When do you need to label an AI visual, and when not?

This article explains the deepfake test, the form of an AI disclosure, the EU icons and the labelling of AI ads in Google Ads and Meta Ads.

Sabina Sedláčková

Sabina Sedláčková

Certified PPC Specialist & Account Manager

25. October 2026

12 min

Intermediate

Labelling an AI visual means giving visible notice that content was created or modified by artificial intelligence. The AI Act requires a label only for deepfake content. A deepfake is an AI visual that comes across as an authentic depiction of an existing person, object, place or event. Cropping, colour correction and retouching are generally edits that carry no labelling obligation. What decides the obligation is the impression the final visual creates, not the use of AI as such. Article 50 of the AI Act (Regulation (EU) 2024/1689) applies from 2 August 2026. This article explains the deepfake test, the form of an AI disclosure, the EU icons and the labelling of AI ads in Google Ads and Meta Ads.

Contents

Does every AI visual have to be labelled?

No. Using AI to create an ad does not, in itself, trigger a labelling obligation under the AI Act. For image, audio and video content, the AI Act does not separate human-made content from AI-made content. The deciding criterion is whether the result meets the definition of a deepfake.

A deepfake, under Article 3(60) of the AI Act, is image, audio or video content generated or manipulated by AI. This content resembles existing persons, objects, places, entities or events. It may falsely appear to a person to be authentic or truthful.

Who is responsible for labelling AI content?

Two parties are responsible for labelling AI content: the provider and the deployer of the AI system. The provider is the party that develops the AI system and places it on the market. The deployer is the company or organisation that uses the AI system in its activities.

The provider of a generative AI system marks its outputs in a machine-readable format under Art. 50(2). Technical solutions include metadata, watermarking and content provenance technologies. Google Cloud describes the obligations of technology providers in its documentation EU AI Act compliance – Google Cloud.

The deployer, i.e. the company or agency, ensures visible disclosure for deepfake content.

Obligation

Who is primarily responsible

Machine-readable marking of AI output (Art. 50(2))

Provider of the AI system

Visible disclosure for deepfake content (Art. 50(4))

Deployer of the AI system (company, organisation)

A watermark or metadata from an image generator does not replace the deployer's obligation to inform the audience about deepfake content.

Are you interested in PPC campaign management?

Contact us via our web form and we will be happy to take care of it.

Contact form

How does the deepfake test work?

The deepfake test consists of three conditions that must all be met at once. AI created or modified the content, the content resembles something that exists, and the result may come across as authentic. Content that fails even one condition is not a deepfake.

1 Did AI create or modify the image, audio or video?

Content created without AI does not fall under Article 50 of the AI Act. Content created or modified with AI moves on to condition 2.

2 Does the visual resemble something that actually exists?

Besides face swaps, the deepfake definition also covers existing objects, places, entities and events. Condition 2 is met by a visual that resembles:

  •         a specific person,
  •         an existing building,
  •         a shop,
  •         a product,
  •         a company,
  •         a place,
  •         an event or situation.

3 Could the result come across as authentic or truthful?

The third condition is decisive. The criterion is the impression the visual creates in the viewer, not the author's intention. European regulation aims to let audiences tell authentic content from synthetic content. Euractiv describes how the EU's AI transparency rules are changing the content shown online.

Content that meets all three conditions and creates a false impression of authentic reality is a deepfake and must be labelled.

What questions should a marketing team ask before publishing an AI visual?

Before publishing an AI visual, a marketing team checks three questions: whether it depicts a real person, whether it depicts an existing place or event, and whether the product matches reality.

Does the visual show someone who looks like a real person?

A realistic AI person calls for closer scrutiny when viewers could take them for a specific real individual. A stylised illustration, drawing, 3D character or fantasy figure does not create this impression. With these, viewers do not expect a photographic record of reality.

Example: a cartoon AI character presenting a product is assessed differently from a realistic video. In such a video, a well-known personality appears to endorse a product, even though they never made any such recording.

Does the visual show an existing place, business or event?

A photorealistic AI photo of an existing shop after a refurbishment that never took place gives customers the impression of an authentic photograph. A made-up interior scene does not depict a specific flat, hotel or shop. A photorealistic AI scene is therefore not automatically a deepfake.

Does the product shown match what the customer will receive?

