Reshaping Perception Through Technology: From Ancient Script to Large Language Models

As large language models reshape how we create and access information, questions arise about how to frame their role in human creative and cognitive life. We argue that AI is best understood not as artificial intelligence but as a new medium – one that, like writing before it, reshapes perception and enables novel forms of creativity. Drawing on Marshall McLuhan’s insight that “the medium is the massage,” we trace a lineage of technologies – from DNA and the nervous system to symbols, writing, and now LLMs – that mold cognition through a shared logic of flexible unfolding and co-creation. We observe that as technologies become more externalized and decoupled from physiology, they introduce both greater creative potential and greater risk of inauthenticity and manipulation. This tension is acute with LLMs, but not unprecedented: ancient responses to writing reveal a recurring human tendency to project intelligence onto powerful new media. Rather than viewing AI as a competitor, we propose framing it as a medium that foregrounds artistic skills: aesthetic judgment, curation, and the articulation of vision. We discuss implications for education, creative practice, and how society might adapt to this new medium as it did to writing.

💡 Research Summary

The paper reframes large language models (LLMs) not as “artificial intelligence” but as a new cultural medium that reshapes perception, cognition, and creativity. Tracing a lineage from DNA and the nervous system through symbols and writing to LLMs, the authors argue that all these technologies share a “fold‑unfold” logic: information is physically folded into a substrate (DNA helices, neural synapses, ink on paper, digital weights) and later unfolded to generate new forms and meanings. Drawing on Marshall McLuhan’s insight that “the medium is the massage,” the paper shows how each medium’s spatio‑temporal properties directly sculpt human experience, turning the medium itself into the message. Historical reactions to writing—Plato’s warning of “pseudo‑intelligence” and the magical aura ascribed to texts—are presented as early examples of the same tendency to project intelligence onto powerful new media, a pattern that repeats with LLMs.

Neuroscientific evidence is marshaled to illustrate the brain’s constant spontaneous activity, neuroplasticity, and “auto‑sculpture” of synaptic architecture, suggesting that the brain already functions as a self‑writing system. Experiments such as grafting a third eye onto a frog embryo or the rubber‑hand illusion demonstrate how external extensions are incorporated into perception, mirroring how LLMs become extensions of human thought.

Three key implications are derived. First, LLMs amplify the externalization of thought, inviting users to engage in co‑creation rather than competition. Second, as a medium becomes more detached from physiology, its creative potential rises while the risk of manipulation, inauthenticity, and loss of agency also grows. Third, the dominant artistic skills shift toward aesthetic judgment, curation, and articulation of vision—abilities that cannot be outsourced to the model.

The authors conclude with practical recommendations for education, artistic practice, and public policy: treat LLMs as a new sensory device, cultivate critical‑curatorial competencies, and develop ethical frameworks that preserve human agency while leveraging the unprecedented generative capacity of this medium.