Every major design technology in history has left its fingerprint on the culture that adopted it. Gutenberg's press made blackletter typefaces the default, not because they were ideal for reading, but because movable type could reliably reproduce them. Roman letterforms took decades to emerge as an alternative. Photoshop arrived in the 1990s and suddenly every website had lens flares, drop shadows, and glossy buttons. Those weren't collective artistic choices. They were the most impressive things the software made easy. Web 2.0 was Photoshop's default capabilities dressed up as a cultural moment.

Instagram launched with a set of filters, and within two years an entire generation of photography looked like it was shot through the same amber window. Valencia. Nashville. Lo-Fi. Dropdown menu options that shipped with the app, mistaken for an aesthetic movement. Canva showed up in the mid-2010s and made template-driven design available to anyone with a laptop. The catch was instant: everything looked like a slightly rearranged version of someone else's layout. The Canva aesthetic became recognizable. Quietly, nobody wanted to be associated with it.

Before AI entered the conversation, the design world was already deep in a sameness crisis. Fons Mans wrote a piece called "We Made Design Efficient and Killed the Magic," and the argument landed. Apple's design leadership set a standard. Dribbble gave designers a stage to perform polish for other designers. Startups realized design mattered, but the way they adopted it, through shared platforms and trending styles, produced a generation of visually interchangeable products. Scroll through Dribbble's popular page from 2018. Soft gradients. Floating UI cards. Friendly illustrations with identical proportions, identical rounded corners, identical careful emptiness.

AI inherited that convergence. Then it put the whole thing on fast forward.

When a brand tells an AI tool it wants something "clean, modern, and tech-forward," the model doesn't examine that brand's history, competitive landscape, or the culture it's trying to reach. It looks at the statistical average of everything in its training data labeled with those words. The result is predictable because averaging is literally what the technology does. Same tools, similar prompts, converging output. The Design Research Society has a phrase for this kind of work: visually competent but conceptually hollow. It looks like it means something. It doesn't mean anything specific.

Marvel's Secret Invasion on Disney Plus offered a test case. Method Studios used AI to generate the opening title sequence, and the rationale was clever on paper: the show is about shapeshifters, so the morphing, unstable quality of AI imagery fit the theme. The backlash was immediate. Artists saw a signal that studios would replace human labor with algorithms. Audiences sensed something was off. Not ugly. Generic in a way that was hard to articulate. Competent work, missing meaning. The audience could tell.

The brands winning against sameness are the ones no algorithm would produce. Liquid Death sells canned water with branding ripped from a thrash metal album cover: hand-drawn skulls, aggressive typography, a tone that swings between absurd and confrontational. No default model lands there. Duolingo turned a language learning app into one of the most unhinged brand presences on the internet, leaning into meme culture, chaos, and a green owl with an unsettling amount of personality. These brands don't resist AI because they're anti-technology. They resist the average.

AI tools are improving fast. Companies can now fine-tune models on proprietary brand assets, producing outputs that are both machine-generated and brand-consistent. But fine-tuning is expensive. It requires a deep library of existing visual work to train on. Nike can do it. The startup that launched last month cannot. The gap between brands with resources to make AI distinctive and brands stuck with the defaults is going to widen into a chasm. Bespoke sameness for those who can afford to train their own models. Generic sameness for everyone else.

Forbes recently reframed brand identity not as a differentiator but as a survival strategy, and that framing is right. When AI can generate a professional-looking brand package in minutes, "professional" collapses as a category. Looking good is table stakes. Table stakes are free. The only thing left to compete on is specificity: cultural, emotional, the kind of meaning that only comes from having a real point of view that a statistical median could never produce.

The question for the next five years isn't whether AI will get better at design. It will. The question is whether visual culture splits permanently into two tiers, one where brands can afford to mean something and one where they can't afford to mean anything at all. That split is already underway. The brands that survive it won't be the ones with the best tools. They'll be the ones with something so specific to say that no tool could say it for them.

If you're thinking about what sameness is costing your brand, subscribe to The Threads of Culture for the next piece.

Watch the video essay, Why Every Brand Looks the Same Now, on YouTube.