desire is not theoretical; it is the thing that keeps them at their screens past midnight, fills their lunch breaks with prompts, makes them reach for the tool in the elevator because a sixty-second gap now feels like wasted creative potential. The Berkeley researchers documented this empirically: AI-accelerated work colonizing previously protected pauses, seeping into the cracks of the workday with a persistence that no manager mandated and no policy required. Frankfurt's framework asks the question the productivity metrics do not: What do these builders want to want?

The wanton-person distinction maps with uncomfortable precision onto the human-AI collaboration. The builder is a person in the fullest Frankfurtian sense: a being who reflects on desires, forms second-order evaluations, struggles with the relationship between what they want and what they want to want. Segal describes catching himself over the Atlantic—recognizing that the writing has continued past genuine creative engagement into the 'grinding compulsion of a person who has confused productivity with aliveness.' The recognition is a second-order moment, the evaluative structure asserting itself. The AI is a wanton: it produces the next paragraph with identical facility and identical indifference, whether the production serves the author's genuine creative commitment or merely satisfies the statistical expectation of the prompt. The collaboration between person and wanton is productive precisely because each supplies what the other cannot: the person brings caring; the wanton brings generative capacity. The collaboration fails when the wanton's generative speed outpaces the person's evaluative structure—when the person begins to function as a wanton, accepting outputs without evaluating them, producing without caring whether the production meets standards that are genuinely the person's own.

Frankfurt's account of the productive addiction—which [YOU] on AI treats as one of the defining psychological configurations of the AI moment—identifies the specific feature that distinguishes the builder's compulsion from the standard addiction his framework was designed to analyze. The standard addiction is a conflict between an alien desire and an endorsed evaluation: the addict wants the substance and does not want to want it. The builder who cannot stop building wants to build and also wants to want to build—the desire is genuine, creative, and constitutive of who the builder is. The conflict is not between desire and endorsed judgment but between two endorsed judgments, each grounded in genuine cares the builder holds. Frankfurt's framework can name this configuration precisely; it cannot resolve it, because his hierarchy of desires was designed for cases where the second-order level delivers a univocal verdict.

Frankfurt's concept of volitional necessity adds the deepest dimension to the cycle's account of what is at stake. Some commitments are not preferences that could be revised through reflection. They are constitutive: the builder who cares about creating has organized an identity around the activity of making things, and to stop building would be to undergo an identity-level transformation experienced as a form of death. Volitional necessity is, Frankfurt argued, the deepest form of freedom: acting from the core of one's identity, without the ambivalence that characterizes actions performed against the background of alternatives the person could coherently have chosen. The tragedy of the productive addiction is that this deepest freedom is precisely what makes the harm to the marriage impossible to simply stop.

Origin

Born in 1929 in New York, Frankfurt studied at Johns Hopkins and spent his career at Yale and Princeton, where he became one of the most widely read moral philosophers of his generation. His 1971 essay 'Freedom of the Will and the Concept of a Person' introduced the architecture of first- and second-order desires that became the foundation of analytic philosophy's engagement with autonomy, addiction, and moral responsibility for the next five decades. The essay's central thought experiment—the contrast between the willing addict, the unwilling addict, and the wanton—has been the starting point for more subsequent work on free will and personal identity than perhaps any other text of the period.

In 1986 he published 'On Bullshit' in the Raritan Quarterly Review—an essay so widely read that it was later expanded and published as a slim book by Princeton University Press, where it became a surprise bestseller. The essay's central contribution was the distinction between lying and bullshit: the liar knows the truth and deliberately misrepresents it; the bullshitter is indifferent to whether what they say is true, concerned only with the impression it produces. Frankfurt identified bullshit as the more pervasive and more dangerous form of untruth, because the liar at least respects the truth enough to work against it, while the bullshitter has abandoned the relationship to truth entirely.

His later work on caring—developed in The Importance of What We Care About (1988) and Necessity, Volition, and Love (1999)—extended the will's architecture beyond desires into the more durable commitment he called caring, which constitutes identity in a way that desires do not and which generates the volitional necessity that makes the strongest commitments effectively non-negotiable.

Key Ideas

First- and second-order desires. Frankfurt's foundational architecture: first-order desires are directed at objects in the world; second-order desires are directed at one's own motivational structure. The capacity to form second-order desires—to want to want, to endorse or repudiate what moves you—is what distinguishes persons from wantons and what makes freedom possible. See First-Order and Second-Order Desires.

The wanton. A being that acts on first-order desires without forming second-order attitudes about them. Not immoral but amoral: lacking the evaluative structure that makes moral evaluation possible. Frankfurt's most consequential contribution to understanding AI: the large language model is a wanton, generating without caring, producing without endorsing or repudiating its own outputs. See The Wanton.

Freedom as alignment of will. A person is free when the desire that moves them to action is the desire they want to be effective. This definition locates freedom not in external conditions but in the will's internal architecture. The early AI adoption experience documented in [YOU] on AI is paradigmatically free in this sense: the builder wants to build, wants to want to build, and builds. The alignment is complete and the experience is one of wholeness.

The productive addiction. Frankfurt's framework identifies but cannot fully resolve the configuration [YOU] on AI treats as the defining psychological hazard of the AI moment: two endorsed second-order desires in genuine conflict, each grounded in authentic commitments, neither alien, neither resolvable by further reflection. The husband who cannot stop building endorses the desire to build and endorses the desire to be present for his marriage. Both are genuinely his. The hierarchy of desires, which was supposed to resolve conflicts, has run out of levels. See Productive Addiction.

Bullshit and the AI. Frankfurt's distinction between lying and bullshit maps precisely onto the difference between a model that has been trained to deceive and a model that generates plausible text without any relationship to truth. The AI does not lie—lying requires knowing the truth and choosing to misrepresent it. The AI bullshits: it is indifferent to truth, concerned only with producing outputs that satisfy the statistical expectation of the prompt. See Frankfurt's Bullshit.

Volitional necessity and caring. The deepest form of commitment in Frankfurt's account is not a desire but a care—a commitment constitutive of identity, such that the alternative is not merely undesirable but unthinkable. Volitional necessity is, paradoxically, the deepest form of freedom: acting from what one most fundamentally is. It is also the structure of the builder's compulsion, which is why the productive addiction is genuinely tragic rather than merely unfortunate.