The spirits we called...

m&k: What happens to creativity when the algorithm decides what "performs well" and we misinterpret this as creative excellence?

Fabian Bertschinger:

Imagine if a play was no longer written by the author, but by an algorithm that decides which scenes stay and which are removed based on applause measurements. This is what it feels like when algorithms decide what performs well. Creativity becomes an echo of past successes, as algorithms predominantly favor content that is already popular. What stands out is smoothed out to match the familiar patterns that the algorithm knows from the past. The result would be that what brings the most clicks, likes or conversions is suddenly sold as creative. Yet the ground-breaking ideas often come from beyond the predictable.

A venture into the unknown...

Yes, a truly creative campaign always involves a moment of uncertainty. But if we only play it safe, we risk nipping anything new in the bud. We end up with the same images and messages over and over again because the algorithm convinces us that this is optimal. But it's just yesterday's statistical mediocrity. At a time when everyone is suffering from content fatigue, we are missing the opportunity to inspire people through creativity. If everything looks the same, everything becomes interchangeable and nobody is really surprised or moved. This makes it increasingly difficult to stand out from the crowd and create a genuine connection. For me, creativity means swimming against the tide and daring to do something that doesn't yet have a track record. But the algorithm has no gut feeling and no vision, it only knows data. So if we blindly rely on its definition of "excellence", we cut off the very thing that brings creativity to life: the surprising, the unpolished, the leap into the unknown.

"The Sorcerer's Apprentice is a cautionary parable. It shows that you should master a power before you use it."

Conversely, is the greatest risk of generative AI in marketing mediocrity and not manipulation?

Manipulation, deepfakes and subtle influence are extremely dangerous, no question about it. They distort reality and undermine trust. But they attract attention, provoke, outrage and get us moving, precisely because boundaries are being crossed. When we recognize them, we discuss them and take a stand. Mediocrity, on the other hand, remains silent. It doesn't excite anyone, doesn't arouse any emotion and offers no resistance. This is precisely where its treachery lies. Mediocrity slips under the radar until it becomes the norm. It seeps unnoticed into the language of brands, images and claims. Everything sounds correct, but nothing special anymore. Mark Zuckerberg describes this mediocrity not as a danger, but as a goal. In a recent interview with Ben Thompson, he outlines how Meta wants to advertise in the future: companies should simply state their destination and link their bank details. The AI will do the rest. No creation, no targeting, no strategists, no questions. Content without creativity. That sounds like efficiency, but it is a rejection of the creative principle. Because what results is neither surprising nor original. Generative AI optimizes for what is statistically expected, not for what is unique. It delivers what works, not what fascinates. The result is an aesthetic of flawless harmlessness, but also of insignificance. Yet good marketing thrives on profile and edges. But if every headline is generated by an AI according to a formula, we run the risk of sanding off those corners and edges that make a brand unmistakable. Zuckerberg's vision is radical. The risk that every creative act entails is to be eliminated. What remains is a self-optimizing cycle of input, output and feedback. The greatest danger is then not the dreaded scandal, but the big yawn.

If AI only optimizes attention in communication, who actually pays attention to meaning anymore?

Humans, of course, who else? Meaning doesn't come from algorithms, but from experience, intuition and context. AI does great groundwork for this by being able to recognize patterns from huge amounts of data, which we would never be able to do. It provides an excellent basis for differentiating insights. But tools don't build houses, they have to be managed. Developing communication is not just about seeing or recognizing something, but also about understanding the why behind it. Meaning is created in the space between sender and receiver and no system can ever have a complete overview of this space. Fortunately, one might say. Because this is precisely where our task lies: as curators, translators and creators of meaning. With the sovereignty over meaning also comes the responsibility for what we put into the world. Ultimately, it is the end consumer who decides whether the content has meaning for them. And if everything is the same, nothing has meaning anymore.

"What AI lacks is friction and that's exactly what creates connection."

Can you think of an example where AI communication is completely ethically correct, but still culturally toxic?

A recent example of ethically correct but culturally toxic AI communication was provided by Channel 4 in the UK. As part of the "Dispatches" series, undecided voters were confronted with political messages via a specially developed platform in the experiment "Can AI Steal Your Vote?". Partly real, partly AI-generated, partly deepfakes. Everything was transparent, voluntary, clearly labeled, i.e. ethically impeccable. And yet the result is worrying. Because 23 out of 25 participants changed their political stance. Influenced by content that convinced them emotionally, even though it was not real. This is precisely what makes the case culturally toxic, because AI is not used here to mislead, but to be emotionally effective. The communication doesn't break any rules, but it changes the rules of the game. When deepfakes seem more convincing than real statements, communication becomes a simulation. It is no longer about truth, but about effect. The line between the real and the fake becomes blurred. AI communication seems to numb our critical thinking. What remains is a perfectly staged reality that we want to believe. Not because it is true, but because it is so well made.

What values does a brand lose if it relies entirely on AI-based targeting?

A brand that relies entirely on AI quickly learns what the target group wants. And at some point stops asking itself what it actually wants to say. Because AI is good at adapting, but not at differentiating. It recognizes trends, but no direction. It recognizes needs, but has no attitude. And so it optimizes communication for maximum resonance until everything feels relevant, but nothing has meaning or contributes to an overarching goal. Relevance does not arise where you simply give what is asked for. It arises where there is an overlap between what the target group is looking for and what the brand offers out of conviction. This is not a data cut, it is a decision and an attitude. AI can help to recognize patterns. But only humans can recognize when it's time to break them. Successful marketing begins when a brand not only reacts, but also takes risks. If I were an AI right now, I would probably use the statistically most likely quote from Steve Jobs at this point: "People don't know what they want until you show them".

