The long awaited EU AI Act took effect this week with new transparency rules requiring AI-generated content to be clearly labelled.
The European Union’s (EU) AI Act came into effect this weekend, which mandates businesses to disclose when AI has been used to create certain types of content.
The EU hopes the new rules will help people recognise when they are interacting with AI chatbot or exposed to AI-generated content, so they can make informed decisions and better protect themselves from misinformation or deception.
Article 50 of the AI Act makes in mandatory for all AI-generated content, including publications, must be labelled as such and the EU has created a set of icons for this.
Organisations using AI to generate or change text, images, audio, and video must now add machine-readable marks to help users and detection tools identify AI-generated content.
Under the new rules, deepfakes must also be disclosed, as well as emotion-recognition systems and biometric categorisation tools.
AI-generated text that has not been reviewed or edited by a human must also be clearly labelled when meant for the public. All providers and deployers will have to demonstrate compliance and implementation of the new regulations.
Any companies found to not be compliant could face fines of up to €15m, or 3% of global annual turnover, for companies. While institutions, bodies, and agencies could face a penalty of up to €750k.
These rules apply to any deployer (the creator or publisher) based in the EU, and as well as outside the bloc where they are directing, overseeing, or authorising distribution of AI content into the EU. This includes posting “deepfakes on the globally accessible internet”.
Audiences and customers must also be clearly informed when they are interacting with a chatbot, AI agent, or avatar.
The European Commission (EC) has also issued a Code of Practice on Transparency of AI-generated Content to help support compliance with the new rules and build more trust in the information ecosystem
While transparency obligations will be applicable from 2 August, adherence to the code of practice is voluntary, unlike the obligations under Article 50 (primarily disclosure on AI content), which are written into law.
What this means for marketers
Agencies and advertisers who run campaigns with an EU reach should assume these rules apply, especially if it can be seen via publicly accessible internet – even if the ad is not formally targeted at the EU.
A campaign geo-targeted to the UK, for example, that has only reached the EU unforeseeably (like having been shared in a private text chain) is unlikely to fall under the regulations.
Thomas Ives, co-founder of RAAS LAB, said: “The EU AI Act’s new transparency rules are a wake-up call: uncontrolled AI use is a business risk, not just a regulatory one. However, compliance isn’t the whole story. Even the highest-quality AI creative can undermine a premium brand if it’s served in irrelevant, low-quality or unsuitable environments.
“As the volume of AI-generated content grows, context is the ultimate differentiator. Success won’t come from using AI at the greatest scale, but from using it responsibly. Brands that prioritise transparency, relevance and media quality alongside innovation will be better positioned to build trust, protect brand equity and deliver stronger long-term outcomes.”
These regulations apply to both developers and professional users of AI – covering businesses, technology companies, publishers, and most organisations using automated tools, AI chatbots, or publishing AI-generated material.
Any AI-generated text that is published with the aim of informing the public on a matter of public interest must be clearly labelled – with the exception of anything that has undergone human review and an organisation or person accepts responsibility for it.
AI-created advertising copy is likely to be outside of the scope of this new rule, unless it includes any claims relating to health, sustainability, or consumer safety. Exceptions apply to satirical, fictional, and artistic content.
