An Article by M.R. (922 words, 5 min. read)
Artificial intelligence has become part of the daily routine of many writers, journalists, curators and researchers. It finds information within seconds, summarizes books, compares artistic movements and produces fluent texts in almost every language. It can even write about paintings with remarkable confidence. Reading some of these analyses, one may wonder whether art criticism itself could one day become entirely automated.
To explore this question, we conducted a simple experiment. We asked AI to analyze two paintings. The first was a portrait by George Condo. The second was a female figure by Hussein Madi. For each work, we also wrote our personal art journalistic text. The first was a conventional AI analysis. The second was written by a human in a more literary and emotional style. The differences proved surprisingly revealing.
George Condo: Describing a Face or Meeting a Soul?
The first version immediately adopted the familiar language found in countless exhibition texts:
“George Condo’s work presents a fragmented portrait that explores the complexity of human identity. The face is composed of overlapping features, combining multiple perspectives into a single figure. The expressive black lines contrast with the flat pastel color blocks, creating a balance between movement and stability. Rather than depicting a realistic likeness, the painting examines the many layers of personality, memory and emotion that coexist within an individual.”
Nothing in this paragraph is inaccurate. It identifies the fragmentation of the portrait, the formal construction, the relationship between line and color and the psychological dimension of the work. It provides useful information. Yet the painting remains an object observed from the outside. The words explain what the eye already sees.
The second version, written by one of our art journalists, approached the same portrait very differently.
“A face becomes a meeting place. Eyes search in different directions, profiles emerge and disappear, mouths hesitate between silence and speech. Every line seems to remember another life, another feeling, another version of the same soul. Identity no longer belongs to a single image. It grows, shifts, gathers fragments of memory until contradiction itself becomes a form of harmony.”
Instead of identifying Cubist influences or describing the geometry of the face, the text imagines an inner life. The fragmented features become memories. Contradictions become harmony. The portrait ceases to represent a person and begins to speak about all of us. The analysis enters interpretation instead of going into simple dry description.
Hussein Madi: Form or Feeling?
The same experiment produced an equally striking contrast with a painting by Hussein Madi. The conventional analysis by artificial intelligence reads almost like a museum label.
“The composition presents a stylized female figure reduced to bold geometric forms and sweeping curves. The body is simplified into large blue shapes outlined in black, creating a strong sense of rhythm and balance. Set against a vibrant yellow background, the limited color palette enhances the visual impact while emphasizing the harmony between figure and space. The work exemplifies the artist’s ability to merge abstraction and figuration.”
Once again, everything is correct. Geometry, rhythm, color, harmony and abstraction all belong to the painting. Yet one could almost replace Hussein Madi’s name with that of another modern painter and many of the sentences would remain convincing, although many words used in this analysis are very much common and cliches of AI.
The second version written by a human begins somewhere entirely different.
“The body has forgotten its weight. It curves, stretches and folds into itself with the ease of a melody carried by the wind. Every line seems to know where the next one is going, as though the entire figure were born from a single uninterrupted breath. Blue settles against yellow with the certainty of sea meeting sunlight, filling the composition with warmth and clarity.”
Here, geometry softly disappears behind sensation. Curves become movement. Color becomes light. The body becomes music. The painting is no longer analyzed as a composition but experienced as a living presence.
Culture Before Computation
A machine can recognize a face, identify a style and compare thousands of paintings in a fraction of a second. It can remember more information than any individual could accumulate during a lifetime.
Culture follows another path.
When a writer associates blue with the Mediterranean, silence with memory or a curve with the rhythm of a dance, these connections are born from lived experience. They come from childhood, literature, travel, music, conversations, museums, joy and loss. They belong to culture and memory before they belong to language.
Artificial intelligence can reproduce these associations because generations of human beings first transformed experience into words. Every poem, every novel, every exhibition review and every philosophical essay enriches the vocabulary through which AI learns to speak.
The Future of Art Writing
Artificial intelligence is already changing the practice of art criticism. It accelerates research, organizes information, compares sources and even suggests interpretations that deserve attention. Used smartly, it becomes an extraordinary companion for writers.
Yet the deepest encounter with a painting still begins elsewhere.
It begins in the silence between the artwork and the person standing before it.
A machine may identify what is visible. A human remembers what cannot be seen.
Perhaps the future will not ask us to choose between the two. It will ask us to combine them. One brings extraordinary knowledge. The other gives that knowledge a heartbeat. And as long as human beings continue to create, feel and write, artificial intelligence will continue learning its most beautiful language from them.
