Skip to content
On 2 August 2026 a part of the EU AI Act that touches almost every organisation quietly came into effect: the transparency duties. This is the first broadly applicable EU rule on being honest about AI, and it is directly relevant to Irish businesses.
What the rules require
In short, people must be told when they are interacting with an AI system rather than a person, for example a chatbot, and AI-generated or manipulated content, including deepfakes, must be labelled as such. The same date opened the enforcement window for the obligations on general-purpose AI models. Breaches of the transparency duties can draw fines of up to fifteen million euro or three percent of worldwide turnover.
Why this one is different
The heavier, high-risk parts of the AI Act have been deferred to December 2027, so many organisations assume there is nothing to do yet. The transparency duties are the exception: they apply now, and they reach ordinary uses like customer-service chatbots and marketing content, not just exotic high-risk systems.
What to do
It is not complicated, but it does need doing. Check where you use chatbots and make the “you are talking to AI” disclosure clear. Check where you publish AI-generated images, video or audio and label them. Write down how you did both, because being able to show the decision matters as much as making it.
We help organisations meet the transparency duties now and prepare for the high-risk rules coming in 2027. See our guide to the EU AI Act.
Sources: EU AI Act, Article 50 (European Commission) · GPAI model obligations (Latham & Watkins)
This article is general information, not legal advice.
