Algorithmic conformity

AICritical ThinkingConformity
Written By Andrew Tope

In October 2025, WPP launched Open Pro, an AI platform that generates marketing campaigns from start to finish. The pitch was efficiency at scale. The promise: "professional-grade tools" previously reserved for Fortune 500 clients, now available to anyone. WPP had had a pretty tough 2025 with revenue down nearly 5%, share price dropping, new CEO calling results “unacceptable.” Meanwhile Publicis grew 5.5%, powered by AI-driven tools like Epsilon. Omnicom has Omni AI.

The question is what happens when every brand uses the same AI to optimise for the same outcomes? The answer is already visible.

The Harmonisation of Tastes

In 2016, journalist Kyle Chayka noticed something peculiar while travelling. A café in Odessa, Ukraine, looked identical to one he'd visited in Seoul. Raw wood tables. Exposed brick. Hanging Edison bulbs. Neither was part of a chain. Neither shared a corporate parent. Yet they had independently converged on the same aesthetic.

Chayka coined a term for this phenomenon: "AirSpace." Writing for The Verge, he described it as "a harmonisation of tastes" driven by digital platforms. The coffee roaster Four Barrel in San Francisco looks like Toby's Estate in Brooklyn looks like The Coffee Collective in Copenhagen looks like Bear Pond Espresso in Tokyo. You can get a dry cortado with perfect latte art at any of them, then Instagram it on a marble countertop and further spread the aesthetic to your followers.

This wasn't corporate standardisation of the kind McDonald's perfected. It was something stranger: organic convergence mediated by algorithms. Foursquare recommended the same cafés to travellers everywhere. Airbnb hosts optimised their listings for the same photography standards. Pinterest boards circulated the same interior design references across continents.

What began as a curiosity about coffee shops turned out to be a preview of something much larger.

The Age of Average

In March 2023, strategist Alex Murrell published an essay The Age of Average" which documented, with exhaustive visual evidence, how creative fields across every domain were converging on identical styles. Murrell traced the phenomenon back to an art project from the early 1990s, when Russian artists Vitaly Komar and Alexander Melamid hired a market research firm to survey people in 11 countries about what they wanted in a painting. Despite vast cultural differences, every country produced nearly identical preferences: blue landscapes with a few figures and animals in the foreground. "Looking for freedom, we found slavery," Komar later quipped.

Three decades on, Murrell argued, these "people's choice" aesthetics had metastasised across every creative domain. Interiors all look the same—the "International Airbnb Style" of white walls, Eames chairs, and reclaimed wood. Architecture all looks the same—the "five-over-one" apartment blocks spreading across American cities. Cars all look the same—not just their aerodynamic forms, but their colours too. According to data Murrell cited, around 40% of cars sold in 1996 were monochromatic (black, white, silver, or grey). By 2016, that figure had reached 80%.

Even people all look the same. Jia Tolentino, writing in The New Yorker, described "Instagram Face"—a particular combination of features, achievable through makeup and filters, that had become the dominant beauty ideal across the platform. The face looked like nobody and everybody: ethnically ambiguous, youthfully generic, algorithmically optimised for engagement.

Murrell's conclusion was stark: "Distinctiveness has died."

The Formula Problem

Jazz critic, music historian and one time management consultant, Ted Gioia, offers a parallel analysis of what's happened to sound.

In a February 2024 essay titled "The State of the Culture," Gioia argued that music had been captured by the same forces that homogenised coffee shops and interiors. Songs had simplified, structures had shortened, and everything was now optimised for the first 30 seconds—the window before a Spotify listener decides to skip.

"The real challenge in music is the formula," Gioia wrote. "And that's what deadens your musical culture, the repetition of the formula."

The mechanism is economic. Spotify pays per stream, and a stream counts after 30 seconds. So the hook has to hit immediately. The intro is expendable. Any element that doesn't maximise engagement is a liability. The result is music that has been reverse-engineered from the metrics that determine revenue.

Gioia traces this to a broader shift in what he calls the "dopamine culture"—an entertainment ecosystem designed around limbic stimulation rather than meaning. "The culture business is now totally dominated by attention merchants," he writes. The content that wins is the content that triggers an immediate neurological response. Anything that requires patience, complexity, or delayed gratification is at a structural disadvantage.

The implications extend beyond music. In a follow-up essay from March 2025, "The World Was Flat. Now It's Flattened," Gioia argued that the same dynamics were reshaping all forms of creative expression. AI hadn't created this problem; it had inherited and accelerated it.

The Architecture of Averaging

Large language models, by design, excel at interpolation. They produce outputs that sit within the distribution of their training data. When you ask an LLM to generate marketing copy, it draws on patterns from millions of existing examples. The result is competent, plausible, average.

This is not a bug. It's the nature of how these systems work. As AI researcher Andrej Karpathy has noted, LLMs are "compression algorithms for the internet." They distill the patterns of vast corpora into probabilistic models. What they cannot do is value the unprecedented.

Critical thinking questions the patterns. Creative thinking imagines alternatives. Neither can be truly automated, because both require stepping outside received assumptions rather than optimising within them.

Om Malik, the tech blogger and investor, made this point in a January 2026 essay that prompted this reflection. He paired two quotes separated by 135 years—Oscar Wilde in 1891 ("Most people are other people. Their thoughts are someone else's opinions, their lives a mimicry, their passions a quotation") and designer Verner Panton complaining about a world "housed in dreary, grey-beige conformity."

What Wilde and Panton couldn't anticipate was how technology would industrialise this conformity, baking it into the economic infrastructure of the internet itself. Algorithms don't shame you into conformity; they reward sameness with distribution and punish deviation with invisibility.

The Human Remainder

The human contribution that AI cannot replicate is judgment, taste, and the willingness to be wrong.

This is the domain of critical thinking and creative thinking, the capacities that allow us to step outside received patterns, question assumptions, and imagine alternatives. These are precisely the capacities that algorithmic systems, by design, cannot possess.

Spotify research from 2020 found that algorithmic recommendations, while effective at predicting what users would click, actually reduced the diversity of what they consumed. Users enjoyed familiar patterns more, but discovered less. The algorithm gave them what they wanted, and in doing so, narrowed what they could want.

The parallel to WPP's Open Pro is direct. When every brand uses the same AI to optimise for the same metrics, the result is convergence. The tools work. They produce competent outputs efficiently. But competent and efficient is not the same as distinctive, and distinctive is what builds brands.

The Escape Routes

The grey-beige world is not inevitable. But escaping it requires more than individual acts of aesthetic rebellion. It requires building institutions, incentives, and cultures that reward distinctiveness over conformity, meaning over engagement, and human judgment over algorithmic optimisation.

Some approaches are beginning to emerge. Brands that deliberately avoid templates. Publishers that value editorial judgment over A/B testing. Musicians who resist the 30-second optimisation. These are not nostalgic gestures; they're strategic bets that audiences will respond to distinctiveness when they encounter it.

The challenge is structural. Algorithmic distribution creates powerful incentives toward sameness. Breaking out requires either accepting lower distribution (which few businesses can afford) or finding alternative channels that don't apply the same pressures. Neither is easy.

If distinctiveness is what creates value, and algorithmic optimisation destroys distinctiveness, then the tools that promise efficiency may be undermining the very thing that makes creative work worth paying for.

That is a harder conversation than downloading a new app. It is also more necessary than ever.

Andrew Tope
Previous

Is AI killing critical thinking?
Next

Fox with a compass