AI art is more than just a prompt.

AI art is more than just a prompt.

Since I've been involved with AI-generated images, I've heard quite a few opinions. Comments like "AI isn't art," or "How dare you earn likes for prompts?" were among the milder ones. However, there have been significantly more unpleasant remarks. While these don't really affect me, as they obviously come from trolls looking to provoke, their ignorance regarding art is comparable to that of the Pope running an internet startup.

It's true that generating AI art often requires just a prompt. Many online services and mobile apps offer nothing more than that. Generally, a good image is produced, as the AI (barring errors) usually creates quality images because it has been trained with high-quality pictures. The main difference often lies in the complexity of the generated image, the diversity of the composition, and how well the result matches my own vision.

If I'm satisfied with what the AI has generated, the prompt may suffice. But when I have a specific goal or concept in mind, it quickly becomes clear how often the AI misses the mark, no matter how precisely the prompt describes the scene. It's about positions in the image, about consistency in the scene. Often, many different starting points must be tried until a result is achieved that comes close to what I had imagined - if only approximately.

The uniqueness of my workflow lies in the flexibility and the range of tools I use to realize my creative visions. Fundamentally, I use ComfyUI, a tool that might seem daunting at first glance. However, those who take the time to learn it soon appreciate its comprehensive capabilities without significant limitations.

My creative process is dynamic and integrates various software solutions:

  • Idea Drafting: I often start with a quick scribble on the iPad (using ConceptApp) to capture my ideas.
  • Refinement: Then, I refine these sketches in Photoshop, adding details or fine-tuning the design.
  • 3D Modeling: I create a 3D scene using Blender. This helps me generate a depth map that adds additional dimensionality to my final product.
  • Integration in ComfyUI: I use the previously created elements as references in ComfyUI to generate a preliminary image.
  • Detail Improvement: If necessary, I further refine certain sections of the image in ComfyUI or split the image in Photoshop to enrich it with more details.
  • Scaling and Final Refinement: After adding the details, I scale down the image and perform a final refinement with AI.
  • Post-production: Finally, the post-processing is done in Photoshop, including cleaning up and polishing the image.

This workflow may seem complex, but it's a normal part of my work. The end result is an image that precisely meets my expectations and goes beyond the limitations of simple prompts.

Good Art is Independent of the Path

The debate about artistic authenticity and the use of new technologies is as old as art itself. Every time a new technology finds its way into the creative landscape, a wave of skepticism arises among traditional artists. This skepticism is an echo from the past that runs through the centuries, from the invention of photography to the digitization and artificial intelligence in art.

At the heart of the debate is not the complexity of workflows or the novelty of tools, but the question of what constitutes art and who can create it. It's a misconception to think that top-tier artworks necessarily require complex processes. The truth is that the complexity of an image does not always determine its value. Instead, it's about how the artist expresses their vision, regardless of the means used.

A look at history reveals a pattern of initial rejection of new technologies:

  • With the invention of photography, photographs were dismissed by traditional artists as inferior art because they didn't reflect the laborious handiwork required for paintings and sculptures.
  • The introduction of digital cameras met resistance from analog photographers who valued the art of film photography and darkroom work.
  • As smartphone photography became more widespread, these images were often dismissed by professional digital photographers as simple snapshots lacking the depth and technical knowledge required for traditional photography.
  • Digital art has also faced criticism, with claims that works created with digital tools like Photoshop lacked the same "soul" as traditionally created art.

These examples highlight a recurring skepticism towards the new, yet they also show that the definition of what is recognized as art continually expands over time. Art lies in the eye of the beholder, and what ultimately counts as art depends not on the challenges faced along the way, the financial expenditure for tools, or the learning curve. True art is created by creative individuals who understand how to bring their visions to life with the available means. And whether something is viewed as art remains a subjective decision that each observer makes for themselves.

The debate over the value and authenticity of art created with the help of new technologies is a testament to the dynamic nature of art, constantly evolving, driven by creativity and the relentless pursuit of new forms of expression.

In the end, AI will establish itself as another optional building block in the workflows of creative people.