Cognitive Redistribution: Rethinking Agency and Creativity in the Age of AI

Introduction

Across every knowledge-intensive field, a seismic but under-theorized shift is underway: cognitive labour is no longer concentrated inside the human mind but redistributed across human–machine assemblages. This redistribution is not equivalent to “automation,” nor is it another chapter in the familiar story of technical externalization. Rather, it represents a qualitative transformation in how ideation, interpretation, and decision-making are generated and circulated in contemporary practice.

I propose the term Cognitive Redistribution to describe this shift. Where prior technologies extended or preserved human memory and skill, contemporary AI models actively generate, combine, evaluate, and prioritize cultural material. In Bernard Stiegler’s framework, these systems constitute a novel form of tertiary retention—but one whose agency, variability, and dialogical responsiveness exceed the traditional function of exteriorized memory (Stiegler 1998). AI does not merely store or repeat thought; it participates in its production. As a result, ideation becomes partially technical, agency becomes distributed in unprecedented ways, and the grammar of thinking itself—what Yuk Hui calls the “cosmotechnical condition”—is transformed (Hui 2016).

This article develops the conceptual scaffolding for Cognitive Redistribution, drawing on digital humanities, posthumanism, cognitive science, and media theory. It argues that AI creates a new epistemic environment in which:

1. Externalization (what we shouldn’t be doing) is generative rather than archival,
2. Agency is redistributed across human and non-human actants in qualitatively new ways,
3. Ideation becomes a technical recombination of cultural units,
4. Human cognitive labour shifts toward curation, discrimination, and contextual judgment, and
5. Professional fields (such as design) experience disorientation because their self-understanding is still anchored in pre-AI models of authorship.

---

# 1. From Externalization to Generation: Stiegler and the New Retention

Stiegler’s philosophy of technics provides the most generative foundation for this shift. For Stiegler, human cognition has always been supplemented by exteriorization: tools, writing, photography, and digital media store and transmit what no individual mind can retain (Stiegler 1998, 2009). These “tertiary retentions” form a technical milieu in which human attention and memory evolve.

A parallel we may make is: what did portrait artists that pursuit realism thought about the invention of photography? We can speculate they had to admit the realism was particularly sharp (and done faster, more precisely). Thus the aspect of realism wasn’t anymore a concern of the artist. Was it ever, we may ask? Was there a point, in the first place, in using human cognition to replicate something realistically? And when photography came along, something else came alive: photography carried on the art of the portrait, and painters carried on with more expressive distortions of their subjects. In a way, one same artist — the portrait artist — took bifucating paths, and speciated into two types of artists. There was, thus, a cognitive redistribution of the activity of portraits.

Yet AI introduces something categorically new:

• Earlier forms of tertiary retention recorded thought; AI produces it.
• Externalization becomes generative rather than archival.
• The system interacts dialogically with the user, producing outputs that redirect human intention.

Thus, the locus of ideation itself is partially displaced. Where earlier externalizations stabilized knowledge, AI destabilizes and proliferates it. It creates new combinatorial configurations, new proposals, new analogies, new interpretations—at speeds and scales inaccessible to the individual.

1.1 Pharmacology: Skill Atrophy and Cognitive Expansion

Stiegler’s pharmakon—every technology as both remedy and poison—is particularly relevant (Stiegler 2010). Cognitive Redistribution has two pharmacological effects:

• Cure: frees humans from routine synthesis, benchmarking, pattern detection, and lateral combinatorics, making room for higher-order judgment.
• Poison: induces skill atrophy when humans no longer practice the early phases of ideation or research and therefore lose the sensibility that arises from doing so manually.

The challenge is not to resist AI but to reconfigure human roles such that the poisonous effects are compensated by new forms of attention, pedagogy, and interpretive care.

---

# 2. Posthuman Cognition: Hayles and the Cognitive Assemblage

N. Katherine Hayles offers the clearest theoretical support for the idea that cognition is not confined to human minds. In her account, cognition arises in assemblages composed of humans, technical systems, representational media, and environmental constraints (Hayles 2017). AI therefore becomes not a tool but a cognitive partner—an agent participating in the circulation of meaning.

This reframes Cognitive Redistribution:

the question is not “What is the human losing?” but “How does the assemblage reorganize itself under new conditions of distributed competence?”

Hayles’ perspective also clarifies why resistance emerges: professions built on the myth of solitary or exceptional creativity struggle when cognition becomes visibly shared across human and non-human processes.

---

# 3. Beyond Collaboration: Clark, Simondon, and Dialogical Technics

Andy Clark’s extended mind thesis argues that tools become literal functional components of cognition (Clark and Chalmers 1998). However, the AI era goes beyond “extension” or “collaboration,” terms now overused in managerial discourse. What emerges is a dialogical cognitive system:

• The human provides direction, constraints, correction, and judgment.
• The machine provides variation, memory, generativity, and rapid recombination.
• Meaning emerges in the interaction, not in either component alone.

Gilbert Simondon helps articulate this further. For Simondon, humans and technical objects co-individuate: each shapes and is shaped by the other (Simondon 1958). AI introduces a new dynamic of course-correction:

• the system generates profusions of outputs,
• the human filters and adjusts,
• the system re-generates along new trajectories informed by those adjustments.

This iterative loop constitutes a new cognitive rhythm—one defined by feedback, divergence, recombination, and selective convergence.