This question goes beyond the AI Act. Marketing communication is also subject to consumer protection rules and the ban on misleading commercial practices.

Recolouring a sofa to a blue variant the manufacturer actually offers is a permissible technical edit. The risk of misleading advertising arises with these AI interventions:

  •         making the product look bigger,
  •         adding a feature the product does not have,
  •         creating a variant that does not exist,
  •         changing the material,
  •         showing a result the customer will not get.

A “created with AI” label does not make a false advertising claim legal.

Which AI visuals do not need a label?

AI edits that do not change the meaning of the reality depicted do not need a label. Most everyday design and advertising workflows sit outside the deepfake regime. Typical AI edits with no labelling obligation:

  •         cropping a photo,
  •         resizing,
  •         adjusting brightness and contrast,
  •         colour correction,
  •         sharpening,
  •         noise reduction,
  •         removing dust and minor imperfections,
  •         simple retouching,
  •         technically extending an image for another ad format,
  •         composing and arranging products,
  •         recolouring a product to a variant that actually exists.

When does a background swap need a label?

Whether a background swap needs a label depends on the context of the final visual. A product placed in a generic, made-up interior does not depict existing reality. A product placed by AI into a specific, existing hotel creates the impression that it is really there, and may show the hallmarks of a deepfake.

What does a correct AI disclosure look like?

A correct AI disclosure informs the user clearly and distinguishably, at the latest at the first exposure to the deepfake content. The wording reflects the actual extent of AI involvement:

  •         fully generated content: “Visual created using artificial intelligence.”
  •         partially modified authentic content: “Visual modified using artificial intelligence.”

The disclosure does not belong in the general terms and conditions or somewhere reachable only after several clicks.

What are the EU icons for labelling AI content?

The EU icons for labelling AI content are the European Commission's official set of visual marks. The set contains three types of icon:

  •         the basic AI icon,
  •         Fully AI-Generated,
  •         Partially AI-Modified.

Each icon comes in a black, a white and a 50% transparent version. The European Commission describes the use of the icons as voluntary. An icon on its own does not guarantee compliance with Article 50 of the AI Act.

The European Commission recommends placing the icon visibly at the user's very first contact with the content. The icon also stays in place when the content is shared further or downloaded. As an example for the Partially AI-Modified icon, the Commission cites an authentic photo of an empty flat that AI has virtually furnished (EU Icons for labelling AI-generated content).

How is artistic, satirical and fictional content labelled?

Evidently artistic, creative, satirical or fictional content is subject to an adapted transparency obligation. The obligation does not disappear. The deployer adapts the form of the disclosure so that it does not hamper the display or enjoyment of the work. A film or artistic work therefore does not need a large warning across the image.

How do advertising platforms label AI content?

Google Ads and Meta Ads label AI content with their own mechanisms, independently of the AI Act. Platform labelling and the legal obligation under the AI Act are two separate levels. In online advertising, three situations can arise:

  1.       The AI Act does not require a label and the platform does not add one.
  2.       The AI Act does not require a label, but Google or Meta show their own information about the use of AI.
  3.       The AI Act requires a label and the advertising platform also identifies the AI content.

How does Google Ads label AI content?

With Google Ads, a marketing team distinguishes the legal obligation under the AI Act from the platform's technical and transparency mechanisms. Google is developing its own mechanisms for labelling and tracking content created or modified by generative AI.

Does using AI in Google Ads mean the ad has to be labelled?

No deepfake is created when AI makes these edits:

  •         adjusting brightness,
  •         cropping a photo,
  •         extending the background,
  •         removing a small distracting object,
  •         preparing another size of an existing ad.

Google may apply its own AI content labels, metadata and technical signals regardless of the legal deepfake threshold. A marketing team therefore checks two things:

  1.       Does the AI Act require a label? The team runs the deepfake test.
  2.       Does Google Ads require a label? The team checks the current settings and policies of the ad platform.

The answers to these two questions may differ.

How does Meta Ads label AI content?

Meta Ads applies the same principle to ads on Facebook and Instagram. Meta uses its own AI info label and technical mechanisms to identify content that generative AI has created or significantly modified.

Does a minor AI edit trigger a label in Meta Ads?

Not necessarily. A technical edit of an existing photo differs from an intervention in which AI creates:

  •         a photorealistic person,
  •         a synthetic situation,
  •         a realistic event that never happened,
  •         a fundamentally changed meaning of the original photo.

