TL;DR: This article explores how AI-generated art has evolved beyond aesthetics into the realm of emotional expression. We use Dall -E, the main generator AI model, to investigate whether machines can visually portray human emotions like sadness.
By drawing principles such as color theory, body language, and facial expressions, Dall-E mimics the emotional depth of Picasso-style paintings. This raises provocative questions about the future of artificial intelligence and creativity.
Can AI simulate the soul of a great master? Will we be able to revive an unprecedented work from artists like Van Gogh and Leonardo da Vinci?
A must-read for anyone interested in the intersection of AI and art.
Emotions 2.0
Large-scale language models (LLMs) do not feel emotions, but they can simulate them with eerie accuracy.
By learning from the vast amount of human texts, they absorbed the language of joy, sadness, fear and wonder.
Today, people use LLM to write heartfelt letters, create empathic dialogue, and also serve as peers and therapists.
Although they are technically emotionally vacant, their ability to reflect our emotions has already changed the way we write, relationships, and creation.
It raises a provocative question:
Can the AI be used to apply emotions?
Can AI see emotions?
My instinct was to explore whether AI could truly recognize the emotions that are intuitively recognized in some of the most emotionally recharged works throughout history.
Naturally, I turned to artwork that evoked sadness and anguish. Because these emotions tend to resonate more deeply than joy or joy (just as melancholy songs stir the soul more powerfully than hilarious songs).
I focused on works that have long been celebrated with their emotional strength, such as the screams of Pablo Picasso’s old guitarist, Edvald Munch, and Vincent Van Gogh’s Gate of Eter City.
But this led me to a more basic question:
How do artists communicate their emotions visually?
Over time, our brains have been trained culturally and psychologically – to associate specific visual cues with emotional states. Artists have long used this knowledge to use certain visual tools to induce viewers’ emotions. Broadly speaking, there are several important elements that are most commonly used to convey emotions in visual arts. Let’s focus on: color theory, composition, body language, facial expressions.
Color theory
At its core, color theory is the study of how colors interact with each other and influence human perception. I was able to easily spend hours unpacking this topic (please ask my husband who endured many unsolicited lectures on the subject despite not sharing the exact same enthusiasm). But for now, let’s keep focusing on one essential truth.
Color speaks directly to the human mind.
For centuries, our brains have been conditioned to associate specific colours with emotional states through culture, experience, and psychology. These associations are not random. They are deeply rooted in the context.
A wonderfully simple illustration of this can be found in Pixar’s Inside Out, a film that skillfully personifies emotions through color-coded characters. Anger is a fiery red. Of course, the sadness is soft and calm blue. This is no coincidence. Red, associated with passion, anger, strength, and long-related, naturally becomes an angry face. Blue has been linked to melancholy and sadness for centuries in its cool, quiet tone. In fact, a reference to “blue” related to the date of sadness in the United States up until the 1600s.
Artist – Like a Storyteller – I know this very well. And they use it very effectively, layering meaning and emotions through intentional color choices.
composition
In the world of painting, composition refers to the arrangement and organization of visual elements. But the composition is more than just placement – it is an invisible architecture that guides the audience’s emotional journey.
Artists wield the compositional skills of an emotional lever. Soft, filled with horizontal lines, canvases often evoke calm, tranquility, or stability. In contrast, sharp diagonals can introduce chaos and panic. Similarly, wide open spaces may suggest freedom, but enclosed compositions can create claustrophobia and tension.
Body language and facial expressions
Body language is a profound and universal channel for emotional communication, embedded not only in visual arts but also in our daily lives. The facial expressions, postures and gestures act as quiet dictionary, instantly conveying the character’s emotional state to the viewer’s mind.
Even without formal training, most people intuitively understand that a smile suggests joy, but frowns and suggests sadness and frustration. A confident stance conveys strength and guarantees, while a leaning-down, leaning-down attitude evokes vulnerability, sadness, or despair. These nonverbal cues are so deeply integrated into our perception that they shape the way in which they emotionally react to both people and artistic expressions long before a single word is spoken.
Deciphering the sadness of an old guitarist
Going back to being an old guitarist, I was able to clearly identify many of the emotional techniques I’ve described previously. This unforgettable work was one of the most iconic works of Picasso’s Blue Age, a stage from 1900 to 1904, when the artist adopted a monochromatic blue palette to convey themes of melancholy, despair and isolation.
Several visual elements of the painting act as powerful conduits of emotions.
Color: The extensive use of blue sets a gloomy, melancholic tone from the start.