---

# 4. Latour: Agency, Actants, and the Threshold of the Non-Human

Bruno Latour’s actor–network theory reframes agency as something distributed across human and non-human actants (Latour 2005). Tools act; infrastructures act; environments act. Yet AI complicates Latour’s scheme in a way earlier tools did not.

A hammer shapes behaviour by limiting possibility.

A large language model shapes behaviour by proposing possibility.

This marks a shift from constraint to suggestion, from action-limiting to action-expanding. It also raises delicate questions:

• When does the model’s agency become bias reinforcement rather than creative advancement?
• How do we distinguish meaningful guidance from statistical hallucination?
• What responsibilities arise when non-human actants produce the ideational raw material?

Cognitive Redistribution thus occurs on a field where non-human actors do not merely support human decisions but actively shape the option space.

---

# 5. The Changing Nature of Externalization: Bush, Hui, and the Grammar of AI-Thinking

Vannevar Bush imagined external memory systems as amplifiers of human reasoning (Bush 1945). But AI transforms externalization into a two-way channel: humans not only store meaning outside themselves but increasingly speak in ways machines can parse.

This is already visible:

• meeting notes formatted in bullet-point clarity “for the AI,”
• vocabulary shifts toward the language models favour (“delve,” “tapestry,” “unpack”),
• organizational writing adopting structures optimized for machine readability.

As the New York Times has noted, AI systems normalize certain lexical habits and stylistic forms (McCulloch 2023). This is precisely what Yuk Hui calls the grammar of thinking: each technology imposes a cosmotechnics—an implicit grammar that shapes how thought is expressed and what forms of thought feel “natural” (Hui 2016).

Thus, Cognitive Redistribution is not just about labour but about the reconfiguration of thought’s syntax.

---

# 6. Memetics and Technical Ideation: Dawkins Revisited

Richard Dawkins’ original concept of the meme as a unit of cultural transmission (Dawkins 1976) becomes newly relevant. Ideation is recombinatorial: new ideas are mutations and recombinations of cultural units.

For decades, humans were the exclusive hosts of memetic recombination. Now AI systems:

• ingest cultural corpora,
• recombine conceptual units at scale,
• generate variants humans would not find,
• surface analogies previously inaccessible.

This reveals a deeper truth:

ideation is technical.

It was never purely human; only the apparatus has changed.

Thus, Cognitive Redistribution reduces the human footprint in early ideational stages while increasing the human role in curation, situational judgment, prioritization, and ethical discernment. AI even supports decision-making by producing structured breakdowns—risks, parameters, trade-offs—rendering processes once intuitive more transparent.

---

# 7. A Brief Case: Design Teams and the Anxiety of Redistributed Cognition

Though not the focus of this article, design practice offers a telling example of resistance. Designers historically grounded their value in:

• generating ideas,
• synthesizing research,
• producing trend scans,
• recombining insights,
• facilitating divergent thinking.

These are precisely the domains now subject to Cognitive Redistribution.

As generative models accelerate ideation (a “Crazy 8” multiplied by 10), produce exhaustive benchmarking in seconds, and surface patterns across global corpora, designers encounter a mismatch between their inherited identity and the emerging cognitive ecology. Many respond by focusing on guardrails, ethics, and risk avoidance—a defensive posture that obscures the opportunity to redefine their role around higher-order interpretive work.

This resistance is not a design problem; it is a cultural symptom of a broader epistemic shift.

---

# Conclusion: Toward a Theory of Cognitive Redistribution

AI inaugurates a new phase in the history of technics: not merely the extension or exteriorization of cognition but its redistribution across human and technical actors. This redistribution alters:

• the ontology of ideation,
• the structure of agency,
• the grammar of thought,
• the circulation of meaning,
• the role of human expertise.

Understanding this requires moving beyond terms like “collaboration” or “augmentation” to a deeper theoretical account rooted in media philosophy, cognitive science, and posthumanism.

The question that now stands before us is not how to resist or regulate AI but:

How do we redesign our practices, institutions, and intellectual habits to inhabit the new cognitive ecology responsibly and imaginatively?

Cognitive Redistribution is not a threat.

It is an invitation to rearticulate what human judgment, creativity, and responsibility mean in a world where thinking is no longer ours alone.

---

# References (Chicago Author–Date)

Bush, Vannevar. 1945. “As We May Think.” The Atlantic.

Clark, Andy, and David Chalmers. 1998. “The Extended Mind.” Analysis 58 (1): 7–19.

Dawkins, Richard. 1976. The Selfish Gene. Oxford: Oxford University Press.

Hayles, N. Katherine. 2017. Unthought: The Power of the Cognitive Nonconscious. Chicago: University of Chicago Press.

Hui, Yuk. 2016. The Question Concerning Technology in China: An Essay in Cosmotechnics. Falmouth: Urbanomic.

Latour, Bruno. 2005. Reassembling the Social: An Introduction to Actor-Network-Theory. Oxford: Oxford University Press.

McCulloch, Gretchen. 2023. “Why AI Writes the Way It Does.” New York Times.

Simondon, Gilbert. 1958. Du mode d’existence des objets techniques. Paris: Aubier.

Stiegler, Bernard. 1998. Technics and Time, vol. 1: The Fault of Epimetheus. Stanford: Stanford University Press.

Stiegler, Bernard. 2009. Technics and Time, vol. 2: Disorientation. Stanford: Stanford University Press.

Stiegler, Bernard. 2010. For a New Critique of Political Economy. Cambridge: Polity Press.