What the EU AI Act Means for Marketing, Advertising and Creatives
Updated: Aug 22
Here's what marketers, agencies, brands and creators actually need to know in 2026 — and how Fivora is adapting its AI usage standards.
Artificial intelligence has become part of everyday marketing. It writes copy, generates images, edits video, creates voices, analyses audiences, builds campaign ideas and automates workflows.
The evolving AI sector needs moderation and regulations to keep it safe. In response to this, the European Union set out the EU AI Act.
The EU AI Act does not ban businesses from using AI to create marketing content. However, there are rules around how certain AI systems are used, when people need to be told they are interacting with AI, and when AI-generated or manipulated content needs to be disclosed.
Those rules are now particularly relevant because the EU AI Act's Article 50 transparency obligations took effect on 2 August 2026.
For marketing agencies, advertisers, designers, photographers, filmmakers and content creators, understanding the difference between using AI and using AI deceptively is becoming increasingly important.
Here's what the law actually says.
What is the EU AI Act?
The EU AI Act (Regulation 2024/1689, if you enjoy legal jargon) is essentially Europe stepping in to play referee on artificial intelligence. The good news? It doesn’t treat every algorithm like a thing to be scrutinised. Instead, the EU looks at risk.
Some outright shady uses are banned completely. Critical, high-risk systems get slammed with heavy red tape. Mid-tier setups just have to be transparent about what they are. And the everyday, low-risk tools? Barely touched.
If you’re in marketing, you can breathe. Using generative AI to draft copy or mock up social assets won’t land you in regulatory prison. It all boils down to two simple things: what you’re actually making the AI do, and what you’re putting in front of real people.
Does the EU AI Act ban AI-generated content?
Short answer: no. This is easily the biggest panic-fueled myth floating around. The law isn't a secret ban on your favourite AI Model; it isn't making AI-rendered visuals illegal, and it’s definitely not putting marketing agencies out of a job.
You also don't need a giant warning label on every single sentence just because an LLM helped you fix a typo.
Act introduces specific transparency obligations for certain AI systems and certain types of AI-generated or manipulated content. The objective is largely about preventing deception and manipulation and helping people understand when they are interacting with AI or encountering certain synthetic content.
It is to essentially stop people from getting catfished or manipulated by synthetic media. If you are just using it to work smarter, you're fine. That distinction makes all the difference.
So, what does the EU AI Act actually demand from marketers?
We dived deep into the EU AI Act from a creatives' perspective. These were a few areas worth paying attention to.
Don't pretend your automation or AI bot is Dave from Support
If a business deploys an AI system that directly interacts with people, such as an AI chatbot, agent, or digital avatar in front of customers, they need to know it's a machine unless this is already obvious from the circumstances. Trying to gaslight your audience into thinking your automated responder is a live human typing away in a coffee shop will land you in hot water.
Drafting an internal campaign brief with an LLM behind closed doors? Completely fine. The rule only kicks in when the public is the one having the conversation.
The important distinction is who is interacting with the AI and how.
Leave the invisible watermarks alone
If you did not know this, all mainstream AI generative engines (Gemini, ChatGPT, Claude, etc) are obligated to watermark assets produced using their generative engines. They are called provenance tags or content credentials. We've written a blog dedicated to how these work, how to verify them and their use cases.
This is to ensure that AI-generated or manipulated audio, images, video and text are marked in a machine-readable format so that they can be detected as artificially generated or manipulated.
This obligation falls on the companies building the models like OpenAI and Midjourney. Your job as a marketer is simple: use reputable tools and don't go out of your way to strip out provenance tags.
AI-generated images need honest labels
This is where things get spicy. Under Article 50, if your business or brand publishes an AI-manipulated image, audio clip, or video that looks so real it could easily fool someone, you legally have to disclose that the content has been artificially generated or manipulated using a simple "AI-Generated" or "AI-Manipulated" tag. I am pretty sure you'll find a better way to disclose it
We aren't talking about stylised 3D renders here. If you generate a hyper-realistic video of a celebrity endorsing your product when they’ve never even heard of your brand, that's not just "creative marketing".
At Fivora, we call it deceptive synthetic media. This is what the EU AI Act is trying to regulate because this synthetic media could mislead people about what actually happened. The distinction is enormous.
What about AI-generated ad visuals?
This is where the internet says "Drama, Drama, Drama" and loves to lose its collective mind.
If your creative team uses Midjourney to build a dreamy surreal background, concept art, or a sleek fictional product render, that does not automatically mean the resulting advert is illegal under the AI Act.
The EU AI Act recognises that the context does the heavy lifting here. The regulators care about intent, realism, and what your audience reasonably expects to see.
What about AI-generated Copies and Text?
You are free to use AI-Assisted copies or text in a marketing context. However, AI-generated or manipulated text published to inform the public about matters of public interest, political positions, and emergency directives should undergo human review or editorial control.
There is no obligation to mark AI-generated texts or copies.
What does the EU AI Act say about AI literacy?
Under Article 4, if your team uses AI to deliver work, you are legally expected to make sure they actually know what they’re doing with it.
If your business is shipping markeiting campaigns campaigns with AI, they should understand that these tools hallucinate facts, amplify baked-in bias, miss nuance, botch brand claims, leak sensitive data, and happily generate slick, confident lies that look authentic.
What is Fivora changing?
Thinking the EU AI Act is a cue to pack up and swear off AI forever is entirely missing the point. The tech is wildly useful. It chews through brain-numbing grunt work, speeds up messy early-stage ideation, crunches data in seconds, and lets lean, punchy creative teams hit way above their weight class without burning millions on prototyping.
Asking "Should marketers use AI?" is the wrong question—that ship sailed ages ago. You should be using AI. If you aren't, what are you doing?
The real question is: How do we leverage these tools without diluting real creativity, butchering copyright, leaking private data, or throwing human judgment out the window? That's a question worth asking and an answer we believe will always evolve.
Follow us on Instagram to learn more about what we are doing.
Comments