Almost everyone uses or consumes AI content, whether in everyday life, in the company, or at the agency. With the rapid arrival of the technology, however, the challenges are growing too, and the EU AI Act now provides a first, if somewhat superficial, answer to them.
From 2 August 2026, Article 50 of the EU AI Act introduces a staggered labelling obligation for AI-generated content. This is not a fringe topic for tech corporations, but affects nearly every company that uses AI visibly in its external communications. High time, then, to set a few things straight and to explain why this can be understood as an opportunity.
What Happens from 2 August 2026?
The AI Act already entered into force on 1 August 2024, but is being activated in stages. The transparency obligations under Article 50 become directly applicable across the entire EU from 2 August 2026, without requiring any national implementation (Official Journal of the European Union).
Article 50 is not a single obligation, but bundles four different requirements. For our world, meaning marketing, communication, and brand management, three of them are particularly relevant:
- Chatbots must identify themselves. Anyone interacting with an AI system on a website must be made aware of this, unless it is obvious anyway (anwalt.de).
- Deepfakes must be disclosed. Realistic, artificially created images, sounds, or videos that simulate real situations must be clearly and distinguishably labelled (TÜV Rheinland).
- KI-Texts on matters of public interest are subject to labelling if no editorial control has taken place.
And then there is the technical, machine-readable marking of synthetic content under paragraph 2. For systems that were already on the market before the cut-off date, this deadline is, according to the "Digital Omnibus" draft, expected to be postponed to 2 December 2026 (artificialintelligenceact.eu).
Important to know for anyone now nervously eyeing their latest ChatGPT-assisted LinkedIn posts: the labelling obligation for texts does not apply if a human has reviewed the content substantively, meaning for accuracy, plausibility, and sources, and a clearly named person bears editorial responsibility. A mere "read-through" or a spelling check does not suffice for this, however (LAUSEN Rechtsanwälte). Classic marketing copywriting and a company's own communications generally do not fall under the "public interest" criterion anyway. Anyone now breathing a sigh of relief and believing that social media is a rule-free zone, however, is gravely mistaken. As soon as companies post AI-generated advertising videos, photorealistic graphics, or artificial voices to promote their own business, the labelling obligation applies in full force. The line is drawn sharply along commerciality here: pure private individuals are indeed exempt, but as soon as companies or commercially operating influencers build reach with AI content, synthetic media must be marked as such. An AI video on Instagram or LinkedIn without a label thus becomes a direct risk of cease-and-desist letters and fines.
To make clear what is at stake: the fines for violations reach up to 15 million euros or 3 percent of global annual turnover, whichever amount is higher (TÜV Rheinland).
And … that’s a good thing?!
Now one could dismiss the whole matter as a compliance burden that creates work for us and unsettles clients. We see it differently, and for a reason that lies deep within our self-image.
A brand lives on trust. It lives on the fact that what a company promises externally aligns with what it does. This is precisely where our brand promise comes in: we develop ideas that capture your identity and give it a stage with genuine, authentic action. Authenticity and concealment go together poorly.
Anyone who communicates openly that an AI avatar is involved or that an image was artificially created is more likely to be perceived as innovative and transparent. The greater risk is not disclosure, but being caught and then standing exposed as dishonest. Transparency is quite simply the better brand strategy.
For you as our clients, especially in the technology- and industry-adjacent B2B environment, this means something very concrete. You sell to people who are supposed to trust you over years; to buyers, to technical decision-makers, to partners with long investment cycles. In this environment, trust is not a soft factor, but the very basis of business. Honest, clearly labelled AI use therefore not only protects against fines, but pays directly into what holds your brand together at its core.
How we at MINT handle this
So that you understand we don't just talk about it: we have long since aligned our own processes accordingly. AI is a tool for us, not a secret ghostwriter. Every text, every concept, every design undergoes genuine human review, in which a person from our team bears responsibility. This is not only legally sound, but corresponds to our quality standard: a good idea does not arise at the push of a button; it needs the right context and a trained eye.
At the same time, we are actively building the topic into our consulting. When we speak with you about a website or a campaign, the question of AI labelling will henceforth be a self-evident part of it, from the choice of labelling position to clean documentation in the CMS.
The labelling obligation has come to stay. We can see it as a burden or as what it actually is: an invitation to be honest about the way we work today.
If you want to know where your company stands and how to implement AI labelling in a way that strengthens your brand rather than weakening it, then let's talk. That is exactly what we are here for.