Meta may apply its own transparency system even to content that does not reach the deepfake threshold under the AI Act.

Can Meta detect AI content created outside its own tools?

Yes, Meta can identify AI creatives made in an external AI generator. Modern systems use metadata, content provenance standards and other technical signals. A company should therefore not assume that Facebook or Instagram will treat an external AI image as an ordinary photo.

Google Ads vs. Meta Ads: comparing how AI ads are labelled

Situation

AI Act

Google Ads

Meta Ads

Colour correction using AI

Generally no label

Possible technical AI signals

For a minor edit, generally no visible AI label

Removing a small object

No label if it does not change the meaning of reality

Depends on how the asset was created

Depends on the extent of the edit

Generic AI scene

Not a deepfake if it does not depict existing reality

The platform may record the AI origin

The platform may label the AI content

Realistic manipulation of an existing person

Possible deepfake → label

A platform label does not replace the legal obligation

A platform label does not replace the legal obligation

Fake photo of an existing place or event

Possible deepfake → label

Separate legal assessment

Separate legal assessment

AI-generated product feature that does not exist

Risk of misleading advertising

Risk of breaching advertising policies

Risk of breaching advertising policies

How do marketing agencies check AI creatives?

Marketing agencies check AI creatives on three layers: the AI Act, advertising and consumer law, and the ad platform's policies. Passing one layer does not guarantee passing the other two.

  1.       AI Act: Does the visual present itself as an authentic existing person, object, place or event? A yes calls for the deepfake test and an assessment of labelling.
  2.       Advertising and consumer law: Do the product, service, result and advertising claim match reality? An AI label does not make misleading advertising legal.
  3.       Ad platform: Does Google, Meta or another ad system require or automatically add its own AI label?

Are you interested in PPC campaign management?

Contact us via our web form and we will be happy to take care of it.

Contact form

When should you consult a lawyer about labelling AI visuals?

A legal or compliance assessment is needed in situations that go beyond what a designer or PPC specialist can check:

  •         an AI brand ambassador across multiple campaigns,
  •         an AI character resembling a specific person,
  •         a digital twin of a real person,
  •         a synthetic voice of a specific individual,
  •         AI customer reviews,
  •         fictitious references,
  •         AI testimonials,
  •         a realistic AI reconstruction of an actual event,
  •         a fake photo of a real business,
  •         an AI presentation of a product or service result that differs from reality.

In these cases, the AI Act overlaps with the rules on advertising, consumer protection, personality rights, copyright and personal data protection.

What deadlines and penalties does the AI Act set?

The transparency obligations of Article 50 of the AI Act apply from 2 August 2026. For relevant infringements, penalties for businesses reach up to €15 million or 3% of total worldwide annual turnover, whichever is higher (Art. 99(4)).

The European transparency framework is increasing the number of AI labels, watermarks and content provenance notices that users see online. Euractiv analyses the impact of the AI transparency rules on how people use the internet.

Deadline or area

Meaning

2 Aug 2026

Art. 50 transparency obligations start to apply

Deepfake

Visible notice of content created or manipulated by AI

EU AI icons

Voluntary tool for transparent labelling

Google Ads

Its own technical and transparency mechanisms

Meta Ads

Its own AI info label and AI content detection

Penalty

Up to €15 million or 3% of worldwide annual turnover (whichever is higher)

How do you build AI visual checks into your company workflow?

A company workflow does not need a complex approval process for every photo opened in an AI tool. A six-point checklist takes care of the checks:

  1.       Did the team use AI?
  2.       What did AI change?
  3.       Does the change concern a real person, object, place or event?
  4.       Could the result come across as authentic?
  5.       Do the product and the advertising claim match reality?
  6.       Does Google Ads, Meta Ads or another platform require an additional label?

Using AI does not create an obligation to label every visual. A deepfake does require a label. Google Ads and Meta Ads apply their own AI labels, even beyond the legal obligation under the AI Act.

So the key question for a marketing team is not “Did we use AI?”. The key question is: “What did AI change in the final visual, and what impression will it create for the customer?”

This article is for information only and does not replace individual legal advice.

Recommended reading

The latest trends, proven practices and analyses from the digital world. Find out how to take your technology and marketing growth further.