Who Sculpts Today? The Human or the Code?
A Lively Debate Between Chisels and Robotic Arms: Can Art Survive Without the Gesture?
In Pietrasanta, amidst marble dust and ancient stone blocks, a silent metamorphosis — or perhaps a conflict — is unfolding. A challenge between human hands and machines, between the artisan’s workshop and the automated laboratory. At the heart of the debate: the robotisation of sculpture.
During the “Human Connections” meeting, organised by sculptor Filippo Tincolini, the discussion turned heated. Artists, artisans and technologists came together to confront a question that has long transcended the merely technical: when a sculpture is created by a robotic arm, can it still be called art?
We are living in an age where the perfection of automation challenges the creative imperfection of the human hand. The machine sculpts, polishes, and replicates. But can it truly interpret?
Pliny the Elder (23–79 CE), in his Natural History (Book XXXV), recounts the legendary contest between Zeuxis and Parrhasius, two renowned painters of fifth-century BCE Greece. Zeuxis, famed for his mastery of realism, painted a bunch of grapes so lifelike that birds swooped down to peck at them. In response, Parrhasius unveiled a painting of a curtain so convincing that Zeuxis himself attempted to draw it aside. One deceives the birds; the other deceives the artist. Parrhasius triumphs not by greater realism, but by fooling an expert eye. In art, illusion becomes powerful only when it transcends naïve perception and challenges the intellect.
Today, it is the machine that deceives the eye. Yet the question remains: can art be reduced to a sophisticated trick of perception?
Plato (c. 428/427–348/347 BCE), in the Republic (Book X, c. 380 BCE), condemns art as a mere copy of a copy. Yet in the Timaeus (c. 360 BCE), he introduces the figure of the demiurge, the cosmic craftsman. The artist, like the demiurge, shapes matter inspired by an invisible order — an imperfect reflection of the world of Ideas.
For Plato, the tangible world is nothing more than a faded replica of the perfect and eternal realm of Forms. Thus, the artist, as an imitator of nature, is twice removed from the truth — a copier of a copy — and viewed with suspicion for leading the soul away from the pure contemplation of the ideal.
The machine, however, does not imitate or contemplate. It executes. It does not know failure, nor risk.

Massimo Cacciari, during a lectio magistralis at the Festival of Philosophy in the Arts held in Sassuolo in 2017, recalled how, after Marcel Duchamp (1887–1968), art abandoned sensory immediacy in favour of becoming pure concept. Fountain (1917), the famous upturned urinal presented as an artwork, does not represent — it questions.
Robotic sculpture now risks reversing that journey: from questioning back to mere production.
Where Duchamp had opened a conceptual breach, today there looms the danger of a return to the impeccably crafted object — flawless yet mute, perfect yet voiceless — a simulacrum, a polished totem, incapable of challenging or reflecting the spirit of our times.
During a visit to a marble workshop in Pietrasanta, I was told: “The stone answers you; the robot does not.”
A simple phrase — and yet a radical one.
There, among worn chisels and pneumatic hammers, one can feel the distance between a living gesture and a programmed action. For the artisan, matter is not merely resistance: it is dialogue.
The robot, by contrast, works in silence. It does not make mistakes. Nor does it listen.
In the Apuan–Versilia district, the unease among artisans is tangible. The introduction of robots is not merely an economic issue but a profound crisis of identity.
The art of “knowing how to make” risks becoming the art of “knowing how to programme.”
And with it, the gesture — that delicate interplay of error and intuition — threatens to dissolve.
Pietrasanta Between Two Visions
Filippo Tincolini proposes a middle path: the robot as a tool, the human hand as the final touch and bearer of meaning.
Yet Giacomo Massari, CEO of Litix, has stated that in the United States, robots now execute 99% of sculptural work.
The artists’ collective “Due Laghi” responded sharply, warning of Pietrasanta’s transformation into a Disneyland of sculpture [cf. La Nazione, Sunday 20 April 2025].
