Can AI Art Generators Assist Artists and Others with Aphantasia?

Some people can’t visualize. With AI generators, maybe they can.

What do a top Disney animator and an AI art generator have in common?They both don’t have a visual imagination. I’ll explain, but let’s back up first:

As someone with an acquired disability who finds AI to be a life changing technology everyday—in fact, it has become so second nature as to be nearly symbiotic; a virtual assistant, language therapist, a second brain — I’m always on the lookout for opportunities where AI can benefit humans.

With my majors in communication and psychology (focusing on cognition and perception), a lot of this involves identifying gaps and lacunae in how we express ourselves, and how AI can be an external platform to process and synthesize information. I’ve written at length about how AI text helps me manage my aphasia. But what would be the visual equivalent for AI art?

Aphantasia: A disorder that’s difficult to imagine

Picture this: imagine not having a visual imagination. It’s a conundrum, a headscratcher, right? At least for those of us who can imagine things easily (ironically enough!). But that’s the reality for up to 3.9% of the population.

“Aphantasia” refers to the inability to voluntarily visualize mental imagery.

People with aphantasia can’t see things in their mind’s eye. For example, if you ask them to picture a red apple, they won’t be able to conceive a mental image of it. They know what a red apple is, but they can’t generate imagery.

Is this a problem? Although many people with aphantasia have not let it stop them (and I’m about to discuss a Disney animator with aphantasia next), it does hold certain challenges, including affecting visual spatial working memory, autobiographical memory, and insight. So there are definitely areas where AI-assisted visual imagery could be of benefit.

From Aphantasia to “Fantasia”

Walt Disney famously said “Animation can explain whatever the mind of man can conceive”. So it may come as a surprise that Glen Keane, the lead animator behind Ariel and the Beast—despite being acclaimed as “one of the best animators in the history of hand-drawn animation”, and whose work was pivotal to the Disney Renaissance — has no visual imagination.

How Glen Keane creates art without visualisation.

It’s important to draw the distinction (no pun intended!) between not being able to elicit mental imagery, and knowing how an object, person or scene should look. When an artist with aphantasia creates art, they adjust and refine it in response to a reference or standard. They also have implicit knowledge, which they can draw upon without necessarily visualizing.

For example, Glen Keane doesn’t begin with a blank page; he starts with random scribbles (we’ll compare this soon to the random “Seed” or static used in AI generative art!) and then pares away what shouldn’t be there, and highlights what should, until the form he’s seeking emerges from the noise.

Even without envisioning a red apple, one can understand the concept: its structure, its texture. It’s actually a neat mental exercise — for those of us without aphantasia — to see if we can recall an object without eliciting a mental image (it’s like the classic conundrum: Don’t think of an elephant).

Refining Creativity: The Iterative Power of Feedback in AI Art

The process of creating art with aphantasia is iterative: the artist might compare what’s on paper to their understanding or expectation of the object. The comparison doesn’t have to be visual; it can be a sense that a line is “off” or “right”. Some aphantasics suggest they work more from a “feeling” than a clear image. They might have a sense of proportion, or an aesthetic that guides their hand, even if they can’t “foresee” the end result.

Perhaps aphantasia can even be an advantage when wielded purposefully, bestowing a critical detachment from the disjuncture between the artist’s intent or idea and the actual work being produced. Importantly, Keane is not trying to depict what’s in his head (and inevitably failing to match it, because what we create can never live up to the “perfect ideation”, the Platonic notion that all art is but imperfect copies of an Ideal Form). He is working it out by hand, on the page, in each moment. Each stroke on the canvas is both an act and an observation, and shapes the subsequent art.

The artist, in a sense, is using the canvas for an exterior cognitive process. Instead of wrestling with an internal image, the artist confronts a concrete representation, which can then be critically analyzed, modified, or erased.

“Look at this stuff, isn’t it neat?”

I would never suggest Keane could be “replaced” or “superseded by an AI art envisioning the early stages of animation (although Disney has never shied from technology, even when it impacts creative professions. Case in point: Disney was one of the first companies to embrace Xerox machines, which decimated the old Ink and Paint department, but is how we got 101 Dalmatians, a movie that would’ve been impossible without photocopying).

