By the way... AI images, trust and the question of reality
In our column «By the Way…,» Laura Colledani and Klaus Ammon from Management Tools take turns reflecting on contemporary issues. This time: AI-generated images often seem harmless—yet they can still raise fundamental questions. The recent case involving SBB illustrates this: A clearly labeled, efficiently produced AI winter image was enough to spark a debate about trust, reality, and brand perception.
The decisive factor is how artificial people affect real emotions. And what does the Swiss population think? We have tested this for you.
A perfect moment - that wasn't one
A young man stands in the snow, the light is right, the GA is present. Everything looks real at first glance - until you take a closer look: The man is AI-generated and a small note also reads: "AI generated". The person does not exist. The scene does not exist. It is constructed (see here).
SBB talks about efficiency, cost savings and a balanced mix between AI and traditional photography (source: nau.ch) . Transparency is guaranteed. And yes - formally this is correct. But this is where it gets interesting.
The question behind the question
Photographer and AI expert Markus Mallaun formulates the core of the problem on LinkedIn something like this: Caution is required with people and emotions. Not for nostalgic reasons - but because real people in real situations convey a reality that AI cannot offer.
Brand experiences generally live from this reality - you could also say verisimilitude. Although real photos are usually also staged, they are taken from situations that have actually taken place. There is a moment in which a person actually stood in front of the camera.
But theory and lived perception are rarely congruent. Whether AI-generated or "real" - this does not seem to be equally relevant for everyone. Our data shows something else: of the consumers we surveyed on our deeptrue.com platform, only very few (around 6%) took offense at the AI subject, even though it was hotly debated on social media. The vast majority also stated that the motif fits in with SBB's image:
What brands could lose
When emotional moments are artificially constructed, a silent mistrust can arise. Not because AI is bad - but because AI cannot achieve something crucial: The certainty that what is experienced is real. Trust in brands is not only based on content, but also on the feeling that we are not being fooled. Labels such as "AI generated" are therefore necessary in terms of clarity and transparency.
But what does AI generated mean for the relationship level? The result for this SBB subject is positive for the vast majority of respondents:
The real question is therefore not: Can a brand use AI images? But rather: What is the benefit and what are the risks for a brand if the people shown never existed?
AI motives can be useful. But they change the relationship between expectation, emotion and trust. And this is precisely where it becomes apparent whether "synthetic" people are perceived as credible - or as a disruption to the brand experience. A professional debate is important, but the audience's perception is even more important. That's why we tested the subject with consumers on deeptrue.com by management tools.
Even if consumers are relaxed, there is still a limit: Experience cannot be synthesized.
By the way
... real authenticity is not created in the computer - but where a moment is really experienced.
Laura Colledani and Klaus Ammon have more than 40 years of combined experience in consumer research and marketing research. Over the years, they have developed a keen sense of what moves people - even beyond the numbers. Both authors advocate a communication culture that listens before it broadcasts. With the new approach of consumer listening management tools, they want to support marketing managers in not only questioning consumers in the traditional way, but also entering into a genuine relationship with them.
