GenAI Art Is the Least Imaginative Use of AI Imaginable

Automation, Labor, and “Free Time”

• Discussion starts from John Cage’s quote about tools that create more work versus save it.
• Household tech examples (dishwashers, washing machines) show mixed real-world time savings: some say dishwashers barely help and add ritual; others claim big gains, especially for laundry.
• Several argue that large-scale automation (tractors, combines, Haber–Bosch) is what makes non-survival activities like advanced research or art possible.
• A recurring worry: saved time often gets captured by addictive platforms (TikTok, algorithmic feeds) rather than meaningful creation, and chores themselves can be valuable “thinking time.”

Speed, Interactivity, and Image Generators

• Some want image models fast and dense enough to act like “infinite Google Images” or a Library-of-Babel browser: 20+ variations instantly, interactively.
• Others respond that current systems are already fast enough for many uses, and obsession with volume misses that there’s “millions of useful things” already possible.
• Technical caveats: fastest models have lower quality; upscaling changes images; quality/speed tradeoffs remain.

Is GenAI Output “Art”?

• One camp: GenAI can be a powerful creative tool. Users describe spending significant time prompting, curating, iterating, and feeling genuine ownership and pride, especially in music tools (e.g., Suno, Udio).
• Opposing camp: most AI images/music are generic “slop” with no intention or human struggle; typing a sentence is not the same as developing craft over years.
• Some assert AI cannot create art because it lacks conscious self-expression; any art must come from human post-processing and curation.
• Others contest that definition, noting:
  • Historical views of art often blurred with “craft.”
  • Many human works (stock photos, hotel art, Kinkade, t‑shirt graphics) aren’t deep expressions either but still serve purposes.
  • Patronage, creative direction, DJing, and composition already separate “vision” from fabrication.

Art vs Illustration, and the Web “Slop” Problem

• Distinction proposed: art as self-contained meaning vs illustration as something that derives meaning from accompanying text. Most AI blog images are placed firmly in the “illustration” bucket.
• Content creators admit using AI images mainly to satisfy platform/SEO incentives (“engagement,” hero images), not artistic goals.
• Several readers now actively avoid AI thumbnails, associating them with SEO spam and low-quality text; analogy to recipe pages bloated for Google.
• Concern: generative filler will flood the open web, degrading search more than chatbots themselves.

Process vs Product in Creativity

• Many participants resonate with the view that the process (practice, struggle, refinement) is central—using analogies to sports training, mountaineering, woodworking, and music performance.
• Others challenge this as elitist: the finished artifact and the audience’s experience also matter; AI can accelerate iteration and exploration.
• A nuanced middle view: there’s a hierarchy of satisfaction:
  • High: struggle and succeed (traditional craft).
  • Medium: low effort, good result (GenAI-assisted).
  • Low: struggle and fail.
    For skilled artists, being pushed from the first to the second tier feels like a loss of meaning.

Ethics, Theft, and Displacement

• Some bluntly label prompt-based image generation “theft,” since models are trained on unconsented human work; countered with analogies to photography and remix culture.
• Broader fear: tools are built for managers to replace creative labor, not to empower craftspeople—hence especially strong backlash from artists and musicians compared to programmers.
• Others note that previous technologies (cameras, digital tools) also displaced crafts but enabled new forms (film, mass photography). Debate remains whether GenAI will similarly create genuinely new art forms or just mass-produce surface-level facsimiles.

Desired Future: Assistive, Not One-Click, Tools

• A distinct subgroup has little interest in “text-to-finished-piece” systems but does want granular, interactive assistants:
  • Tools that refine a stroke, fix proportions, suggest variations, or tighten musical timing as editable layers or MIDI.
• They argue the real potential is in collaborative, fine-grained augmentation of human skill, not in fully automated content vending machines.