Key Takeaways

1. The EU AI Act now requires transparency for certain AI uses. Article 50 of the EU AI Act took effect on August 2, 2026, and requires businesses to disclose certain AI interactions and AI-generated or manipulated content, depending on how the technology is being used.

2. U.S. businesses may still be affected. The law can apply to organizations outside the EU when AI systems are provided in the EU or when AI-generated outputs are used within the EU, meaning a company does not need a physical European presence for the rules to potentially apply.

3. Not all AI-generated content requires labeling. Contrary to some misconceptions, Article 50 does not require companies to place AI-generated disclaimers on every blog post, webpage, or piece of content. Meaningful human review and editorial control can exempt certain content from disclosure requirements.

Artificial intelligence is quickly becoming part of everyday business. Companies are using AI to answer customer questions, create content, generate images and videos, personalize digital experiences and automate work that once required a person.

As AI becomes more common, governments are establishing rules for how businesses use it. One of the most significant examples is the European Union’s AI Act, which establishes requirements for companies that develop and use certain types of artificial intelligence.

One portion of the law, Article 50, focuses specifically on transparency. In simple terms, its goal is to make sure people know when they are interacting with AI or, in certain situations, when something they are seeing, hearing or reading was created or manipulated by AI.

These Article 50 requirements went into effect on August 2, 2026.

What is Article 50?

Article 50 is the transparency section of the EU AI Act. At its most basic level, it says companies using certain types of AI need to be transparent about it. The exact requirement depends on how the AI is being used.

For example, someone having a conversation with an AI-powered chatbot may need to be told they are interacting with AI. Certain realistic AI-generated images, audio or video may need to be identified as AI-generated or manipulated. There are also specific rules covering emotion-recognition systems, biometric categorization and some AI-generated text involving matters of public interest.

It does not mean that every business using AI must label everything it creates as “AI-generated.”

Who Could Be Affected?

This is where the law may surprise some U.S. businesses.

A company does not necessarily need a physical office or headquarters in Europe for the EU AI Act to apply. The law applies to companies providing AI systems in the EU regardless of whether they are located inside or outside the EU. It can also apply to providers and deployers outside the EU when the output produced by an AI system is used within the EU.

That means U.S.-based companies doing business internationally should not assume the law doesn't concern them simply because they have no European office.

Having an EU customer by itself, however, does not automatically mean Article 50 applies. Companies need to consider what AI they are providing or using, where it is being used and what the AI is doing.

What Does the Law Require?

The easiest way to understand Article 50 is through a few common examples.

AI chatbots and assistants: When an AI system is designed to have a direct conversation with people, those people generally need to know they are interacting with AI unless it is already obvious. The European Commission says disclosure should occur from the beginning of the interaction.

AI-generated images, audio, video and text: Companies that provide generative AI systems have requirements to make certain AI-generated or manipulated content detectable using machine-readable markings. There are exceptions and limitations, including for certain standard editing functions and qualifying business-to-business or industrial uses.

Deepfakes: If AI creates or manipulates realistic images, audio or video so that an existing person, object, place, organization or event could falsely appear authentic, the content generally needs to be clearly disclosed as AI-generated or manipulated. A technical marker hidden in the file is not enough for this disclosure requirement; people need to be able to perceive it.

AI-generated written content: This requirement is narrower than it may initially sound. The rule covers AI-generated or manipulated text published to inform the public about matters of public interest, such as politics, public health, public safety, consumer safety and certain economic or financial developments.

Importantly, qualifying written content that has gone through meaningful human review or editorial control and for which someone takes editorial responsibility does not need to carry the Article 50 AI label. Simply running spellcheck or checking grammar does not qualify as substantive human review.

So, no: the EU is not requiring every company that uses AI to help write a blog post or webpage to put an “AI-generated” disclaimer on it.

Why Should Companies Pay Attention?