"AI does great groundwork by being able to recognize patterns from huge amounts of data."

Could it be that the greatest creative provocation in 2025 will be not to use AI - as a matter of principle?

Absolutely. At a time when generative AI has become the norm, deliberate renunciation is the real provocation. Not as hostility towards technology, but as an attitude. Dove demonstrated this in April 2024 with its "The Code" campaign, in which the brand explicitly refrained from using AI-generated images of women in its communication. Instead of algorithmically generated flawlessness, Dove shows real faces, real bodies, real stories. It is remarkable that this step did not come out of nowhere. It is the logical continuation of a consistently managed brand. For over 20 years, Dove has stood for "Real Beauty", for an understanding of beauty that is not based on standards, but on people. In a world that increasingly relies on synthetic emotionality, Dove is defending its brand essence: authenticity as a design principle. What makes the campaign really strong is not just the renunciation itself, but the public appeal to do the same. Dove is calling on the industry to also commit to not using AI-generated images of women. This is no small gesture, but a strong signal that responsibility is not just a matter for the market, but a cultural issue. In the midst of a flood of generated content, the renunciation of AI is not a step backwards, but a conscious resistance against the uniformity. Dove shows that the courage to embrace imperfection does not mean renunciation, but cultural leadership.

Which myth or fairy tale do you think would make a good parable for today's AI communication in marketing - and why?

Funny question, perhaps Goethe's "Sorcerer's Apprentice"? An inexperienced apprentice orders a broom to fetch water using a magic spell. At first he is delighted with the new efficiency - at last he no longer has to carry the buckets himself. But soon everything is under water because he doesn't know how to stop the broom. "I can't get rid of the ghosts I called." What does this have to do with AI in marketing? Not that we want to get rid of AI, but for me the moral of the story pretty much reflects our situation in communication. With algorithms that automatically produce, distribute and optimize content, we have conjured up a powerful force. We celebrate content at the touch of a button, marvel at efficiency, scalability and text quality. At the same time, however, we feel our control slipping away. The communication landscape is filling up with generic messages and automated interactions. Soon we are knee-deep in content water and looking for the magic word to stem the tide. The sorcerer's apprentice is a cautionary parable. It shows that you should master a power before you use it. In the same way, we should learn to use AI consciously and at the right moment. Otherwise we will drown the brand along with its audience. But the challenge does not end with technology. While Europe is attempting to set guidelines for the use of AI with the AI Act, it is clear how inconsistently AI is handled around the world. Many brands operate globally, but their communication rules are negotiated locally.

In your opinion, what aspect of the human being is too contradictory, too chaotic, too flawed to ever be meaningfully captured by AI in marketing?

People are contradictory, not by accident, but by principle. They want to fit in and still stand out, preserve and change at the same time, plan and remain spontaneous. What was true yesterday seems embarrassing today. What comforts us can bore us at the same time. These inner tensions are not a flaw, but the essence of the human condition. This is precisely where artificial intelligence struggles. It thinks in patterns, in probabilities. What it cannot grasp is the improbable, namely the moment when someone acts against their profile, makes a gut decision or even contradicts themselves with conviction. AI tries to divide our nature into segments. But what to do with a customer who buys organic vegetables and loves trash TV? AI recognizes the combination, but not the reason. It understands the what, but not the why. AI also lacks the ability to grasp the subtle nuances of human emotions. Emotional intelligence remains the domain of humans, as AI can recognize patterns but cannot empathize with real feelings. Effective marketing, however, thrives on a feeling for what cannot be calculated. From those illogical moments when a person overtakes themselves. Brands that recognize this do not meet us as a mirror image, but as a counterpart. What AI lacks is friction and this is precisely what creates connection. These unpredictable human impulses are also what make great campaigns. Moments in which the audience is surprised and genuinely touched.

"AI is good at adapting, but not at setting boundaries. It recognizes trends, but not direction."

If you had to combine AI with a philosophy, for example Zen, Dadaism or Situationism, which would you choose and what would the result be?

Dada. Not because it fits, but because it stands in the way. Where AI seeks order, Dada brings disorder. Where algorithms love patterns, Dada focuses on disruption. The combination seems absurd, but that is precisely what would make it fruitful. Because the hallucinations, semantic aberrations and formal lapses that we dismiss today as AI errors would become artistic potential in the light of Dadaism. At a time when generative AI is increasingly being used to produce disinformation, a Dada AI offers an alternative approach. By deliberately dispersing meaning and subverting the expectation of coherent statements, this AI evades the game of truth claims and deception. If you cannot claim truth, you cannot lie. The result would be a machine that does not work, but inspires. One that does not try to provide the right answers, but rather triggers thought processes. In an age where predictability is becoming the ideal, Dada AI would be one that does not make itself useful, but one that enriches life.

The power of AI.images: Midjourney

Fabian Bertschinger is Executive Creative Director at Farner, Switzerland's leading communications agency. He has more than 20 years of experience in branding and communication and is a member of the Art Directors Club Switzerland (ADC). He played a key role in Switzerland's first AI-generated advertising campaign and is currently driving the integration of generative AI in the creative industry with various projects. In addition to his agency work, he teaches at several universities and is currently continuing his own education in the field of AI management. He is a regular speaker on topics at the interface of AI, design and brand strategy.