Design, taste, and the cost of sameness
Why Authenticity, Curiosity, Aesthetics And A Sense Of Humour Are Disappearing From Graphic Design, And Humans Are Putting Up With Slop
Taste has become the billion dollar keyword. But what do we mean by taste?
What a person finds beautiful, funny, sacred carries their culture inside it, their beliefs, the memories, the tiny details they listened and observed. Nobody arrives at an aesthetic sense from nowhere. It comes from things we like and don’t like, things we argue with, rebel against, and slowly make personal. Be it aesthetics we appreciate or values we live by.
Authenticity is what happens when our taste shows up honestly in the work, instead of getting edited out, a decision that feels true and could be defended even when it isn’t the safest option.
Graphic design used to run entirely on this. A poster wasn’t just information arranged well, it was one person’s specific point of view, a provocation put out bravely carrying an emotional risk of being judged. A logo wasn’t just a mark, it was an act of vulnerability, a designer’s private sense of humour or belief, offered up for a stranger to judge. And it was always a negotiation, between what the designer believed was true and what the audience was willing to feel. Graphic design worked because it connected two people, a maker with a point of view and a viewer who appreciated it. That’s the part disappearing now, not the skill, the authenticity and vulnerability. And it’s disappearing into something with a name: slop.
What actually makes a designer’s work theirs isn’t a skill they learned. It’s everything else: the life they’ve lived, what they’ve been through, what they believe, what makes them angry, what makes them laugh, the ideas they can’t let go of, the cause they give up their convenience for, how they see the world, how they talk, even sense of fashion. All of it keeps shifting, week to week, year to year. That’s not a file an AI can copy. It’s an evolving signature, and humans are far messier, far less predictable, and far harder to replace.
The internet is increasingly becoming a place for agents to hangout, as much as humans
How many things stopped your scroll yesterday? Made you pause. Made you engage and reflect? The internet is increasingly becoming a place for agents to hangout, as much as humans. What once was a place for knowledge, stimulation, provocation, learning, questions and answers today is becoming fatiguing with content generated by AI for AI.
A 2026 paper in the Journal of Aesthetics and Art Criticism formally names it “aesthetic bias,” AI models systematically favour beauty, symmetry, and saturation, while marginalizing everything that doesn’t fit that mould. The Copenhagen Institute estimates that by 2030, 99% of internet content could be AI-generated. If that’s true, taste itself narrows to whatever the model already believes is safe.
There’s a classy font. You know the one. There’s a gorgeous gradient, blue sliding into purple. You’ve seen this exact hero section on a hundred products that have nothing to do with each other. Nobody chose any of it. It’s what AI chooses, and by mid-2026, design writers had a name for the pattern: model collapse. Generative models pick what’s likely (what’s safer) over what’s distinctive (what matters). Eventually as more designers are primed to be content with “safe, approvable” outputs, the more AI learns to give more of it than anything new.
And it’s not just about aesthetics. In December 2025, sellers used AI-generated product images to make cheap goods look far better than they were. By January 2026, buyers were returning what they’d started calling “AI slop gifts,” once they felt duped by the gap between the image and the object. Recognition is fast now too: the moment a viewer’s brain tags something as AI-made, in under a second, they assign it less effort and less credibility, and scroll on. There’s a trust tax now, and slop is what triggers it.
Weird (And Vulnerably Imperfect) Used To Be The Whole Point
Adobe’s 2026 design trends report found 73% of designers now deliberately add “imperfect” elements on purpose, a wonky serif, an off-centre curve, specifically to prove a human made it.
Landor’s global creative director named the counter-trend “Anti-AI Crafting” in December 2025, hand-built sets, stitched texture, visible imperfection, now commanding a 10 to 50x premium over anything AI-polished.
Allianz’s brand value grew 20%, to $28.2 billion, in 2025, credited directly to what researchers are calling a “flight to trust”. People trust a brand that feels intentional, not the one that looks flawless.
Why Efficiency Isn’t Always Better Than Authenticity
By April 2025, 74.2% of new webpages already contained AI-generated text. AI’s share of Google’s top-20 search results nearly doubled in fourteen months. The loop doesn’t just repeat. It feeds on itself.
Each pass through the loop narrows the range of what gets made, and what gets made next.
A 2026 meta-analysis pooling 19 studies and 61 separate experiments found a statistically significant homogenization effect from AI-assisted creative work. Whether people are creating art, writing stories, brainstorming ideas, or designing anything at all: individual output seems to look better, faster. Collective output seems to look bland and safe, lacks novelty. Researchers publishing in Nature in 2024 documented “model collapse,” when AI-generated content is fed back into training the next generation of models, the models progressively lose their grip on rare, unusual, distinctive choices and drift toward “bland central tendencies.”
This is no longer a design problem. Every idea that shapes our future, a product nobody’s built, a business nobody’s tried, a way of living nobody’s experienced yet, starts somewhere as something unreasonable, weird, statistically unlikely. If the tools training the next generation of tools only reward what’s already common, the unreasonable idea gets filtered out politely. In fact, it never gets generated in the first place. In a way, our future becomes a smaller bubble, decided quietly, by whichever model happened to be popular the year the training data was collected.
Authenticity, ingenuity, and inventive thinking aren’t nice-to-haves sitting next to efficiency. They’re the only mechanism left for injecting a genuinely new idea into a system that otherwise just recirculates its own safe and average.
Being Weird Is A Practice, Not A Personality
3 small, simple ways to practice greater authenticity as a designer:
- Create your “signature” with everything that defines you. The values, beliefs, quirks, humour, anger, experiences, insights, perspectives, preferences, fashion sense, social style and everything else that makes you you. Train AI on your signature, and to start with first principles. And let your signature evolve with you, as you live life.
- Look for references and inspiration from outside your category. A textile pattern, a culturally rooted word, a decades-old print ad, a joke or an analogy from a completely different industry. Think in first principles, drill down multiple Whys to get to bare truths. Strip everything down and then design from there.
- Use reverse brainstorming to think of what could be the worst way to screw it all up? And then think of the most audacious, unreasonable idea that you’ll bet everything on, even if you’re the only person to believe in it. Think of the status-quo you’re challenging, the argument you’re constructing.
Weird was never the mistake. It was the fee for making something that mattered. Slop is what happens when nobody’s willing to pay it anymore.
At Codewave, we’re building a collective of designer signatures, so businesses don’t have to suffer from more AI slop.
Codewave is a design thinking led digital transformation company enabling organisations with playful innovation using AI & ML, IoT & Edge, AR, VR, Cloud, Blockchain, and Data.
