28/07/2023
Deciphering the Aesthetics of AI Art
A few weeks back, we at the Swiss AI Association launched an interesting challenge — to use the prowess of Artificial Intelligence in creating art. The task was simple yet captivating: create, compete, and be rewarded. With a plethora of AI tools at disposal, from heavy prompting software like Midjourney and Stable Diffusion to advanced graphics editing tools, the horizon of creativity was unlimited.
However, we soon stumbled upon an interesting dilemma: "What makes one piece of AI-generated art superior to another?"
To confront this challenge, we invited a diverse panel of experts from various backgrounds in Art and AI.
Our esteemed panel included:
Victoria Marchenkova, Artist and owner of an Art Gallery in Zurich
Denis Belkevich, General Partner at Fuelarts, Art Economist and Art Advisor
Rea Stark Rajcic, Design Evangelist & Entrepreneur
Rok Zaloznik, AI Expert and Swiss AI Association Co-founder
Sergali Adlibekov, Banking expert and consultant, Swiss AI Association Co-founder
The brainstorming session with the experts unraveled a blend of anticipated and surprising criteria that serve as a roadmap to evaluate AI art. Here's a look at these key standards:
Market Presence and Accessibility: Evaluates the artist's presence within the AI art marketplace. This criterion includes measuring the frequency and diversity of their art listings, whether they are present on curated platforms that are accessible to the majority of art collectors, and whether their work is featured on open marketplaces that reach wider audiences. It also assesses the regularity of the artist's participation in art fairs, exhibitions, or online events, further increasing their accessibility.
Narrative Depth: Analyzes whether the AI-generated art tells a compelling and successful story. Art is more than just aesthetics, it's about the story it tells and the emotions it evokes.
Technological Mastery: This measures the artist's understanding of the generative model and their ability to effectively use prompts to achieve desired results. It considers if there's a comprehension of the AI's processes.
Emotional Resonance: Evaluates the emotional depth and impact of the AI-produced art. Good art should stir emotions in its audience.
Conceptualization & Innovation: Assesses how the AI-generated art reflects contemporary art concepts, and how it integrates new ideas such as "deep art injection" and hidden knowledge within the image.
Brand Continuity: This judges the consistency in the artist's work, their ability to produce a series of related images, and the development of a recognizable brand. A single image is insufficient; the artist must demonstrate a coherent body of work.
Audience Expansion: Determines how successful the art is at attracting and involving a new audience, beyond traditional art appreciators. It measures the art's potential for audience growth and diversification.
Reference & Originality: Scrutinizes the choice of reference images for historical art understanding, and the use of original graphics or photos as a starting point. This criterion blends respect for artistic tradition with the innovation of new visuals.
What do you think of the proposed criteria?
Would you add some other criteria? Do let us know in the comments!
