Artificial intelligence is already part of the day-to-day work of marketing teams. It is used to generate images, write ads, create videos, optimise campaigns, analyse results and personalise content at a speed that would have been unimaginable just a few years ago.
However, as AI becomes increasingly integrated into business processes, a new need also emerges: using it responsibly and transparently.
With this objective in mind, the European Union approved the European Artificial Intelligence Regulation (AI Act), the first regulatory framework designed to establish common rules for the development and use of artificial intelligence.
Although the regulation covers many different areas, one section directly affects marketing and digital advertising. From 2 August 2026, the transparency obligations set out in Article 50 will begin to apply, introducing new requirements for certain content generated or modified using AI.
What really changes for brands? Will all AI-generated creatives need to be identified? How are platforms responding? In this article, we answer the main questions.
What is the AI Act?
The AI Act is the European Union regulation that establishes a common framework for the use of artificial intelligence according to the level of risk it may pose to individuals.
Its aim is not to limit innovation, but to ensure that AI is used safely, transparently and in a way that respects fundamental rights.
The regulation classifies AI systems into different risk levels and establishes specific obligations for each one. In marketing, the main focus is on the transparency obligations set out in Article 50.
What changes from 2 August 2026?
From this date, new transparency obligations will begin to apply to certain AI systems and AI-generated content.
This means that, in some cases, users will need to be informed when they are interacting with an AI system or when content generated or modified using artificial intelligence could be misleading, or when the platform on which it is published requires it.
It is important to understand that the regulation does not prohibit the use of AI to create campaigns or content.
What changes is how certain uses must be communicated.
What does Article 50 say?
Article 50 establishes different transparency obligations depending on the type of AI system and how it is used.
When a person interacts with an AI system
If a user has a conversation with a chatbot, virtual assistant or AI-generated avatar, they must be informed that they are interacting with an AI system, unless this is already obvious from the context.
The aim is to prevent someone from believing they are speaking to another person when they are actually interacting with a machine.
When synthetic content is used
The regulation also establishes obligations for certain content generated or modified using artificial intelligence.
This may include:
- AI-generated images;
- videos created or modified using AI;
- synthetic voices;
- digital avatars;
- manipulated content that could appear real.
This is particularly relevant when the content could be misleading or when platform policies require it to be identified.
The case of deepfakes
One of the best-known aspects of Article 50 concerns so-called deepfakes.
When an AI-generated image, video or audio recording depicts a person saying or doing something that never happened, it must be identified as artificially generated or manipulated content, except in the circumstances provided for by the regulation, such as certain artistic or satirical uses or content subject to editorial oversight.
Will all AI-generated content need to be labelled?
The short answer is not necessarily.
One of the most common misconceptions is that any content created using tools such as ChatGPT, Gemini, Copilot or Midjourney will need to display a visible label.
The AI Act does not establish a general labelling obligation for all AI-generated content.
The obligations depend on the type of content, the context, the system used and, in some cases, the requirements established by each platform.
For example, a marketing team may use AI to create a first draft of an article or advertising copy that is subsequently reviewed, corrected and approved by a person. In these cases, human intervention continues to play a fundamental role.
What does this mean for marketing teams?
Beyond regulatory compliance, the AI Act represents a change in how artificial intelligence is managed within organisations.
Simply introducing new tools is no longer enough.
Processes must also be established to ensure that these tools are used in a controlled and transparent way.
In practice, this means:
- understanding which AI tools are used within the organisation;
- identifying which creatives include AI-generated content;
- defining human review and oversight processes;
- protecting the information and data being used;
- adapting to the specific requirements of each platform.
In other words, moving from the isolated use of AI tools to a proper governance strategy.
How are platforms responding?
The main advertising platforms are already introducing mechanisms to adapt to the new regulatory framework.
Although each platform is implementing these measures differently, they are all moving towards the same objective: providing greater transparency regarding the origin of certain AI-generated content.
For this reason, before launching a campaign, it is advisable to review the specific requirements of each platform and check how the use of AI must be disclosed when necessary.
How are the main platforms adapting their policies?
Digital platforms are not applying transparency requirements in the same way. Although they are all moving towards clearer identification of content generated or modified using artificial intelligence, the format of the disclosure, the declaration process and its visibility may vary depending on the channel.
