The Work That Refuses to End: Creativity, Authorship, and the Limits of Artificial Intelligence

 

The Work That Refuses to End: Creativity, Authorship, and the Limits of Artificial Intelligence

Carl Jean

Abstract

This essay examines how artificial intelligence transforms creativity by separating creative output from creative formation. It introduces the concept of creative displacement, defined as the transfer of generative processes from human struggle, revision, and embodied judgment into automated systems capable of producing sophisticated artifacts without undergoing the conditions that traditionally form originality. Drawing on Margaret Boden’s theory of transformational creativity, Walter Benjamin’s account of technological reproduction, Roland Barthes and Michel Foucault’s critiques of authorship, and contemporary analyses by Kate Crawford, Matteo Pasquinelli, and Andy Clark, the essay argues that AI does not end creativity but alters the conditions under which creativity can be meaningfully attributed. While creativity has always been distributed across tools and cultural systems, this essay contends that distribution becomes displacement when systems replace rather than extend the formative processes through which creators develop judgment, style, and imaginative authority. It proposes a creative threshold, defined by process displacement, formative absence, and selection dominance. The essay’s central contribution is to distinguish between the extension and displacement of creative processes, and to formalize this distinction through the concept of the creative threshold, which clarifies when technological systems support creativity and when they begin to replace the conditions that make it possible. The essay concludes that artificial intelligence is creatively legitimate only when it preserves creative formation as an epistemic, aesthetic, and ethical necessity.

‎

I. Creativity as Formation, Not Output

Creativity is often mistaken for the production of novel artifacts, but its deeper structure lies in the processes through which those artifacts emerge. To write, compose, design, or build is not merely to produce an outcome; it is to undergo a transformation in perception, judgment, and possibility. This transformation is epistemic, in that it generates new understanding; aesthetic, in that it refines sensitivity to form and meaning; and ethical, in that it cultivates discipline, attention, and responsibility toward one’s work. Margaret Boden’s distinction between exploratory and transformational creativity clarifies this structure: the former generates variation within established conceptual spaces, while the latter restructures those spaces themselves. Artificial intelligence excels at exploratory variation, but whether it produces or merely simulates transformation remains uncertain. Walter Benjamin’s analysis of technological reproduction shows that changes in production alter not only the artwork but the conditions under which it acquires meaning. Matteo Pasquinelli extends this insight to AI, arguing that machine learning systems reorganize cultural production by extracting patterns from collective labor and recombining them at scale. Artificial intelligence brings these pressures together: it produces variation at scale while weakening the link between creative output and formative transformation.

‎

Artificial intelligence becomes creatively dangerous not when it produces poorly, but when it produces well enough to make creative formation appear unnecessary.

‎

II. Creative Displacement and the Transfer of Process

From this transformation emerges creative displacement, the transfer of generative processes from human cognition to automated systems. This displacement does not eliminate human participation; rather, it repositions it. The creator becomes a selector, curator, or evaluator of outputs generated elsewhere. Roland Barthes and Michel Foucault demonstrate that authorship has always been mediated, but mediation is not equivalent to displacement. In mediated creation, tools extend the creator’s capacity while preserving formative engagement; in displacement, systems perform generative work independently of that engagement. Kate Crawford emphasizes that AI systems are built on vast infrastructures of human labor and data, meaning that their outputs reflect accumulated cultural production rather than newly formed understanding. When generative stages are transferred to the system, the creator’s role shifts from producing meaning to selecting among possibilities.

‎

The central question is not whether humans remain involved, but whether their involvement still requires them to become capable.

‎

III. The Counterargument: Distribution vs. Displacement

The strongest objection is that creativity has always been distributed. Language, tools, traditions, and collaborative systems have always shaped creative work. Andy Clark’s extended mind thesis formalizes this by showing that cognition depends on external supports. From this perspective, AI represents a continuation rather than a rupture. Creativity has never been purely internal, and tools have always extended human capability.

‎

The question is not whether cognition is distributed. It is whether distribution still requires the creator to develop.

‎

The distinction between extension and displacement is therefore decisive. Extension enlarges capacity while preserving formation; displacement produces outputs while making formation optional.

‎

IV. The Limits of Expansion: Process and Meaning

The counterargument holds only if process is incidental rather than constitutive. Across domains, creative development depends on sustained engagement with resistance. Writers revise drafts, designers iterate forms, and musicians rehearse structures not because they lack outputs, but because these processes form judgment. When such processes are bypassed, the artifact may still appear, but the formation it signifies may not occur. This claim is supported by research on expertise development, where sustained engagement with difficulty has been shown to be essential to the acquisition of high-level skill, suggesting that the removal of formative struggle alters not only performance but the underlying capacities that make it possible (Ericsson).

‎

When variation replaces transformation, novelty persists but the capacity to recognize meaningful difference diminishes.

‎

V. Formation Under Pressure: A Case of Creative Development

Historical evidence shows that creative work emerges through sustained engagement with resistance, but to move beyond illustration, it is necessary to examine how this process operates in detail. Consider the compositional notebooks of Ludwig van Beethoven, which document not simply the drafting of musical ideas but their transformation through repeated revision. Sketches of the Eroica Symphony reveal themes that are fragmented, recombined, abandoned, and reconstructed across dozens of iterations before reaching their final form. What is visible in these notebooks is not the execution of a pre-existing idea, but the gradual emergence of musical judgment through constraint, failure, and reconfiguration.

‎

This process matters because it demonstrates that the artifact—the finished symphony—cannot be separated from the formation of the composer who produced it. The revisions are not incidental; they are the means through which Beethoven develops the capacity to hear, evaluate, and transform musical structure. Each failed variation is not wasted effort but a condition of possibility for the final work. The composer does not simply produce the music; the music produces the composer’s capacity to create it.

