At first, governance is almost always perceived as a hindrance—an extra review step, another approval, one more rule in a process that used to be faster. That usually doesn’t change until it becomes clear what the rule is actually protecting against. At coeo, that’s exactly what happened with regard to AI-generated images.
As of August 2, 2026, new transparency requirements under the EU AI Act will apply to certain AI-generated and manipulated content. For many companies, this is the moment when they are seriously addressing the issue for the very first time. At coeo, the deadline is more of a confirmation than a starting point: The company already had its own implementation guidelines for AI transparency in place beforehand, developed in response to a question that arose independently of the legal situation.
The reason for this lies in the way modern image generators work today. Artificial intelligence can generate images that look deceptively real. Isabell Neubert, Manager of AI/IT Governance & Regulatory Tech at the coeo Group, sums up what this means for a company like coeo:
“For a company that operates on the basis of trust—as coeo does in its dealings with customers—this is not just a neutral tool, but a matter of attitude.”
Every face, every situation, on demand
What makes this technology special becomes particularly clear in the context of debt collection. When a company needs images that convey financial hardship, AI opens a door that wasn’t available before: It allows you to generate any face or situation you want, in real time. Through profiling, it would theoretically even be possible to capture the characteristics of the viewers themselves in order to simulate their situation.
AI image generators can produce, on command, exhausted faces, overflowing mailboxes, stooped postures, or even people of a specific ethnicity or social group. This is precisely where the real risk lies: not in the technical capability itself, but in how easy it is to use without anyone consciously making a decision about it. An image can be generated in seconds, but the question of what it actually depicts and who it affects takes time—time that is often in short supply in day-to-day business. The responsibility that arises from this has little to do with the quality of the image and much to do with the impact it has on viewers.
Guardrails Instead of a Mere Marking Requirement
coeo has therefore defined its own content guidelines that go beyond mere labeling requirements. The core idea behind this is that financial hardship is a social reality, not a stage for drama, and certainly not a source of stereotypes. Visual imagery should reflect reality, not exaggerate it, and should not turn anyone into a symbol of failure.
Internally, there is not only a policy with specific guidelines on images, but also guiding questions that encourage employees to put themselves in the viewers’ shoes, rather than simply asking whether an image works or looks good. This shift in perspective—away from technical considerations and toward human ones—is the very essence of the guideline. It requires employees to consciously ask themselves, with every image, how it would feel if they themselves were depicted in that situation.
To ensure that this isn’t just a one-time decision, the review is embedded in the process as a recurring step, not as a one-time approval at the beginning. New image formats, new campaigns, and new channels are evaluated against the same guiding questions, regardless of how urgent a project may be at the moment.
Labeling is mandatory; attitude is optional
The key point of this approach is that the use of synthetic images does not relieve anyone of their responsibility for content and ethics. AI generates the image, but the decision about what it conveys and about whom remains a human one.
When it comes to synthetic images that involve financial implications, the issue is therefore not just whether an image is recognizable as AI-generated, but whether it should be created in the first place. Labeling is mandatory; taking a stance is optional, and coeo treats both as inseparable.
For a company whose business is built on trust, this is more than just a compliance exercise. It’s the difference between a rule that’s followed simply because it must be followed, and a mindset that holds true even when no one is watching.
Cover Image © Funster