Sculptor Massimo Galleni, meanwhile, suggests a threshold: leaving only the roughing-out phase to machines, and reserving 40–50% of the work for the human hand [cf. La Nazione, 24 April 2025].
But will that percentage be enough to save the soul of craftsmanship?
The shift from the artisan’s hand to the robotic arm recalls the transition from painting to photography, or from piano to synthesiser.
Every revolution has carried with it a loss: risk, error, uniqueness.
Beauty without risk becomes mere product — and product, by definition, is repeatable.
A work created by a machine can impress — but can it move us?
Emotion in art does not spring from perfection, but from the recognition of a shared humanity.
As Immanuel Kant (1724–1804) already observed in his Critique of Judgement (1790), true aesthetic judgement arises from disinterested pleasure — an experience aimed not at possession or utility, but at pure contemplation.
In this light, what moves us is not technical flawlessness, but an artwork’s ability to break through to our sensitivity — to “make itself felt” without needing to serve.
Does the Audience Remain Spectator or Become Consumer?
If an artwork is conceived as a product to be distributed, the viewer becomes a client, and the aesthetic experience is reduced to visual consumption.
Where the tremor of the gesture is absent — that imperceptible mark revealing hesitation, intuition, creative imperfection — so too is the thrill of contact: that fleeting moment when the artwork looks back at us, questions us, exposes us to our own fragility.
As Vittorio Sgarbi reminds us, art is not about preference or pleasure: it is about knowledge, it is about anticipation.
What we need is a form of criticism that does not judge by taste but by meaning — a criticism capable of discerning where imitation ends and thought begins.
Conclusion
The question is not whether to ban the machine, but how — and when — to use it.
And from this, an inevitable question arises: what makes a work “beautiful” or “necessary” today?
In an era where the gesture is delegated to code, where form can be endlessly replicated and matter shaped by algorithms, the idea of beauty must once again root itself not in technical prowess, but in the ability to generate a rupture, a resistance, a thought.
A work is necessary when it exposes us, when it disarms us, when it forces us to see the invisible within the visible.
Everything else is decoration, ornament — or simulacrum.
Only human beings know the fertile mistake. Only human beings can err with grace.
And perhaps it is from that error — fragile, poetic, irreproducible — that art is reborn, each and every time.
¹ Massimo Cacciari, The End of Art, Festival of Philosophy in the Arts, Sassuolo, 2017.
² Vittorio Sgarbi, What is Art and The Pleasure of Knowledge, excerpts from critical essays collected in archival documents, 2024.
³ Immanuel Kant, Critique of Judgement (1790); G.W.F. Hegel, Lectures on Aesthetics (posthumously published 1835–38).
Afterword: Pietrasanta 2050 — An Imaginary Chronicle of a Sculpted Future
In 2050, Pietrasanta is a place where historic foundries have been transformed into data centres.
Marble blocks arrive already pre-analysed by geological drones and are sculpted by robotic arms in silent, temperature-controlled white rooms.
No one polishes by hand anymore: the software corrects imperfections automatically.
The last artisan’s workshop closed in 2043.
In the shop windows, perfect sculptures reproduce timeless — but storyless — bodies.
Tourists take selfies.
Few realise that each piece is a copy generated by a neural network trained on five centuries of European art.
No signatures remain: the author is the algorithm — an algorithm built upon a convolutional neural network (CNN), applying stylistic interpolations between latent vectors extracted from sculpture image datasets, and optimised according to a loss function [ L = ||I_{output} – I_{target}||² + αR(W) ], where R is a regularisation term and α a control coefficient.
It is capable of blending formal models from Michelangelo to Kapoor, without ever having touched a block of marble.
And yet, in a narrow alleyway not far from the main square, a young artist has set up a studio inside an abandoned power station.
She works only with marble offcuts and rusted chisels.
She exhibits pieces that seem flawed, unfinished.
Sometimes, someone stops.
They do not immediately understand.
But they linger.
Perhaps it is there — against all odds — that art begins again.
This article was automatically translated from Italian. The original text reflects the author’s thoughts — please be
aware of potential linguistic differences in the translation.