‎

Contemporary creative practice exhibits the same structure. Writers develop voice through revision, designers refine judgment through iteration, and musicians cultivate perception through rehearsal. Tools such as word processors, digital audio workstations, and design software extend these processes, but they do not eliminate the need to undergo them. They amplify formation rather than replace it.

‎

Artificial intelligence differs because it can generate outputs that appear complete without requiring the creator to pass through comparable stages of development. The artifact may exist, but the formative trajectory that would make its emergence meaningful may not. This is the critical distinction: AI does not merely accelerate production; it risks severing the relationship between production and formation.

‎

When the process that produces the work no longer produces the worker, creativity persists but formation declines.

‎

VI. Derivative Saturation and the Restructuring of Value

As AI-generated content proliferates, derivative saturation emerges: a condition in which variation increases while meaningful transformation becomes harder to identify. Matteo Pasquinelli’s account of statistical learning explains this phenomenon: AI systems recombine existing patterns rather than generate new conceptual spaces. Novelty persists, but its significance changes.

‎

When novelty becomes abundant, formation becomes the only scarce resource.

‎

VII. Authorship, Responsibility, and the Problem of Unformed Creation

Creative displacement not only destabilizes authorship but raises a deeper problem of responsibility. If a creator produces an artifact without undergoing the formative processes traditionally associated with its creation, what does it mean to be responsible for that work? Responsibility in creative practice has historically implied more than attribution; it has implied a relation between the maker and the process through which the work emerges. To be responsible for a work is not simply to claim it, but to have been formed by the conditions that produced it.

‎

This distinction becomes clearer when contrasted with adjacent practices. In collaborative creation, multiple contributors share responsibility because each participates in the formative process. In ghostwriting, authorship may be transferred, but formation still occurs—only in a different agent. In forms of appropriation or found art, the act of selection itself can be formative when it recontextualizes material in a way that requires judgment and transformation. In each case, responsibility is preserved because formation remains present, even if distributed.

‎

Creative displacement introduces a different condition. When outputs are generated independently of the creator’s formative engagement, responsibility risks becoming detached from process. The creator may still select, refine, or present the work, but these actions may not require the development of the capacities that the work appears to express. Responsibility becomes a matter of attribution rather than transformation.

‎

Responsibility, in this sense, depends not only on authorship but on the presence of formation, since it is through formation that creators acquire the capacities for judgment, evaluation, and accountability.

‎

When formation disappears, authorship becomes attribution without responsibility.

‎

This shift has both ethical and aesthetic consequences. Ethically, it raises the question of whether individuals can be accountable for outputs they did not meaningfully produce. Aesthetically, it challenges the basis on which works are evaluated, since the connection between artifact and capacity becomes uncertain.

‎

VIII. The Creative Threshold

The creative threshold marks the point at which AI transitions from extending creativity to displacing it. It is defined by three conditions: process displacement, formative absence, and selection dominance. These conditions operate as a continuum rather than a binary. The threshold is crossed when all three converge to make formative engagement unnecessary.

‎

Below the threshold, tools extend creativity. Beyond it, they replace the processes that make creativity meaningful.

‎

Conclusion

Artificial intelligence does not eliminate creativity. It transforms the conditions under which creativity occurs. The central question is not whether AI can produce creative outputs, but whether human formation remains necessary within systems that no longer require it.

‎

What is at stake is not novelty, but the necessity of becoming capable of creating.

‎

Final Doctrine

Artificial intelligence is creatively legitimate only when it preserves creative formation as an epistemic, aesthetic, and ethical necessity—requiring creators to develop the capacities that their work expresses.

Related Reading

If this essay examines how artificial intelligence reshapes creativity by separating output from formation, the next essay extends this inquiry into the ethical domain. The Diffusion of Responsibility: Moral Latency and the Ethics of Artificial Intelligence explores how AI systems redistribute responsibility across networks in ways that can weaken accountability. Together, these essays trace a shared problem: as artificial systems take on increasingly complex functions, both creative formation and moral responsibility risk becoming displaced rather than developed.

Join the Conversation

If you are working at the intersection of creativity, technology, and education, I invite you to stay connected with this ongoing inquiry. Subscribe to The Carl Jean Journal for future essays, and consider sharing your perspective in the comments. How do you see artificial intelligence reshaping creative formation and responsibility in your own field?

Works Cited

Barthes, Roland. “The Death of the Author.” Image–Music–Text, translated by Stephen Heath, Hill and Wang, 1977, pp. 142–148.

Benjamin, Walter. “The Work of Art in the Age of Mechanical Reproduction.” Illuminations, translated by Harry Zohn, edited by Hannah Arendt, Schocken Books, 1968, pp. 217–251.

Boden, Margaret A. The Creative Mind: Myths and Mechanisms. 2nd ed., Routledge, 2004.

Clark, Andy. Supersizing the Mind: Embodiment, Action, and Cognitive Extension. Oxford UP, 2008.

Crawford, Kate. Atlas of AI: Power, Politics, and the Planetary Costs of Artificial Intelligence. Yale UP, 2021.

Dornis, Tim W. “Artificial Creativity and the Law.” Yale Journal of Law & Technology, vol. 22, 2020, pp. 1–61.

Ericsson, K. Anders, et al. “The Role of Deliberate Practice in the Acquisition of Expert Performance.” Psychological Review, vol. 100, no. 3, 1993, pp. 363–406.

Foucault, Michel. “What Is an Author?” Language, Counter-Memory, Practice: Selected Essays and Interviews, edited by Donald F. Bouchard, translated by Donald F. Bouchard and Sherry Simon, Cornell UP, 1977, pp. 113–138.

Pasquinelli, Matteo. The Eye of the Master: A Social History of Artificial Intelligence. Verso, 2023.