Contact us
Want to talk about how AI fits into your brand's creative work? Get in touch, and let's figure out where judgment matters most for you.
The interesting story in AI right now isn't the technology. It's the gap that's opened between two kinds of brand: those using AI to do the same work cheaper and faster, and those using AI to do better work than was previously possible. The first group is producing a lot. The second group is winning.
Getting to that point has taken a couple of years of collective learning, some of it painful. Here's where we are, what we've picked up, and where things are heading.
Matt HareSenior Designer
The death of the bottleneck
For most of modern creative work, making was the constraint. Skill was scarce, software was hard, and teams were expensive. If something looked polished, someone had probably spent serious time on it. That relationship has broken. AI tools now produce high-fidelity output in minutes. Work that once needed a week, a team and a developed craft vocabulary can be assembled by someone who has never opened Figma.
AI doesn't just make bad work. It makes bad work look finished, and finish used to be the signal that someone cared.
Superside's May 2026 survey of 211 enterprise creative and marketing directors found 97% believe the creative role will look fundamentally different in five years, and 97% say human creative output still matters. Sprout Social found 56% of social media users now see low-effort AI content "often or very often". Gartner found 50% of US consumers would prefer brands that don't use generative AI in customer-facing content. The most AI-literate consumers, Klaviyo found, are also the most sceptical of brands using it carelessly.
Supply and demand are pulling in the same direction. What audiences react to isn't AI itself. It's the feeling that nobody was really there.
The anti-AI campaign became a category
The brands responding to that sentiment have built a recognisable movement. Adweek called it directly in its 3 March issue: if 2025 was the year brands began cautiously expressing views on AI, 2026 is the year they're taking clear, public stances.
Blue Diamond Almond Breeze launched a Jonas Brothers campaign explicitly framed as a "playful jab" at AI-generated advertising. Polaroid placed anti-screen messaging near tech-heavy urban spaces. Apple's Vince Gilligan series Pluribus closes its credits with the line "this show was made by humans". Absolut Vodka shot its Tabasco campaign on location in Iceland, with a 40-foot volcano built in a studio, and made the AI-free production part of the story.
Audiences now grade these statements with the rigour of a fact-check. Saying you're human-made isn't enough. You have to show your work, the way Absolut did, where the studio volcano wasn't just production value, it was evidence.
What we've learnt: authenticity has moved from claim to proof
The most visible shift is that "human-made" has stopped being a tagline and started becoming a certification. The BBC reported in March that organisations worldwide are racing to create a universally recognised "AI-free" label — the equivalent of Fair Trade for human creativity. It identified at least eight competing initiatives across the UK, Australia and the US. It's early and noisy, but the direction is clear.
Transparency is quietly becoming the default. In late 2025, Google began adding AI transparency labels to ads and image search when content had been generated or significantly edited by AI. From July 2026, Google is going further: advertisers will be able to add text or visual AI labels directly within image and video creatives, and Google Ads will automatically apply labels to certain assets generated by its own AI tools.
For creative teams, the takeaway is simple: build a workflow where someone accountable reviews the output, and be straightforward when disclosure applies. Transparency about AI use is becoming table stakes, not a differentiator.
Knowing when to use AI matters more than whether to use it
The brands moving ahead aren't using AI the most. They're using it most deliberately. AI has made production cheap and judgement expensive, and judgement is increasingly what clients pay for.
Judgement isn't a feeling. It's attention under pressure, staying with a problem until the easy answer wears thin, earning the call with a clear rationale rather than "trust me". As James Hurst, VP Global Executive Creative Director at HelloFresh, puts it: "AI is really good at being deductive. The thing it can't do is think of a novel way of solving problems... that layer of human intelligence is, and will be, radically important."
Superside's research shows where creative leaders expect the line to land: within three to five years, over 40% expect AI to handle most image generation, editing workflows and creative strategy without human input. AI is absorbing execution. Judgement is moving up the value chain.
We saw this in practice on a recent campaign for L'Oréal's Essie. The Liquid Diamonds launch could easily have become another generic Midsummer scene: a flower-crowned model on a meadow, prompted in seconds. Instead, we built it around close-up intimacy: hands, textures, the small rituals of a long Nordic day. We ran a dedicated photoshoot with content creator Petra Tungården, localised across Meta, TikTok, Snapchat and YouTube for Sweden, Denmark, Finland and Norway.
As Liridona Mulolli, Nordic Product Brand Manager at Essie, put it: "When the collaboration works, the result shows".
That's the choice facing every brand: not whether to use AI, but where to spend the time it frees up. High-volume, low-judgement tasks, AI handles well. Others sit at roughly 70%, where human refinement makes the difference. And some work, campaign creative, brand systems, strategy, has to stay human-first, because that's where the value lives. AI can render the image, but it can't decide what it means.
The brands gaining ground aren't working faster. They're spending the time AI frees up on the judgement that produces work people remember.
What comes next: human-verified and accountable
Writing in Entrepreneur, Scott Baradell argued that "human-verified" and transparent AI governance is now a trust-building differentiator. The point isn't whether AI was used, but whether a human took responsibility for the output, and whether the audience can see that.
HubSpot now openly acknowledges generative AI in its content workflow while keeping human editorial oversight. The acknowledgement isn't a weakness; it's a credibility signal. Hidden automation destroys confidence; disclosed automation, paired with visible judgement, doesn't.
The brands gaining ground use AI deliberately, are transparent about how, and have a system that makes that transparency repeatable. None of that is incompatible with scale. All of it is incompatible with shortcuts.
AI-generated work is producing what GQ called a "very good average": competent, polished, indistinguishable. Which is exactly why taste and accumulated judgement have become the scarce, valuable things. Taste isn't a feature you can ship. It's built through repetition, observation and time.
The bottom line
Everyone has access to the same AI tools. That's what makes them tools, not advantages.
What separates the brands pulling ahead is what those tools are pointed at, who reviews what comes out, and who's willing to put their name to it. As designer Simon Dixon put it recently, "a choice no one will answer for isn't a decision. It's a suggestion".
Production is cheap, but judgement is expensive. Accountability is what turns a machine's output into a brand's work. The brands that understand that and build the workflows, disclosures, and human review to prove it will navigate the next phase of AI. They'll define it.
Gabriella FritzsonCreative director
Gabriella BjörnbergManaging director, Stockholm
Mikko PeltomäkiManaging director, Finland