There may even be a paradoxical advantage to having aphantasia. Many of the Pixar artists, according to co-founder Ed Catmull’s survey of 540 of his colleagues, have what he describes as a “blind mind’s eye” (as opposed to their more visually-minded production managers). Growing awareness of has resulted in many people in creative fields “coming out” as aphantasic.

Additionally, the externalized, iterative process of creating with aphantasia can lead to works of art that are highly original, as they are not strictly tied to a predefined internal image, memory, or cultural expectation. The art becomes a living, evolving entity, constantly improving its own direction.

However, there could be more potential artists out there, people who feel they are not creative because they can’t “see” — or have been taught or told to believe they aren’t— who could flourish with the help of AI visualization.

Why encourage AI art when there are successful aphantasic artists?

Clearly, people with aphantasia can be professional artists. There is even an exhibition at the Centre For The Study Of Perceptual Experience that celebrates artists’ works with aphantasia. History is filled with countless artistic individuals who have overcome or even leveraged their unique neurologies to achieve greatness. But it would be impossible to measure how many others have been deterred. With AI art generators, people with aphantasia have one more tool to explore, create, and express themselves.

That’s the purpose of this article, the mission statement: to awaken artists. And if you recognise yourself in the description of aphantasia as someone who cannot visualize but wants to be a visual artist, I hope you’ll shrug off the preconceived traditional notions that artists must work from a “vision of genius”, and explore AI art as a way of getting something on the canvas.

If you want to describe your artistic concept in ways that are meaningful to you, and watch it develop before your eyes, who’s to say that isn’t creative Let’s get into how simpatico AI art is with the aphantasic creative process.

The affinity between aphantasia and diffussion

Last year, I wrote an article about how AI image generates work, and while much has changed since then, the basics remain the same. What I like to emphasize is that AI is not conjuring up images—there’s no “imagination” pulling pictures out of thin air. Instead, its more like an image “restorer”:

Diffusion is important to how AI art generators don’t “steal” artists’ work; there is a public misconception that AI art is like a collage, riffling through a stockpile of images to find pictures that match the prompt, then cutting them out and and sticking them back together to make a composite image.

This couldn’t be further from the truth. Think of diffusion like dropping ink into water; it spreads out, becoming more chaotic and random. A diffusion model in AI art does the reverse: it takes this “spread out” randomness and unspreads it, reconstructing it into a clear image.

From diffusion confusion to coherent design

So AI starts with a chaotic, meaningless image; a pattern of pure Gaussian noise represented by a Seed number—and works to restore this ‘damaged’ image to what it knows about how the subject should look. It does this by training on a dataset of real images and an understanding of how to reverse the process of turning an image into noise. Over a series of steps, the noise is gradually altered and refined to match the prompt (which is unrelated to the Seed — the image’s appearance is not directly decipherable from the seed value itself. The seed is essentially a starting point to the process).

The process of diffusion in AI art starts with a canvas of pure noise — like a sheet of paper covered entirely in random doodles. There’s no recognizable shape or form, just chaos. Sound familiar? It’s like Glen Keane’s scribbles.

The AI can’t envision interiorly what a cat, a tree, or a face should look like. Yet, through iterative feedback, it can generate convincing representations of these subjects. AI transforms a random pattern of noise into a coherent, structured image. It doesn’t “cut and paste” but rather “shapes and refines.”

Much like the aphantasic artist who doesn’t rely on an internal image but instead iteratively refines an external work, the generator doesn’t have a preconceived “picture” in mind. AI relies on feedback to refine creations.