AI is being added rapidly to websites, marketing platforms, customer service tools and digital experiences. In many organizations, different departments may be adopting AI independently, making it increasingly important to understand where and how AI is actually being used.

For companies with business or users in Europe, Article 50 provides another reason to inventory customer-facing AI experiences and establish clear internal processes around AI-generated content.

A few straightforward questions can provide a useful starting point:

•      Are customers interacting directly with AI anywhere on our website or digital platforms?
•      Are we creating realistic AI-generated or manipulated images, voices or videos?
•      Are we publishing AI-generated content about matters of public interest?
•      Do we have meaningful human review and editorial responsibility for AI-assisted content?
•      Are any of these AI systems or their outputs being used in the European Union?

The answers can help identify where closer legal and compliance review may be appropriate.

What Happens if a Company Doesn't Comply?

The EU AI Act includes potentially significant penalties.

According to the European Commission, violations of the Article 50 transparency requirements can result in fines reaching €15 million or 3% of a company's total worldwide annual turnover for the preceding financial year. The law also provides for proportionality considerations involving smaller and medium-sized businesses.

That doesn't mean every mistake will result in a multimillion-euro fine. It does mean the requirements are significant enough that organizations with relevant European operations or AI use should understand whether they apply.

A Good Time to Review Your AI Practices

Article 50 has applied since August 2, 2026, so this is no longer simply a future regulatory issue.

For organizations doing business internationally, the practical first step doesn't have to be complicated: understand where and how your organization is using AI.

Identify customer-facing AI experiences. Review how AI-generated content is created and approved. Understand where those experiences and outputs are being used. Then determine whether additional transparency, labeling or governance measures may be required.

As AI becomes a larger part of the digital experience, organizations that build these practices into their normal digital governance processes will be better prepared as AI technology — and the rules surrounding it — continue to evolve.

Frequently Asked Questions About the EU AI Act and Article 50

Does the EU AI Act apply to U.S. companies?
It can. A company does not necessarily need an office or other physical presence in the EU. The Act can apply to companies outside the EU when they provide AI systems in the EU or when the output of their AI systems is used within the EU.

Does having customers in the EU automatically mean Article 50 applies to my company?
No. Having an EU customer alone does not determine whether Article 50 applies. Companies need to consider what AI systems they provide or use, what those systems do and where their outputs are used.

Do companies have to disclose when they use AI to write website content or blog posts?
Not in every case. Article 50's text disclosure requirement focuses on AI-generated or manipulated content published to inform the public about matters of public interest. Qualifying content that receives meaningful human review or editorial control and has a person or organization taking editorial responsibility may also be exempt from the labeling requirement.

Do AI chatbots need to tell users they are AI?
Generally, people interacting directly with an AI system need to be informed that they are interacting with AI unless that fact is obvious from the circumstances. The European Commission says this disclosure should occur at the beginning of the interaction.

Do AI-generated images and videos need to be labeled?
Certain AI-generated or manipulated images, audio and video are subject to transparency requirements. Deepfakes generally require a disclosure that people can perceive, while providers of generative AI systems also have requirements related to machine-readable detection of AI-generated or manipulated content.

What should businesses do to prepare for Article 50?
A useful starting point is to inventory where AI is used across customer-facing websites, applications and digital experiences; identify AI-generated content; understand where those systems and outputs are used; and establish appropriate human review and governance processes. Companies should then work with qualified legal counsel to determine which requirements apply to their specific circumstances.

Disclaimer: This article is provided for general informational purposes only and does not constitute legal advice. It reflects our interpretation and understanding of the EU AI Act and related guidance as of the date of publication. AI regulations are complex, and how they apply can vary based on an organization's specific circumstances and use of AI. Organizations should always consult qualified legal counsel with questions about their legal or regulatory obligations.

Sources: European Commission, Transparency obligations under Article 50 of the AI Act | EU AI Act Service Desk, Article 2 — Scope