Artificial intelligence has long been more than just a trend. It is transforming the way digital products are developed, websites are operated, and business processes are automated.

This opens up enormous opportunities for resellers, agencies, domain investors, and online entrepreneurs: content can be created more quickly, customer inquiries can be automated, and digital projects can be scaled more efficiently.

At the same time, the legal framework is changing. With the EU AI Act (Regulation (EU) 2024/1689), the European Union is establishing comprehensive rules for the use of artificial intelligence for the first time.

August 2, 2026, is a particularly important date. Starting on that date, additional transparency requirements will apply to certain AI applications and AI-generated content.

The main message is:

AI may continue to be used extensively, but users must be able to recognize in certain situations when they are interacting with AI or viewing AI-generated content.

For digital service providers, this means that AI expertise is not only becoming a legal requirement but also, increasingly, a mark of quality.

The EU AI Act: Why the AI Regulation Will Be Relevant to Digital Business Models

The EU AI Act takes a risk-based approach. This means that not every AI application is regulated in the same way. The key factor is the level of risk posed by an application.

Among other things, the regulation distinguishes between:

Prohibited AI Systems

This includes applications whose risks are deemed unacceptable and which may therefore be prohibited.

Examples:

• Social scoring systems that evaluate people based on their behavior, social characteristics, or personal traits, and use this information to place them at a disadvantage.
• Manipulative AI systems designed to influence people in specific ways to make them take decisions they might not otherwise have made.
• Certain biometric categorization systems that classify individuals based on sensitive characteristics.
• Emotion recognition in certain areas, such as when it leads to inappropriate judgments or decisions about individuals.

For traditional online retailers, agencies, or domain investors, these systems generally aren’t relevant. What matters more are the transparency requirements for everyday AI applications.

High-Risk AI Systems

These relate to particularly sensitive areas, such as certain applications in personnel decisions, critical infrastructure, or biometric procedures.

Examples:

• AI-powered candidate selection, in which systems analyze resumes or evaluate candidates.
• AI systems in the field of education that, for example, play a role in decisions regarding admission, evaluation, or classification of individuals.
• AI applications in healthcare, such as systems that assist doctors with diagnoses or medical decisions.
• AI in critical infrastructure, such as for controlling key energy, transportation, or utility systems.
• Certain biometric identification systems, such as those used for the automated recognition or identification of individuals.

For most resellers, agencies, and website operators, these applications are less relevant—unless they develop or integrate specialized solutions of this kind.

AI Systems Subject to Transparency Requirements

This category is particularly interesting for digital companies because it involves many everyday AI applications.

The goal is to protect users from being misled by the use of AI.

These include, among other things:

AI-generated or AI-manipulated content

Examples:

• a photorealistic AI image of a person who does not actually exist,
• a doctored video in which a real person appears to make statements that they never actually made,
• an AI-generated audio recording that sounds like a real person’s voice.

The main point here is that users should be able to recognize when content has been artificially generated or altered.

AI Systems for Direct Communication with Users

Examples:

• an AI chatbot in the online store,
• a virtual support assistant on a website,
• an automated sales advisor.

In these cases, users must be able to recognize that they are communicating with an AI and not with a human.

Certain generative AI applications

Examples:

• AI tools for creating images, videos, or audio content,
• Systems capable of generating realistic-looking synthetic media.

This category is particularly relevant for resellers, agencies, and domain investors because such applications are increasingly being used in marketing, e-commerce, and content production.

Low-risk AI systems

AI systems with low or no risk

Many everyday AI applications fall into this category.

Examples:

• Writing assistants that improve the language of texts or correct spelling errors.
• Translation tools that automatically translate content into other languages.
• Spam filters that detect unwanted emails.
• Recommendation systems that, for example, suggest relevant products or content.
• Automation tools that simplify workflows.
• Analytical tools that summarize data or identify patterns.

Even though these applications are generally less heavily regulated, companies should still use them responsibly and monitor the results.

For resellers, agencies, and domain investors, transparency requirements are the primary focus. The following are therefore particularly relevant:

• AI-generated images, videos, and other media content,
• AI-powered chatbots and assistants,
• automated content processes,
• Transparency regarding the use of AI with customers and users.

The key question here is not: Was AI used?

Rather:

Can the use of AI lead to users being misled about the source, authenticity, or nature of a communication?