Google Ads allows advertisers to indicate whether an asset has been generated or edited using AI. This information may appear in the “How this ad was created” section of My Ad Center and, for campaigns targeting the European Union, it may also be displayed through a visible disclosure on the ad itself.
Google may also automatically label certain assets created using its own AI tools or when it detects signals originating from other systems. Since July 2026, it has also allowed advertisers to include transparency text or visual elements directly in images and videos without these being considered a breach of its policies on overlays or watermarks.
In the case of Google Merchant Center, AI-generated images must retain metadata that identifies their origin. Titles and descriptions created using AI may also require specific structured attributes within the product feed.
Meta
Meta identifies certain AI-generated or AI-modified content using technical signals, industry-wide standards and information provided by users or advertisers.
In advertising, Meta already labels ads that have been created or significantly modified using its generative AI tools. The company also uses notices such as “AI info” or equivalent labels to provide additional context about the origin of the content.
The visibility of the notice may depend on the type of modification and where the ad is displayed. It should therefore not be assumed that every creative will receive exactly the same label.
TikTok
TikTok allows advertisers to declare through Ads Manager that an ad contains images, videos or audio that have been fully generated or significantly modified using AI.
When the advertiser activates this option, TikTok adds a text disclosure indicating that the ad includes AI-generated content. For certain in-house creation tools, such as Symphony Creative Studio, the platform may add the disclosure automatically.
TikTok states that correctly activating transparency disclosures does not in itself reduce content distribution. However, commercial content that should have been declared but was not may lose eligibility to appear in the For You feed.
Pinterest may identify content generated or modified using AI through file metadata, information provided by the owner and its own detection systems.
For ads, the disclosure does not necessarily appear directly on the Pin. Users may access it alongside other ad details through the options menu. Pinterest may indicate that content has been modified using AI, that it contains an artificially generated person or that it may have been altered using AI.
LinkedIn is focusing on content credentials and digital provenance standards, such as C2PA, to provide information about the origin of a piece of content and any modifications made to it.
These credentials can retain information about how a file was created or edited and which tools were involved in the process. Rather than being a simple visible label, they provide a technical layer of traceability that accompanies the content.
Snapchat
Snapchat uses different indicators to identify content created using its own generative AI features. Some images generated or edited using tools such as AI Snaps may include a watermark featuring the Snapchat ghost icon accompanied by sparkles.
The platform may also display context cards to inform users that an image was created using a generative AI feature. This measure mainly applies to content produced within Snapchat’s own tools and should not be interpreted as an automatic watermark for every external asset uploaded to the platform.
What does this diversity mean for brands?
The transparency obligation may be shared across platforms, but its implementation is not fully standardised.
The same creative may require:
- a declaration within the ad manager;
- a notice accessible through the ad information section;
- a visible label;
- technical metadata;
- provenance credentials;
- or a combination of several mechanisms.
Brands therefore need to identify from the outset which assets have been generated or modified using AI and communicate this information to the team responsible for activating the campaigns.
For static campaigns, this identification can be managed directly for each creative. For dynamic or catalogue-based campaigns, it will need to be incorporated into the data flow and feed management process.
At Adsmurai, we recommend reviewing each case according to the type of content, market and platform. Declaring the use of AI does not, in itself, result in a distribution penalty. However, an incorrect declaration or the omission of required information may lead to reviews, restrictions or requests for modification from the platform.
Transparency is no longer purely a creative consideration. It also affects campaign configuration, feeds, metadata and approval processes.
Best practices for brands
The introduction of these obligations represents a good opportunity to review internal processes related to artificial intelligence.
Some recommendations include:
- establishing an internal policy on the use of AI;
- maintaining human oversight within creative processes;
- identifying which content uses generative AI;
- documenting cases when necessary;
- keeping up to date with the policies of the platforms on which campaigns are activated.
More than a legal matter, this is about building trust.
AI needs transparency to build trust
Artificial intelligence will continue transforming how brands create content, optimise campaigns and make decisions.
The AI Act is not intended to slow down this evolution. Its objective is to ensure that innovation is accompanied by transparency and responsibility.
For marketing teams, the challenge is no longer simply to incorporate AI into their daily work, but to do so through clear processes, human oversight and a strategy that makes it possible to harness its full potential without losing users’ trust.
Because the future of marketing will not depend solely on who uses the most artificial intelligence. It will depend on who is capable of integrating it most effectively into a governable, transparent growth system that is prepared to evolve alongside regulation.