How AI creates images without a visual imagination

Here’s a breakdown of how GANs work and how similar it is to aphantasia:

1. Two Neural Networks: GANs consist of a generator and a discriminator. The generator creates images, while the discriminator examines them.
2. Generator’s Task: The generator doesn’t have an “internal visualization” of what an image should look like. Instead, it generates a random Seed.
3. Discriminator’s Task: The discriminator evaluates the image produced by the generator and compares it to its training. It provides feedback to the generator about how “off” or “right” its generated images appear.
4. Iterative Feedback Loop: The generator and discriminator compete. Through iterations and feedback, the generator improves its output.
5. End Result: After many iterations, a well-trained GAN produces images that can be nearly indistinguishable from real images to the human eye.

(Another pivotal element is ‘reinforcement learning from human feedback’ (RLHF). This is basically where people gauge output and reward the model)

This process underlines the exterior-focused, feedback-driven nature of both human and machine learning in creating images in the absence of true visualisation. But how might the consanguinity between aphantasia and AI art be used in the art-making process to bring the artificial to life?

The real thing: Using artificial art in authentic ways

The way AI approaches generating images is uniquely simpactico with the way aphasphic artists develop their art. AI art could be incorporated into the aphantasic creative process to make sure that they’re not staring at a blank canvas—because that’s what AI imagery should be: a canvas, not necessarily the end result. Which is an important consideration for whether or not people who use AI art are indeed artists, or have any creative ownership over the images. There’s a process of refinement. It’s like sculpting; you don’t criticize the sculptor for not creating the marble.

How AI art could assist the aphantasic art-making process

• No More “Blank Canvas”: The person with aphantasia can provide the AI with detailed descriptions, colors, shapes, or even emotions they wish to capture. The AI can then generate images of these inputs.
• The Iterative Process: The ability to rapidly iterate with AI helps artists experiment with visual concepts quickly. Each iteration brings the creator closer to a visual that resonates with their stated intent.
• Collaborative Creation: Aphantasic artists may find it challenging to start with a visual concept. With AI, they can feed it their thoughts, however abstract or verbal, and see how the AI interprets them. The artist reacts to these suggestions: refining, or discarding them. The artist can mold and shape their vision externally with the help of AI.
• Expanding the Feedback Loop: Involving AI in the creative process expands the traditional feedback loop. Rather than relying solely on personal intuition or the critique of others, aphantasic artists can use the AI’s interpretation of their input as a form of objective feedback.
• Tangible Feedback: The externalization of the feedback loop provides tangible, immediate feedback. Instead of wrestling with an ephemeral, mutable internal image, the artist deals with a concrete representation.
• Mirror, Mirror: The external artwork — in reflecting back the artist’s words as an image — acts as a mirror to their internal state. It might reveal things about the artist’s subconscious, emotions, or their aesthetic inclinations that even they weren’t explicitly aware of.
• Diverse Aesthetic Possibilities: AI art generators can introduce any artist to styles they might not have considered, pushing them out of their comfort zones, leading to innovative and unexpected results.
• Liberation: By not being bound to an internal visualization, the artist might feel both the freedom of endless possibilities and the undefined.
• Empowerment: AI helps artists explore visual ideas that previously might have seemed inaccessible due to their inability to visualize mentally. It brings the unseen and unimagined into clear view.

For artists grappling with aphantasia, AI art generators could bridge over the visualization gap. For those who choose to embrace AI, it can enable the exploration of visual concepts that might have once felt out of reach.

Seeing is believing: AI art in action

I’ve seen this for myself. I’ve coached an AI art student in her 80s who was finally able to get a character she’d “thought about for years” down on the canvas, and it was a privilege to see it come to life. Recently I had a lovely comment from a reader about how her late mother would’ve used AI art as a creative outlet when her rheumatism got too bad to her artistic hobbies— admittedly this wasn’t aphantasia, but highlights another good application.

And it’s not just for artists

It’s not just potential artists that AI imagery can help with visualization. I’m a huge reader (I used to teach English lit) and one of my adult friends shyly asked me how to “get into reading”. She’d never enjoyed it, and at first I wondered if the obstacle was dyslexia. However, it turned out she could read fluently; she just could never “see” what was described on a page.

Everything clicked into place: she didn’t have a reading problem, it was a visualization one. I suggested to my friend that she try reading books that had been adapted to films, after she’d watched the movie first. And that she try to hold the memories of what she’d seen in her head as she read as a jumping off point — even if they conflicted with descriptions in the book.

