Brave new world. How we use AI in the studio. And how we don't.

Yeah, another article about AI. Everyone is writing (or singing) about it. Everyone has an opinion. Everyone sits somewhere between euphoria, wonder and the quiet unease you cannot quite shake in the evening when you scroll the feed. Why work at all, if everything can be “one-shotted” in perfection? Everyone is on the hype train, absorbed in finding out how we want to deal with this technology, professionally, personally, morally.

We, as of today, have no answer either as to how all this turns out. But we can say how we at fjnland currently deal with it. What helps us, what we refuse, and why the more interesting question for us is not what AI can do, but what we do not want to leave to it.

We use AI. Seriously, and without a bad conscience.

We are not nostalgic refuseniks. Curiosity is in our DNA. We have used AI intensively since Sam Altman tweeted quite unpretentiously in 2022 “We’re launching ChatGPT today. Try talking with it here: …”. After 4 years now it is fair to say: in certain areas it has changed our work fundamentally. Not only faster, but actually better.

Research is one of them. What used to mean half days sunk in dossiers, reports and sector sources is today, in minutes, at a level that is good enough as a starting point. No substitute for real understanding, but a damn good springboard.

With enough context, AI as an intellectual sparring partner is surprisingly honest. It does not flatter. It contradicts you when you formulate nonsense. When a brief is not yet sitting, or a thesis is still wobbly, AI steers thought in directions you might have avoided on your own.

In coding, with Cursor or Codex, our working has changed at the root. What used to mean long documentation and alignment with external developers now happens in conversation with AI-generated plans, agentic loops and the codebase “developed” that way. This area probably holds the most magic right now.

And then there is our own AI OS. Our working memory. Holding context, documenting developments, writing briefs, thinking along with the editorial plan, structuring emails. AI not as a machine that delivers, but as infrastructure that creates clarity, so we can do the work that counts. What joins all these fields of use: AI is our tool. The directive stays with us.

What we do not do. And why that is not a romantic stance.

Generative AI for creative work in the actual design sense we refuse. “One-shot designs”, AI slop, image worlds generated in seconds: none of that fits us. Not because we are afraid of the competition or of the general development. But because that is simply not what clients engage us for.

Whether Form follows Narration or Form follows Function, design means making the specific visible. That which distinguishes a brand from all others. AI produces the statistically probable. What comes out when you think all previous inputs together. That is by definition the opposite of unmistakable. The opposite of creativity. The opposite of the thought of the future that lives in our work. It is the average. Sometimes useful, for brands in the great majority of cases a death sentence.

That is why we do not automate the decisions that carry our ideas. Not the moment in which a direction becomes a conviction. Not the last sentence. Not the choice of the image that does not only fit, but hits. And not the conversation with clients, the understanding and questioning of context that is written nowhere and changes everything. For us that is not a romantic stance. It is a strategic one.

The actual problem is a different one.

If social media has turned us into goldfish, what does AI make us? We are living through an acceleration that reinforces itself. Every month a new “groundbreaking” model. Every day a hundred new tools that promise the same thing. Content in masses, quality on average falling. That is not an observation from outside. We experience it ourselves, in our own feed, in our own impulses. The temptation to produce more because it is now so easy is real. And it is dangerous.

The problem here is not primarily aesthetic, although AI slop still looks truly terrible. When everything gets easier, everything also gets more arbitrary. When decisions are delegated, you lose the ability to make them yourself. When drafts are generated instead of designed, the eye atrophies. Depth needs friction. That goes for craft, for human thought and, last but not least, for creativity.

And then there is the moral dimension we all like to look past, because it is uncomfortable. Which work is disappearing right now? Whose craft is being devalued? Who really profits from this acceleration, and who pays the price? We have no real answers to that. But we think it is wrong to pretend these questions do not exist. We sit, like many in our field, somewhere between euphoria, scepticism and genuine moral unease.

A look into the crystal ball.

The amount of generated content will keep rising. Average quality will keep falling. And at some point, perhaps earlier than we think, what was made by humans will count again as a feature. As a quality signal. As a reason to pay more.

Market logic, not utopia. When everything sounds the same, looks the same, was generated the same, then the self-willed becomes more valuable again. The studio, the designer, who can say they made it themselves. Thought themselves, designed themselves, decided themselves. Not generated.

We want to stand on the right side of that. Not because we are against technology. But because we understand what we stand for: brand strategy as a process of weighing that needs human judgement. Creativity that is not exhausted in prompts. A stance you cannot delegate.

How we live with it.

AI is a tool. An extraordinarily powerful, sometimes frightening, sometimes genuinely thrilling tool. It changes our work every day and it will keep changing it. We do not refuse that. But we have decided what we want to protect. The creative work at the core. The conversation with clients. The human decision that in the end makes the difference. The process in which quality stands not only as a result, but as a shared path.

If you are asking yourself where your own line sits: do not start with tools. Start with what your work is at the core. What about it, in your view, has to stay human so that it is what someone needs you for? The rest a machine can (and will) perhaps take on. But the rest you probably did not want to do that much anyway.