The Timeline for the EU AI Act: Why Companies Should Act Now

The EU AI Act is not a law that will take effect only in 2026. The regulation already entered into force on August 1, 2024, and is being implemented in phases.

Key dates:

August 1, 2024
The AI Act officially takes effect.

February 2, 2025
The first regulations take effect. These include, among other things, guidelines on prohibited AI practices as well as requirements regarding the AI proficiency of individuals who work with AI systems.

August 2, 2025
Additional requirements, including those for certain general-purpose AI models, will take effect.

August 2, 2026
A large portion of the transparency requirements under the AI Act will take effect. These include, in particular, guidelines for certain AI-generated or AI-manipulated content, as well as requirements for certain AI systems where users must be informed that they are interacting with an AI.

For companies, this means:

Anyone who doesn’t start preparing until 2026 is wasting valuable time. Professional AI processes should be established well before then.

Who is affected by the AI Act?

A common misconception is that only AI developers or large technology companies are affected by the AI Act.

In fact, the regulation also applies to companies that use AI systems and make their results available.

These include, for example:

• Web agencies that integrate AI features into their clients’ websites
• Resellers who offer digital products with AI components
• Online Store Owners with AI Assistants
• SEO Agencies with AI-Powered Content Production
• Domain investors with automated content projects
• Operators of platforms or digital services

What matters is not who programmed the AI model.

The key factor is who uses AI and what impact that use has on customers, visitors, or users.

Provider or Operator? The Important Distinction in the AI Act

For professional providers, there is an important distinction to note: A company that develops and brings an AI system to market is different from a company that uses an existing AI solution.

An agency that, for example, integrates an external AI chatbot into a client’s website does not develop the chatbot itself. However, as a user or operator of an AI system, it may have to comply with certain obligations.

This raises some practical questions:

• Who informs users about the use of AI?
• Who monitors the print quality?
• Who is responsible for ensuring proper integration?
• What information must be disclosed to customers?

Therefore, especially for agency projects, the proposal or project contract should specify which AI components will be used and who will assume which responsibilities.

AI Labeling Requirement: Not All AI Content Must Be Labeled

One of the biggest mistakes in the current discussion is the statement: “Starting in 2026, everything created using AI must be labeled.”

That’s not correct.

The AI Act does not require a blanket labeling requirement for all AI-generated content.

The transparency requirement is particularly relevant in the case of so-called deepfakes.

What is a deepfake under the AI Act?

A deepfake is defined as image, audio, or video content that has been created or altered using artificial intelligence and bears a deceptively close resemblance to real people, objects, places, institutions, or events.

The decisive factor is the potential for deception.

The question is:

Could a user believe the content is real, even though it was artificially generated or altered?

Examples:

• an AI-generated photo of a person who appears to be real
• a doctored video of a well-known person
• an artificially created recording of an alleged event

The following are not automatically affected:

• clearly recognizable illustrations,
• Fantasy depictions,
• obviously computer-generated graphics,
• Creative designs that are not intended to be realistic.

Section 50 of the AI Act: The Key Provision on Transparency for AI-Generated Content

The relevant transparency requirements are set forth, in particular, in Article 50 of the EU AI Act.

The regulation has a clear objective:

People should be able to tell when they are interacting with AI or when certain content has been artificially generated or manipulated.

For digital companies, this means that it is not the mere use of AI that matters, but rather the specific way it is applied.

A company that uses AI to support a creative process should be viewed differently from a company that publishes realistic-looking synthetic content that users might mistake for the real thing.

Real-World Examples for Resellers and Agencies

AI-Generated Employee Photos on Websites

An agency creates a company website and uses AI-generated images of supposed employees.

This can be problematic if visitors might assume that these are actual team members.

AI-generated product images

A store creates photorealistic product images without using actual photos.

In this context, it must be determined whether users could be misled as to the actual origin or reality of the depiction.

AI-generated testimonials

Artificial reviews, customer testimonials, and experience reports are particularly sensitive.

Even aside from the AI Act, competition law risks may arise here if artificial content is presented as genuine customer experiences.

How is AI-generated content labeled?

If labeling is required, it must be clearly visible to users.

Examples of suitable instructions include:

• “AI-generated image”
• “Created using artificial intelligence”
• “AI-manipulated content”

Important:

The information should be provided directly within the content or in its immediate vicinity.

A general note in:

• Legal Notice,
• Privacy Policy,
• Footer

It is not sufficient for specific content to be involved.