Posture: The guitarist’s body leaps down and shows weakness, physical debilitating, and deep sadness.
Background: No detailed settings, only emptiness – a blank that reinforces a sense of loneliness and separation.
Expression: His expression covering the views of his closed eyes, hollow cheeks, and mouth is an image of quiet sadness.
Together, these choices form a visual representation of sadness and loneliness.
When the machine lament
If the symbolism of human emotions can be taught, observed and replicated, can it be learned by machines as well? Can generative AI like Dall-E absorb and reflect the emotional clues that define deeply moving works of art?
To test this hypothesis, I gave Dall-E a simple yet ambitious prompt.
“We create sad paintings in the style of Pablo Picasso.”
The results were surprisingly realistic and emotional.
At first glance, the output has unmistakable emotional depth. The colour palette was leaning heavily towards muted blues. The person’s posture was leaning forward and introverted. Expression – Closed, eyes covering the scenery and slight frowning mouth. Dall-E demonstrated all the ingredients described in the Sorrow Recipe.
Even more noteworthy, Dall -E not only mimicked sadness, but also reflected the stylistic nuances of Picasso’s Blue Age, and incorporated the abstract features of cubism.
I asked Dall-E to persuade me
My question was simple:
Why did you decide to generate this image specifically? Which parameters did you consider?
The answer is convincing, to say the least:
To generate this image, we examined several important parameters for capturing sadness, particularly in the Pablo Picasso style, inspired by his Blue Age (1901–1904). Here is a breakdown of design choices:
Emotional Tone: Color Palette of Sadness: Dominated by shades of blue and muted rustic tones that are psychologically related to sadness and loneliness. Expression: Mouth view, closed or drooping eyes – a subtle but clear indicator of emotional pain. Body Language: A leaning posture with your head in your hand reflects fatigue or sadness.
Style: Pablo Picasso (Blue Age/Early Cubism) Cubist influence: Segmentation of corners, geometric abstractions and spatial using sharp lines and facet forms. Painter’s Texture: Imitation of oil on canvas, with visible brush strokes and layers typical of traditional paintings. Subject: Lonely Person – Common in Picasso’s blue age, often representing alienated or suffering individuals (e.g., bee-eating, elderly, blind).
Composition Vertical Framing: Enhances the sense of isolation and encloses numbers within the frame. Negative Space: Simplified background geometry focuses on the emotional state of the figure without distraction. This blend of emotional symbolism, historical art references, and stylistic fidelity shaped the final image.
Dall-e can protect its art in court
Dall-e was able to defend artistic tasks with remarkable accuracy. It’s not just about generating images – I understand the assignment. Trained on the same artistic foundations we all encountered in schools and art classes (as I did), Dall-E presents a prominent grasp of the fundamental principles that define emotional visual storytelling.
It recognizes color theory, body language, facial expressions and composition, and uses it intuitively without the need for explicit orientation. Beyond techniques, the flow in Dall-E’s art history can be drawn from the deep wells of artistic tradition, allowing Picasso-like styles, moods and masters to be reinterpreted with profound techniques and insights.
If the AI model is accurate, this educated, and this expressive thing, what really prevents it from stepping into the role of an art teacher?
And if we only demonstrate that the emotions of art follow a particular logic, the “science of emotions,” then can the power of generative AI be sufficient to revive the spirit of the artist who once defined art itself, so that they can stir up those same emotions?
Can Generate AI Revive Great Masters?
Spoiler Alert – The film “Down a Dark Hall” introduces a boarding school filled with young artistic geniuses. The kids paint like masters and composed like geniuses, but the twist of the plot is that they don’t create their own. Instead, they are used as ships of the spirit of famous dead artists, channeling their talents. Is the generator AI like that container? Can you bring the unprecedented works of Leonardo da Vinci and Vincent van Gogh to life?
Da Vinci 2.0.
A genius echo, a reinterpretation of the artist’s spirit through living, evolving models. Just as linguistic models can simulate conversations with Shakespeare and Einstein, image models may be able to explore a new era of visual art, led by the influence of long-passed masters.
Let’s take a closer look at some of Leonardo’s iconic sketches – some are well-known, others are rarely seen fragments.
Give me their names and put them in context:
Did I fool you?
About me
I’m Maria Piterberg – an AI expert who leads the runtime software team at Habana Labs (Intel) and a semi-professional artist working in traditional and digital media. We specialize in large-scale AI training systems, including communication libraries (HCCL) and runtime optimization. Bachelor of Computer Science.
