AI and the Myth of Creativity 3/4: Is Creativity Algorithmic?
We defend our mind as a sanctuary of originality. What if it were actually a factory? An extraordinarily sophisticated plant whose main business is to imitate, compress, and recreate the world.
In the previous articles, we explored the historical debate on machine creativity and deconstructed creativity itself as a social phenomenon. But one question remains, lurking in the shadow of our inner fortress: what about our own mind? We defend it as a sanctuary of pure originality, a magical place from which ideas spring forth. What if this sanctuary were actually a factory? An extraordinarily sophisticated plant whose main business is not to invent but to imitate, compress, and recreate the world.
Before comparing the human mind to AI, we must have the courage to look at our own functioning. What we're going to discover is that the processes we attribute to machines – calculation, simulation, generation from models – are at the very heart of what it means to be human.
The Lesson of the Rose: Thought as Compression
Let's start with the simplest experience: seeing a rose. This rose, in the world, is a thing of almost infinite complexity: the texture of its petals, the play of light on each drop of dew, the quantum state of every atom that composes it. When you "see" it, what really happens? Your eye and brain don't capture the "thing" itself. They perform a work of radical compression.
Your mind doesn't store the rose. It creates a model of it: an efficient summary, a lightweight version that retains the essential features (color, shape, scent) and ignores 99.99% of the information. To perceive is already to limit and imitate.
The Word Is Not the Thing
Now you want to share this experience. You say: "I saw a rose." The word "rose" is a second-level compression. It has further reduced your mental model to a simple linguistic symbol. When your friend hears this word, their mind does the reverse work: it decompresses the symbol and generates its own mental image of a rose.
Imagine the communication process as a game of telephone pictionary: there's the real-rose, the infinitely detailed original. You make a quick mental sketch of it (your mental-rose), a first imitation. You describe this sketch with a simple word: "rose." Your friend, from this word, draws their own mental sketch (their mental-rose). At no point was the original transmitted. All our communication rests on a process of compression and re-generation of imitations. We are machines that simulate each other's reality.
Memory: Replaying the Simulation
And memory? Remembering the rose you saw this morning is not "reviewing" a photograph stored in your brain. Memory is not a hard drive. Remembering is an act of re-creation. Your brain reactivates the compressed model it created and generates a new image, a new simulation of the original experience. And this process is imperfect.
Try to remember in detail the face of a loved one. Is the image as sharp as reality? Probably not. It may be faded, blurry, distorted. Some details are vivid, others have disappeared. Your mind "fills in the blanks," sometimes by inventing. This process of generation from an imperfect model is fundamentally similar to the way an AI generates an image from its own statistical model.
The "Digital" in Our Neurons: The Calculating Mind
"Alright," one might say, "but our mind is biological, organic, not a binary machine!" This is another myth of the fortress. Let's look more closely at a neuron.
A neuron receives thousands of electrical and chemical signals from its neighbors. It does only one thing: it sums these signals. If the total sum exceeds a certain threshold in a very short time, it "fires": it in turn sends an electrical impulse. If it doesn't reach the threshold, it does nothing. It's an "all-or-nothing" process. This is, at a functional level, a binary calculation. The information is not in a single "firing" but in the frequency of firings (the number of electrical spikes per second). It's frequency coding, a way of transforming analog information (the intensity of a stimulus) into a digital signal.
Our brain is not a computer in the sense we understand it, but at its most fundamental level, it is a calculating machine. It processes information constantly, in a massively parallel way, and most of these calculations are completely unconscious.
But I'm Not Just a Pile of Neurons!
This is the objection that surely comes to mind. And you're absolutely right. This feeling of being a unified, conscious, and creative "whole" is much more than the simple addition of billions of binary calculators. This mysterious "more" is precisely the phenomenon of emergence.
And this is where AI offers us spectacular proof. Artificial neural networks are based on a crude imitation of this same principle: simple "neurons" that sum signals and activate (or not) according to a threshold. The result? By connecting millions of these simple calculators, we see capacities emerge that no one programmed: recognizing a face, translating a language, generating an image. AI is experimental proof that reproducing the calculation function is sufficient to produce emergence and behaviors that we qualify as intelligent.
The argument "my mind is more than my neurons" is therefore not an argument against the computational vision of the mind. It's the most powerful argument in its favor. It shows that the essence of intelligence, whether biological or artificial, does not reside in the nature of its components but in the magic of their collective interaction.
The Myth of the Pure Original
We have reached the end of our introspection. The conclusion is troubling for our fortress-identity: our perception is a compression. Our language is a compression of this compression. Our memory is a re-generation from compressed models. The substrate of all this, the neuron, functions as a calculator.
Our mind is not a passive mirror of reality. It's an active model generator, an imitation machine, of whose functioning we are absolutely not conscious. The idea that we would be "original" creators and that AI would be merely a "copier" suddenly seems very fragile.
Now, and only now, having dared to look at our own mechanics, we are ready for the great confrontation. In the next and final article, AI and the Myth of Creativity 4/4: Beyond the Creative Fortress, we will examine what truly remains of the boundary between human and machine.
Article by Matthieu Ferry, clinical psychologist, published in French in Intelligences Plurielles on 20 July 2025, adapted in English for Copryce Lab in December 2025.
Copryce Perspective
Expertise as a Library of Models
This vision of the human mind as a “machine for compressing and generating” sheds new light on the nature of design expertise.
If our brain functions through compression and generation of models, then expertise is nothing other than a richer, more nuanced, more finely calibrated library of models. Faced with a client brief, the accumulation of experiences allows for the instant activation of dozens, even hundreds of models: what worked in one context, what failed in another, that subtle variation that made all the difference, that emerging trend glimpsed in the background. Over time, this mental library enriches and refines itself.
This expert “compression” allows them to immediately see what is relevant in a complex brief, to filter the essential from the superfluous, to generate solutions that “ring true” because they resonate with thousands of hours of experience compressed in their neural networks.
The observation that our memory “re-generates” rather than “replays” also explains why two designers with the same training can produce radically different solutions. Each re-generates from their own compressed model, shaped by their unique journey. This singularity in re-generation is precisely what makes the value of a particular expert perspective.
Understanding that our brain “calculates” constantly, unconsciously and massively in parallel, also removes guilt from the creative process. Inspiration isn’t magical; it’s the fruit of invisible neuronal work that mobilizes all available models. The richer these models (through experience, culture, practice), the more the unconscious “calculations” produce relevant results.
If intelligence is a question of emergence from the interaction of simple components, isn't expertise ultimately the quality and richness of the models we've built over time? And doesn't this progressive construction naturally justify that accumulated experience be reflected in the valuation of work?