The AI Act does not prescribe any specific standard wording.

Technical Identification: Metadata, Watermarking, and Content Credentials

In addition to visible cues, many companies are grappling with another question:

Does AI-generated content have to be technically labeled?

The AI Act establishes transparency requirements for certain AI-generated or manipulated content. The specific technical implementation of these requirements can take various forms.

Possible technical methods include, for example:

• Metadata in media files,
• digital provenance information,
• Content-Credentials scheme,
• Technical markings by AI providers.

The following applies:

A technical label does not automatically replace the necessary user information. What remains crucial is that people be able to recognize when they are encountering relevant AI-generated or AI-manipulated content.

For agencies and resellers, this means:

When selecting AI tools, one should not focus solely on the quality of the results. The question of which transparency features a provider supports is also becoming increasingly important.

Professional selection criteria for AI tools may therefore include:

• Does the tool support origin information?
• Can content be labeled in a way that is easy to understand?
• Is there any documentation on the use of AI?
• Are updates regarding regulatory requirements provided?

Especially for larger client projects, this technical transparency can become a true mark of quality.

AI-Generated Text and SEO: What Content Professionals Should Keep in Mind

The good news:

Not all AI-generated text needs to be automatically labeled.

Using AI as a tool for writing does not automatically mean that every article, product description, or landing page needs to include a note to that effect.

Typical applications of AI:

• Topic research,
• Concept development,
• Creation of initial drafts,
• linguistic optimization,
• Structuring content.

However, human responsibility remains the decisive factor.

Professional AI-generated content should:

• be reviewed by experts,
• be tailored to the target audience,
• be checked for accuracy,
• be supplemented by our own expertise.

This is leading to an important development, particularly in the field of SEO:

The volume of automatically generated content will increase. At the same time, high-quality content that demonstrates genuine expertise, cites verifiable sources, and offers a unique perspective will become more important. While AI makes it possible to build larger content portfolios, in the long run, it won’t be the projects with the largest volumes that succeed, but rather the most trustworthy ones.

A special exception: AI-generated texts used to shape public opinion

The AI Act takes a particularly critical view of AI-generated texts when they have the potential to influence public opinion.

One example might be politically related content. In such cases, a transparency requirement may apply when AI-generated content is published that is likely to influence public debate.

However, this special rule does not automatically apply to traditional corporate communications.

A product description, an advice article, or a company website is not required to be labeled simply because AI was used in its creation.

AI Chatbots: Transparency in Customer Interactions Becomes Mandatory

AI chatbots are now among the most important practical applications of artificial intelligence.

They are used for:

• Customer Service,
• Product advice,
• Lead generation,
• Scheduling an appointment,
• Support Automation.

The AI Act requires that, for certain AI systems, users be able to recognize when they are communicating with an AI. This information should be provided at the beginning of the interaction.

Here’s an example:

“Hello, I’m XYZABC’s AI assistant, and I’m here to help you with your questions.”

What matters is not the specific wording, but clear communication.

The user must be able to recognize:

• Is the person I’m talking to a human being?
• Or is it an automated AI system?

What are the rules for platforms and marketplaces?

Many resellers and retailers use external platforms.

An important distinction must be made here:

If a platform provides its own AI assistant or chatbot, the responsibility for labeling it correctly generally lies with the platform operator.

A merchant who merely uses the platform is not automatically required to ensure the transparency of this third-party AI system.

The situation is different when a company integrates its own AI chatbot or provides its own AI feature. In that case, the implementation itself should be reviewed.

Third-Party AI: Why Tool Selection and Contracts Are Becoming More Important

Many companies use AI services from third-party providers.

Examples:

• AI chatbots as a SaaS solution,
• Image generators,
• Text assistants,
• AI plugins for e-commerce platforms,
• Automation tools.

A common mistake:

“The provider is responsible, so this doesn’t affect me.”

This assumption can be problematic. Even if a company does not develop its own AI, it should consider the following:

• What features does the tool offer?
• What data is processed?
• What options are available for transparency?
• What are the provider’s responsibilities?
• What information does the customer need?

Agencies are advised to describe AI components transparently in project proposals and contracts.

Possible points:

• AI technologies used,
• Responsibilities,
• Testing processes,
• Dealing with AI-generated content.

AI and Data Protection: The AI Act Is Not the Only Legal Challenge

In addition to the AI Act, companies must continue to comply with other legal requirements when using AI, particularly data protection regulations.

