Why Does AI Art Look Soulless and Does It Matter for Buyers

“That looks AI.” Three words. And in 2026, they function as a consumer-level insult.

This isn’t aesthetic philosophy anymore. It’s a business problem with measurable consequences. CPG brands are losing shelf attention. Indie game studios are watching Kickstarter backers walk. Comic festivals are writing formal exclusion policies. The soullessness isn’t imagined — it’s detectable, even by people who can’t explain what they’re seeing.

AI image generation has become cheap and fast. But “cheap and fast” is exactly what modern consumers have learned to smell. The same way “bad Photoshop” became shorthand for low effort in the 2000s, AI-generated visuals are now triggering a nearly identical disgust response — often subconsciously.

So what’s actually driving that perception gap? What does the data say about how buyers respond? And where does legitimate AI workflow support end and brand-damaging over-reliance begin?

Key Takeaways

• Consumer perception research shows buyers detect AI-generated visuals even when they can’t articulate why — triggering associations with cheapness and inauthenticity.
• According to Dyno Creative, CPG packaging has roughly three seconds to earn a second look — a window that closes faster when visuals blend into the AI-generated sameness saturating an entire product category.
• The Perth Comics Arts Festival formally banned AI-generated materials in 2026, signaling that institutional resistance has moved from informal pushback to enforceable policy.
• Art director Shane Lewis won back a client considering free AI tools by competing on craft and production reliability — not price — demonstrating that human authorship carries a measurable premium buyers will actually pay for.

The Current Moment Didn’t Appear Overnight

Midjourney launched publicly in mid-2022. DALL-E 3 followed in 2023. By 2024, AI image generation was embedded in Adobe Creative Cloud, Canva, and dozens of no-code tools. The barrier to producing a “decent-looking” image collapsed almost entirely.

That accessibility created a glut. Social media feeds, product listings, indie game storefronts, and brand marketing materials flooded with AI-generated imagery. Volume scaled. Quality — in the craft sense — didn’t.

The critical shift happened in 2025–2026. Consumer familiarity with AI outputs crossed a recognition threshold. People started categorizing AI aesthetics the same way they once categorized stock photos — identifiable, interchangeable, slightly off. According to Australian visual artist Jess Harwood writing in The Guardian, creative communities began adopting the term “CRAP” — Computer Rendered Artificial Pictures — as an internal descriptor, signaling how the professional art world was processing the shift.

Institutional responses followed. The Perth Comics Arts Festival issued a formal policy in 2026 stating it “will not knowingly promote AI-generated materials, nor will we allow any such work to be a part of our festival.” That’s not a blog post. It’s enforceable gatekeeping from a credentialed institution.

Meanwhile, counter-market signals started appearing in consumer data: sold-out IMAX screenings, vinyl record sales growth, collectible packaging editions, limited art prints. Each represents consumers actively paying a premium for perceivably human-made work. The pattern is consistent enough to be strategic, not coincidental.

The question for brands, studios, and product teams right now: does the cost savings from AI-generated visuals actually outweigh the brand-trust erosion that comes with it?

Why the “Soulless” Perception Exists

The word sounds subjective. The mechanism behind it isn’t.

Human-made art carries embedded evidence of decision-making. Brush pressure variation. Color choices that break convention deliberately. Compositional risks. Imperfections that signal the artist chose to leave them. These aren’t just aesthetic features — they’re signals that a mind worked through a problem and made specific, non-obvious choices.

AI image generation works differently. It produces statistically likely outputs given a prompt. The result is the visual equivalent of averaging thousands of reference images — technically plausible, but drained of the micro-decisions that make human work feel specific to a moment or a person.

According to Harwood’s piece in The Guardian, when a researcher offered to train an AI model on her specific artistic style, she declined entirely — because separating the artist from the artistic process fundamentally undermines what makes art meaningful. The emotional connection audiences feel flows through the human struggle behind creation, not just the output itself.

That’s not sentimentality. It’s signal theory. Buyers reading “AI” in an image are reading the absence of that struggle.

The 3-Second Shelf Problem

This is where the question moves from philosophy to P&L.

According to Dyno Creative, CPG packaging gets roughly three seconds to earn a second look. When a product’s visual language blends with the AI-generated sameness of competing products, it doesn’t just look uninspired — it’s effectively invisible.

