The European Act on Artificial Intelligence came into force on Sunday 2 August 2026. What obligations does it impose on creators who use AI in their work, including teachers and trainers? What are the risks involved if you fail to comply with this new regulation? In this article, I provide an overview of the situation and suggest some resources to help you implement the requirements.

The European Act on Artificial Intelligence: a commitment to transparency

In December 2023, I proposed a label to identify articles written without the use of artificial intelligence. In the interests of transparency: I felt it was only right to inform my readers about the use of AI, whether in writing articles or creating media – photos or videos – etc. This suggestion actually sparked quite a lot of discussion on LinkedIn after I posted this article on that social media platform. Other authors, such as Jacques Rodet, whom I had the pleasure of to invite onto our podcast, have proposed icons to indicate the extent to which generative AI is used in media production. The European Union, for its part, has introduced legislation on this matter and, from 2 August 2026, the European Act on Artificial Intelligence will require you to provide a certain level of transparency regarding your use of AI.

Article 50 of the Digital Omnibus imposes certain rules on creators of content produced with the aid of artificial intelligence. This article does not concern media created entirely or with the aid of AI after 2 August 2026. Everything that has been created before that date is excluded from these regulations.

The obligations under the European Act on Artificial Intelligence

Article 50 imposes certain obligations on content creators:

Mapping of your content created or modified by AI after 2 August 2026

You will need to list and map all your content that has been modified or created by artificial intelligence:

• text content : articles, newsletters and other texts created or edited using ChatGPT, Claude, Copilot, etc.
• audio content : artificial voice-overs, speech synthesis, etc.
• image and video content : photos or videos created or edited using tools such as Flexclip or Wall-e, etc.
• customer relations and conversational tools : your website’s chatbot, automated messaging systems or virtual support agents.

Use of the icons from the European Visual Identity Guide

The European Union provides a set of icons to help you inform your audience about the use of artificial intelligence in your creations. They are available in two formats – PNG and SVG – and can be downloaded for free from this page.

This European Commission page also explains when and how often you will need to display the icon in a video or include the statement ‘created with AI’ in audio content.

Labelling exemptions

There are two main types of labelling exemption:

• Artistic or parodic content : if the content is intended as a parody, labelling is not compulsory
• Content reviewed by one person : if a text has been fully reviewed by a human, it may be exempt from labelling, but please note: a person must be identified as the publisher responsible for this content (just like the leaflets handed out in public places). And it must be an ‘in-depth examination’: A simple spelling correction is not regarded as proofreading by a human.

Code of Practice on Transparency

Since June 2026, you have been able to sign a ‘Code of Practice on Transparency’. This is interesting, as it involves the ‘ presumption of conformity“ . In other words, your company will be deemed to be in compliance with Article 50 of the European Act on Artificial Intelligence.”

You can download here Code of Practice on Transparency of AI-generated Content in PDF format

Penalties for non-compliance with the European Act on Artificial Intelligence

As is often the case with European regulations, the penalties are financial in nature:

• Up to 15 million euros
• A maximum of 3 per cent of the company’s global turnover

For a training centre or a self-employed trainer, this could spell financial disaster.

Conclusion: the European Act on Artificial Intelligence promoting transparency

As always, some companies will continue to ignore Article 50 of the European Act on Artificial Intelligence. As always, some will complain about these new obligations and the new administrative burden they place on businesses.

As far as I’m concerned, this transparency meets a need on the part of the public. We already live in a world saturated with disinformation and – I’m going to use an Anglicism here – ‘misinformation’. Sometimes this is deliberate, sometimes unintentional, such as when you share a friend’s post on social media without checking it first.

I believe that transparency will help honest businesses build closer relationships with their audience. For a training centre, trust is an essential foundation. Being transparent about one’s use of AI is part of maintaining that relationship of trust. On that note, the main image for this article was created using Canva for Work’s artificial intelligence. The one featuring the chatbot was created using Canva for Work’s ‘Mockup’ feature, based on a photo of a chatbot superimposed onto an image of a person looking at a smartphone. It only took me two minutes to point this out to you. But those two minutes bring me closer to you by letting you know exactly what you’re seeing on this page. What do you think?

Did you enjoy this article? Subscribe to our newsletter Enjoy this article? and don't miss out on any more publications! It's free and guaranteed spam-free!