AI chatbots and automated assistance systems, in particular, can process personal data.

Examples:

• Customer names,
• Email addresses,
• Support requests,
• Ordering Information,
• user-specific entries.

Companies should therefore consider the following:

• What data does the AI system process?
• Where is the data stored?
• Is the data used for training purposes?
• Are there any necessary agreements with the provider?
• Will the privacy policy be updated?

For agencies, this means:

Integrating AI is not just a technical task. Data protection, security, and transparency are also key to professional implementation.

AI Literacy: Why Training Is Becoming More Important

The AI Act also emphasizes the importance of ensuring that individuals who use AI systems have sufficient AI literacy.

This doesn’t just apply to developers.

Employees in:

• Marketing,
• Sales,
• Content creation,
• Customer Service,
• Web Development

may be affected if they work with AI systems on a regular basis.

A professional AI strategy should therefore include:

• Training courses,
• internal guidelines,
• clear responsibilities,
• Quality controls.

For agencies, this could even become an additional service offering.

Those who not only build websites for clients but also help them use AI safely create additional value.

What are the potential consequences of violations?

The AI Act provides for penalties in the event of violations.

The specific amount of any fines depends, among other things, on:

• Type of violation,
• Severity of the impact,
• Company size,
• other circumstances of the specific case.

These regulations are not intended solely for international corporations.

Even smaller companies can be affected if they use relevant AI applications and fail to comply with transparency requirements.

In practice, this means that AI compliance should not be addressed only when an audit takes place. Professional companies establish the necessary processes early on.

Voluntary EU Code of Conduct: Establishing Standards Early On

In addition to the mandatory requirements, the European Union also offers companies the opportunity to voluntarily follow additional guidelines.

A voluntary code of conduct can help companies:

• Implement transparency standards,
• to establish internal processes,
• to raise awareness of the responsible use of AI.

This can be particularly interesting for agencies and resellers, because trust is increasingly becoming a key factor for success in digital business.

FAQ: The Most Important Questions About the 2026 AI Labeling Requirement

Do I have to label every AI-generated image on my website?

No. Not every AI-generated image is automatically subject to a labeling requirement.

What is particularly relevant is whether the content is AI-generated or manipulated in such a way that it bears a deceptive resemblance to real people, places, objects, or events and could therefore be mistaken for them.

Do I need to label ChatGPT text on my website?

No, not automatically.

The use of AI to assist with text creation does not, as a general rule, result in a requirement to disclose its use.

Quality assurance, human review, and the specific requirements for content that shapes public opinion are important.

Does an AI chatbot need a hint?

Yes, when users interact with an AI system.

Communication should be designed in such a way that it is clearly evident that the response is coming from an AI, not a human.

Is a note in the legal notice sufficient?

No.

If specific content needs to be labeled, the information should be visible directly on the content or in its immediate vicinity.

Is an agency required to inform clients about the AI it uses?

A general answer depends on the specific application.

From a professional standpoint, however, it is advisable to transparently document the use of AI in projects and to clearly define responsibilities.

AI Act Checklist for Resellers, Agencies, and Domain Investors

Companies can prepare by asking themselves the following questions:

Track AI Usage

✓ What AI tools are used?
✓ What content or features result from this?
✓ Which client projects use AI?

Define Processes

✓ Who verifies AI results?
✓ Who decides whether something is published?
✓ How is AI-generated content documented?

Ensuring Transparency

✓ Are AI chatbots easily recognizable?
✓ Is AI-generated content that appears realistic reviewed?
✓ Are the necessary instructions provided?

Prepare employees

✓ Are users of AI systems trained?
✓ Are there internal guidelines?
✓ Is it clear when human oversight is necessary?

Ensuring the Success of Client Projects

✓ Are AI features described in the listing?
✓ Have responsibilities been defined?
✓ Are the right tools being used?

Conclusion: AI isn’t going away, but knowing how to use it professionally will be crucial

The EU AI Act does not change the fact that AI is becoming a central component of digital business models. However, it does change what is expected of companies. Resellers, agencies, and domain investors who use AI do not need to slow down. They need to become more structured.

The successful companies of the coming years will be those that don’t simply use AI for automation, but combine it with strategy, transparency, and quality control. After all, digital users are increasingly accepting AI. What they will not accept is the feeling of having been deceived by AI.

Transparency is therefore not just a legal requirement.

It becomes a competitive advantage.