Art director Shane Lewis, whose credits include key art direction for Weird Al Yankovic’s Black & White & Weird All Over art book, draws a direct parallel to the post-Nike-swoosh era: when every brand adopted the swoosh aesthetic, the aesthetic stopped working for any of them. Widespread AI adoption is approaching the same saturation point. Visual differentiation — the actual function of brand design — erodes category-wide when everyone pulls from the same generative pool.

Lewis retained a client who was considering free AI tools by pitching production expertise and reliability. He competed on craft, not cost. He won. That’s a concrete business outcome tied directly to the perception gap.

This approach can fail when the product category is low-consideration and price-driven — commodity goods where visual craft genuinely doesn’t influence purchase decisions. But for anything requiring brand loyalty or emotional affinity? The calculus flips fast.

The Authorship Problem Nobody Has Solved

Three parties have a claim on AI-generated art: the system that generated it, the person who wrote the prompt, and the developers who trained the model on existing human work without uniform consent or compensation.

According to Harwood’s analysis, this creates unresolved authorship questions that extend beyond ethics into commercial risk. If an AI model was trained on a living artist’s work without permission, the brand using that output carries some exposure. Legal frameworks around this are still forming — but they’re forming fast, and brands that leaned hard on AI-generated output during the gray period won’t be able to manage that retroactively.

The Perth Comics Arts Festival’s 2026 policy isn’t just aesthetic preference. It’s risk management. Festivals, publishers, and platforms accepting AI work without provenance verification are increasingly exposed as litigation precedents develop.

AI Tools That Help vs. AI Tools That Hurt Brand Trust

Not all AI use in creative workflows carries the same risk. The distinction matters.

The pattern is clear. AI performs well as a support layer for human creative decision-making. It damages brand trust when it replaces that decision-making entirely — specifically in customer-facing contexts where visual differentiation carries commercial weight.

What to Actually Do With This

For product and brand teams: The question isn’t “can we use AI art?” — it’s “where in the customer journey does visual craft signal brand investment?” Packaging, hero imagery, and character design sit in high-perception zones. Using AI-generated output there is a cost saving that shows. Internal decks, mood boards, and rapid concepting are low-perception zones where AI adds genuine speed without trust erosion.

For indie game and publishing studios: The BoardGameGeek community discussion around AI art reflects a real purchasing behavior shift. Backers on crowdfunding platforms are increasingly treating AI-generated game art as a negative signal about project quality and creator commitment. If your campaign art reads as AI-generated, you’re not just facing aesthetic criticism — you’re managing a trust deficit before the campaign even closes.

For the design industry’s pipeline: Dyno Creative flags a structural concern that gets underreported: AI is absorbing entry-level design work — flyers, brochures, simple layouts — which eliminates the apprenticeship pipeline that develops skilled senior designers. Studios relying on AI for junior-level output now may face a craft deficit in three to five years when they need senior talent and find the training ground no longer exists.

What to watch next: Legal frameworks around AI training data consent are tightening in the EU and Australia. Adobe’s Content Credentials initiative is moving from opt-in to default on major platforms. When buyers can see a “made with AI” tag on packaging or editorial art, the implicit signal becomes explicit — and brands that have leaned hard on AI-generated output will face a reckoning they can’t manage retroactively.

Where This Lands

Consumers detect AI aesthetics through the absence of human decision signals, even when they can’t name what they’re seeing. The 3-second shelf window closes faster when visuals blend into category-wide AI sameness. Institutional gatekeeping is formalizing what was previously informal resistance. And AI tools add value as internal support layers — they erode brand trust as customer-facing replacements for human creative judgment.

The craft premium is real. Shane Lewis winning a client on reliability over free AI tools isn’t an outlier. It’s the market starting to price human authorship back in.

Audit where AI appears in your customer-facing creative. If the answer is “everywhere,” that’s the risk exposure. Human judgment directing AI tools is a workflow. AI replacing human judgment entirely is a brand liability.

What’s one customer-facing asset in your stack that’s currently AI-generated — and what would it actually cost to replace it with something that could differentiate?

References

1. AI ‘art’ is boring, soulless theft – and when I see it as an artist I see red | Jess Harwood | The G
2. Soulless projects - a problem with AI art | BoardGameGeek
3. AI Art vs. Real Art: Why Your Customers Can Tell (Even When They Can’t Say Why) - Dyno Creative

Photo by Igor Omilaev on Unsplash