By comparing and contrasting something she’d already seen, she was able to rely on her knowledge to “get into” the book and enjoy it; she now reads constantly (especially books that have been adapted by Netflix). Imagine if e-readers could provide tailored illustrations for you in real time as you read. [Hey, any developers/entrepreneurs want to work on this together?]

Let me outline some other practical applications for the little-known, but common phenomenon of aphantasia. Here’s how AI art could be helpful:

• Externalized Visualization: For those struggling to form internal visual images, AI can provide an external “visualization” tool. For instance, if someone with aphantasia wanted to remember or picture a particular scene, they could describe it to the AI, which could depict it for them.
• Educational Aids: In educational settings, AI-generated visuals can help students with aphantasia grasp concepts that are often taught with a reliance on visual imagination. For example, understanding complex spatial concepts or historical events can be enhanced with generated visual aids tailored to an individual student’s description or questions.
• Therapeutic Applications: Engaging with AI art mindfully might have therapeutic value. It may help those with aphantasia express feelings, ideas, or memories in a visual medium, aiding in personal reflection.
• Enhancing Memory: By turning memories into visual representations using AI, individuals with aphantasia can have a tangible “snapshot” of moments or ideas they can keep, aiding recall or sharing memories with others. By describing, refining, and adjusting inputs, they can arrive at a visual representation that resonates with their memory.
• Aiding Communication: For people with aphantasia who are trying to communicate a visual idea (like describing a dream, or a design idea), AI can help translate their verbal descriptions into more visual ones.
• Interactive Narratives: Storytelling and narrative exploration can be enriched using AI visuals. Instead of merely listening to, or reading a story, those with aphantasia can interact with AI tools to generate accompanying visuals, enhancing immersion. I’ve actually been creating this with a series of Custom GPTs called “Chat You Own Adventure”. There are three books so far. Please check them out!

Books that self-illustrate to enhance the reading experience

Currently I have a Gothic fantasy entitled “The Castle of Shadows”, a race against time called “Save The Titanic”, and my favourite, a visual detective story set in a phantasmagorical living art world titled “Framed for Murder”.

Framed For Murder is the most visual; the other two contain images that generate spontaneously, but you can also ask them to self-illustrate. All images arise from the reader’s own choices in the interactive narrative.

I love reading about your adventures in my interactive books, like this one and this one here. So if you do decide to play one, please feel free to blog your results, and your images! Especially on Medium. And tag me — I’m excited to read more of your experiences in these story worlds I created.

While AI art tools can’t replace or replicate the internal visual imagination, they can offer a valuable external platform for exploration, expression, and communication. By leveraging these tools, individuals with aphantasia can better navigate a world that places significant emphasis on visual concepts.

And finally, a manifesto on the “State of the Art, Art”:

AI art gets a bum rap. But it isn’t a cheat; it can be an active collaborative partner in the creative process, igniting a synthesis of ideas and visuals, and functioning as a catalyst for creative acts. It’s akin to having a muse that translates the abstract whispers of imagination into vivid, tangible forms, broadening the horizons of artistic expression and opening new avenues for creative exploration. AI brings the unseen and unimagined into clear view, encouraging artists to enter new territories of creativity.

Whether you can visualize in your mind or see only darkness, art is about expressing yourself. It’s about capturing a feeling, an idea, or a story. By using AI, anyone can align their unique perception of the world with the art they wish to create. While there’s still work to be done in forging fair agreements with artists whose work has been used to train AI, we must encourage and support these endeavors, championing a future where everyone can partake in the joy and fulfilment of visual connection.

Who is Jim the AI Whisperer?

Jim the AI Whisperer offers private coaching on how to write original and compelling content, as well as how to use AI generators to create stunning visuals. If you’re interested in discovering more, feel free to contact me.

I’m also available for podcasts, interviews, fine-tuning AI prompts, and creating prompt libraries and professional AI artistry for companies.

You might enjoy these related articles from Jim the AI Whisperer: