How Should AI Images Be Labelled?

This article is Part 2 of our series on the EU AI Act and digital marketing.

In Part 1, we looked at the bigger picture and explored how the EU AI Act may affect marketers from 2 August 2026.

Now let’s focus on one of the most visible questions for brands, agencies and marketing teams:

How should AI-generated images be labelled?

AI image generators become more powerful and the line between real and synthetic content is becoming increasingly difficult to see.

That creates exciting opportunities for brands, but it also raises important questions about transparency and trust.

When does an AI-generated image need a label?
When should businesses disclose that a visual was created or modified with AI?
And how do C2PA and Content Credentials fit into the conversation?

Let’s break it down.

 

The practical answer

 

Not every AI-generated image creates the same transparency risk.

The key question is not simply:

Was AI used?

The better question is:

Could someone reasonably believe this image is real?

If an image shows a realistic person, place, product, event or situation, transparency becomes much more important.

If an image is clearly fictional, illustrative or fantastical, the context may already make it obvious that the image is not a real photograph.

For marketers, this distinction matters.

AI image transparency is not about putting warning labels on every creative idea. It is about helping people understand what they are looking at, especially when an image could be mistaken for reality.

In practice, the more realistic and customer-facing an AI-generated image is, the more carefully businesses should think about disclosure.

 

AI-generated images are everywhere

 

A few years ago, creating a realistic image required a photographer, a studio, a location or significant design work.

Today, a marketing team can generate a professional-looking image in minutes.

Brands are already using AI-generated visuals for:

• Advertising campaigns
• Website content
• Social media posts
• Product concepts
• Internal presentations
• Creative brainstorming

This speed is powerful.

It allows teams to test ideas faster, create more visual variations and explore concepts that may previously have been too expensive or time-consuming to produce.

But the same speed also creates a new responsibility.

The challenge is not only creating better visuals. The challenge is helping audiences understand what is real, what is synthetic and how content was created.

 

What is a deepfake?

One of the key concepts in the EU AI Act is the term deepfake.

Under the Act, a deepfake is AI-generated or AI-manipulated image, audio or video content that resembles real people, objects, places, entities or events and could reasonably appear authentic.

In this context, the central issue is whether people may believe the content represents reality.

Example

Imagine a company publishes an image showing a CEO speaking at an event that never happened.

The image looks authentic.

Nothing in the image suggests it was generated by AI.

In that situation, transparency becomes important because viewers could reasonably assume the image represents a real event.

Now compare that to an illustration of a robot walking through a futuristic city floating above the clouds.

Nobody is likely to believe the image is real. 

The context is very different.

 

Why image transparency matters

 

For decades, photographs have been treated as evidence.

When people saw a photograph, they generally assumed something actually happened.

AI changes that assumption.

Today, images can be generated that are almost indistinguishable from traditional photography.

That doesn't make AI-generated images bad. Far from it.

Many brands use AI responsibly and creatively. AI can help teams explore ideas, create visual concepts and produce content faster.

That is where transparency matters.

Psychologically, transparency helps reduce doubt. When people understand what they are looking at, they are less likely to feel misled.

The goal is not to make AI-generated content feel suspicious. The goal is to make the brand feel honest.

At Les Creatives, we believe transparency is not about limiting creativity.

It is about maintaining trust.

 

What is C2PA?

 

If you've been following discussions around AI-generated content, you've probably come across the term C2PA.

C2PA stands for the Coalition for Content Provenance and Authenticity.

It is an open technical standard designed to help establish where digital content comes from and how it was created.

Think of it as a digital history record for an image.

Depending on how it is implemented, C2PA information may show:

• Which tool created the image
• Whether AI was involved
• When the image was created
• Whether edits were made
• Whether the information has been cryptographically verified

The goal is simple: to provide trustworthy information about how content was created.

For marketers, this matters because transparency will not only happen through visible labels. It may also happen through technical information attached to the content itself.

 

What are Content Credentials?

 

Adobe uses the term Content Credentials for its implementation of content provenance technology.

If you've used recent Adobe products, you may already have seen this feature.

Content Credentials allow viewers to inspect information about an image's creation and editing history.

For example, they may indicate:

• Whether AI tools were used
• Which editing software was involved
• Whether significant changes were made

You can think of Content Credentials as a user-friendly way of viewing provenance information.

In simple terms, they help answer the question:

Where did this image come from?

 

Will the EU AI Act require C2PA?

 

Not at the moment.

This is an important distinction.

The EU AI Act does not require businesses to use C2PA, Content Credentials or any other specific technology.

The law focuses on transparency outcomes rather than prescribing a particular technical solution.

That said, many technology companies and industry groups see provenance standards as an important part of the future digital ecosystem.

Businesses that understand these developments today will be better prepared as standards continue to evolve.

 

The hidden challenge: metadata disappears

 

This is one of the most overlooked issues in AI transparency.

Even when provenance information exists, it may not survive the publishing process.

Images often pass through:

• Content management systems
• Marketing platforms
• Social media scheduling tools
• Image compression services
• Content delivery networks (CDNs)

Many of these systems automatically optimise images for speed, size and performance. During that process, metadata can be removed.

This means an image may lose information about its origin before it reaches the final audience.

For marketing teams, this is worth investigating now rather than later.

Transparency is not only a policy question. It is also a workflow question.

 

Why marketers should care

 

It can be tempting to think that AI image transparency is only relevant for journalists, governments or large technology companies.

In reality, many brands already use AI-generated images every day.

The question is not whether businesses should use AI-generated visuals.

The question is how those visuals are used and whether audiences understand what they are seeing.

As AI-generated imagery becomes more common, transparency will increasingly become part of brand trust.

Businesses that embrace transparency early are likely to be in a stronger position than those that treat it as an afterthought.

 

What this means for marketers

 

You do not need to stop using AI-generated images.

You do need to understand how images are created, how they are published and what audiences might assume about them.

The more realistic and customer-facing an AI-generated image is, the more carefully businesses should think about transparency.

The businesses that succeed with AI will not be the ones hiding its use.

They will be the ones using it openly, responsibly and creatively.

 

The bigger picture

 

The future of digital content will almost certainly include a mix of human-created and AI-generated media.

That is not something to fear.

It is simply a new reality for marketers, designers and communicators.

The opportunity is to build better workflows now, before transparency becomes urgent.

At Les Creatives, we believe the most successful brands will be those that combine creativity, innovation and transparency.

Technology will continue to evolve.

Trust will remain one of the most valuable assets a brand can build.

 

Preparing your marketing for AI transparency?

 

Understanding the regulation is one thing.

Implementing it across websites, content workflows and digital channels is another.

At Les Creatives, we help businesses prepare for the next generation of digital marketing by reviewing how AI is used across content, websites and marketing workflows.

If you would like to discuss how AI transparency may affect your website, content strategy or marketing operations, we would be happy to help.

 

Continue reading

In Part 3 of this series, we tackle one of the most common questions surrounding AI and content marketing:

Do AI-written blog posts need to be disclosed under the EU AI Act?

June 2026