Generative AI.

Generative AI is the umbrella term for AI systems that generate new content: text, images, voices, music, or videos. You enter a description—a prompt—and receive a result that didn’t exist before. Article 50 of the EU AI Regulation refers to AI systems that generate synthetic audio, image, video, or text content. For businesses, generative AI has dual significance: as a time-saving tool and as a subject with its own labeling requirements.

How Generative AI Works

Generative models learn patterns from large amounts of data and use them to create something new. When it comes to text, this is usually powered by an LLM that predicts, step by step, what comes next. Image, audio, and video models learn from images, audio recordings, and videos, respectively.

What Your Business Should Keep in Mind

Determine whether you are a provider or an operator. Regulation (EU) 2024/1689 divides the transparency obligations set forth in Article 50 into two roles. Providers—that is, the manufacturers of AI systems—must, according to paragraph 2, ensure that generated content is machine-readable and identifiable as artificially generated. According to the recitals, a “operator” is anyone who uses an AI system in a professional capacity—which includes your business. Under Article 113, the Regulation generally takes effect on August 2, 2026.

Label deepfakes and AI-generated text on topics of public interest. According to Article 50, paragraph 4, you must disclose when images, audio, or videos are deepfakes—that is, when they bear a deceptive resemblance to real people, places, or events. The same applies to texts that inform the public about matters of public interest, unless they have been reviewed by a human and someone bears editorial responsibility. We explain what this means for blog posts, images, and chatbots in “What the New AI Labeling Requirement Means for Your Website.”

Let visitors know when they are talking to a chatbot. Article 50, paragraph 1, requires that people be informed that they are interacting with an AI system, unless this is obvious. This obligation applies to the system provider, but a clear notice in the chat window never hurts.

Don’t mass-produce pages without added value. In February 2023, Google made it clear that it evaluates content based on its quality, regardless of how it was created. However, Google’s spam guidelines explicitly classify the use of generative AI to create many pages with no added value for users as abuse. In our projects, we’ve seen that AI-generated drafts only become pages that stand out from the rest through our own experience and real-world examples from actual operations.

Frequently Asked Questions About Generative AI

Do I have to label every AI-generated text on my website? According to the wording of Article 50, not automatically. The requirement for texts applies to publications on matters of public interest and does not apply if the content is reviewed by a human with editorial responsibility. If in doubt, consult a legal advisor to clarify your specific situation.

Is ChatGPT generative AI? Yes. ChatGPT generates text based on a language model and is therefore a form of generative AI.

Where is generative AI most useful in a business setting to start with? For drafts that a human would review anyway: answers to standard inquiries, variations on ad copy, and image ideas for a briefing.

If you want to use generative AI in a way that saves time and stays legally compliant, check out our approach to AI for businesses.