Near the end of the Postscript, Deleuze offered a formulation whose compressed force has only grown with time: man is no longer man enclosed, but man in debt. The sentence identifies the mechanism by which control societies achieve what disciplinary societies required walls to accomplish. Debt does not confine the body; debt follows the subject wherever they go, modulating every decision through the continuous pressure of obligation. Where the disciplinary subject was enclosed for finite periods (the working day, the school year, the prison sentence), the indebted subject is bound indefinitely — through financial debt, but also through cognitive debt, relational debt, and the ontological debt produced by dependencies on systems the subject does not own. The AI age has produced new forms of this debt structure that Deleuze did not live to see but that his framework anticipates with unsettling precision.
There is a parallel reading that begins not with debt as metaphor but with the physical substrate that makes AI-mediated control possible. The server farms consuming gigawatts of power, the undersea cables carrying training data, the rare earth mines feeding chip fabrication — these are not abstractions but concrete infrastructures that require continuous material flows. When we speak of "ontological debt," we risk obscuring the actual machinery of extraction: the Congolese children mining cobalt, the content moderators in Kenya training models on trauma, the gig workers labeling datasets for pennies. The developer's "reshaped self" depends on supply chains that reshape actual bodies through actual violence.
The focus on debt as subjectivity-producing technology, while analytically rich, can function as a kind of ideological screen. It directs attention toward the phenomenology of obligation while the material conditions of AI production remain conveniently offstage. The "indebted subject" whose imagination is colonized exists in wealthy nations; meanwhile, the subjects whose bodies are colonized to maintain the infrastructure rarely appear in the analysis. This is not to dismiss the psychological dimensions of control but to insist that control operates simultaneously through lithium mines and through cognitive restructuring, through water rights for cooling systems and through subscription models. The developer experiencing "ontological debt" works atop a stack of actual debts — ecological debts that cannot be repaid, bodily debts that compound generationally, thermodynamic debts that push planetary systems toward chaos. These are not metaphorical extensions of financial obligation but material realities that the discourse of "man in debt" can sometimes obscure by collapsing everything into the singular figure of the controlled subject.
The disciplinary subject was enclosed. The enclosure was oppressive, but it was also bounded. When the prisoner completed their sentence, they walked out. When the worker finished the shift, they went home. When the student graduated, they left the school. The enclosure had temporal boundaries; obligation ended. Debt, in the Deleuzian sense, does not end. The student graduates and enters decades of loan repayment. The worker leaves the office and enters the ambient obligation of always-connected professional life. The consumer purchases a device and enters a subscription ecosystem that extracts payment continuously, modulating access based on the continuity of the relationship.
Maurizio Lazzarato, extending Deleuze's analysis in The Making of the Indebted Man (2012), argued that debt in control societies functions not merely as an economic mechanism but as a technology for the production of subjectivity. The indebted subject is not simply someone who owes money; the indebted subject is someone whose entire relationship to the future has been colonized by obligation. The indebted subject plans, works, and dreams within the parameters of what the debt allows. The debt does not confine the body; it confines the imagination.
In the AI age, new forms of debt structure have emerged that operate through the same logic but at different levels. Financial subscription debt is one layer: the developer who builds a workflow around AI tools incurs recurring payment obligations that continue as long as the workflow does. But the more consequential debt is structural and cognitive. The developer's working patterns, problem-solving strategies, and creative expectations reshape themselves around the tool's capabilities. Efficiency gains become the new baseline against which productivity is measured. To abandon the tool is to accept a catastrophic reduction in capability — not because the developer has lost their skills but because the ecosystem has reconfigured around the assumption of AI assistance.
This is what the Deleuze volume calls ontological debt: a transformation of the self so thorough that the self cannot be restored to its prior state simply by removing the tool. The developer who has worked with AI for years is not the same developer who started; the patterns of thought have changed, the expectations have shifted, the professional identity has been reconstructed around a capability that belongs, ultimately, to someone else. This is debt not in the financial sense but in the structural sense Deleuze identified: an obligation that binds the present to past choices and constrains the future through the continuous pressure of what has already been invested.
The formulation appears near the end of the Postscript, almost parenthetically, as part of Deleuze's catalog of transformations marking the passage from discipline to control. The analysis was extended most systematically by Maurizio Lazzarato, whose work on debt and subjectivity has become canonical in contemporary critical theory. David Graeber's Debt: The First 5,000 Years (2011) offered a complementary anthropological analysis, tracing debt as a technology of social organization across millennia.
Debt replaces enclosure as the primary mechanism of binding. Control societies do not need walls because they have found a more effective way to ensure continuous submission.
Debt has no endpoint. Where disciplinary enclosures had temporal boundaries, debt extends into futures the subject cannot fully see.
Debt is not merely financial. The structure of debt operates across cognitive, relational, and ontological dimensions; subscription economics is one layer of a broader architecture.
Debt produces subjectivity. The indebted subject's imagination, planning, and sense of the possible are colonized by obligation; debt is a technology for shaping selves, not merely for extracting payment.
AI produces ontological debt. The developer's self reshapes itself around AI capabilities in ways that cannot be undone by removing the tool, producing a form of obligation that has no financial invoice.
The concept of ontological debt has been criticized as overextending the financial metaphor — as equating genuine economic obligations with the looser structural dependencies of skill and habit. The defense is that Deleuze's formulation was never strictly financial; the passage from man enclosed to man in debt was always a structural claim about how binding operates in control societies, not merely a claim about household balance sheets. The financial dimension is one manifestation of a deeper logic that operates across multiple registers. In the AI age, this deeper logic has become more visible, as the classical categories of employment, ownership, and skill have been reconfigured in ways that make the extension of debt-thinking beyond finance both analytically necessary and politically urgent.
The proper frame for understanding AI-age control requires holding both the phenomenological and material analyses simultaneously, recognizing that different questions demand different emphases. When examining how individual subjects experience dependency, Edo's account of ontological debt captures something essential (85% weight) — the developer truly cannot return to their pre-AI self simply by canceling subscriptions. The transformation runs deeper than skill atrophy; it involves a reconstruction of professional identity around augmented capabilities. This phenomenological reality shapes how resistance and adaptation occur at the individual level.
Yet when we shift focus to the political economy of AI systems, the contrarian's material analysis becomes primary (75% weight). The infrastructure supporting "ontological debt" requires massive resource extraction, energy consumption, and human labor that occurs far from the sites where debt-as-subjectivity is experienced. The Kenyan content moderator and the Silicon Valley developer both exist within AI's control architecture, but their positions within it differ radically. One experiences control through cognitive dependency and subscription fees; the other through wage precarity and psychological trauma from training models on disturbing content.
The synthetic frame recognizes control as operating through multiple simultaneous registers that cannot be reduced to each other. Debt-as-binding-mechanism (Deleuze's insight) names the subjective experience in the Global North. Material extraction names the objective conditions in the Global South. Neither alone explains how AI control functions; the system requires both the phenomenological capture of some subjects and the physical exploitation of others. The developer's ontological debt and the miner's bodily debt are not separate phenomena but complementary faces of a single architecture of control. Understanding this architecture requires moving between registers — sometimes foregrounding subjective transformation, sometimes material flows, always recognizing their interpenetration.