The EU AI Act transparency rules take effect on 2 August 2026
From 2 August 2026, Article 50 of the EU AI Act requires you to tell people when they are dealing with an AI system and to label AI-generated content. The high-risk deadlines were pushed back. This one was not.
On 2 August 2026, just over a month from now, the transparency obligations in Article 50 of the EU AI Act start to apply across the EU. If you run a chatbot, generate marketing copy or images with AI, or use it to talk to customers, this is the deadline that touches you first.
What Article 50 actually requires
- Tell people when they are interacting with an AI system, such as a chatbot or voice assistant, unless that is already obvious from the context.
- Mark AI-generated or AI-manipulated content, including text, images, audio and video such as deepfakes, as artificially generated in a machine-readable way.
- Inform the people exposed to an emotion-recognition or biometric-categorisation system that it is in use.
Why this date still stands
The digital omnibus, agreed politically on 7 May 2026 and since adopted by the Parliament on 16 June 2026 and the Council on 29 June 2026, postponed the AI Act high-risk obligations. Stand-alone high-risk systems in Annex III now have until 2 December 2027, and AI built into regulated products under Annex I until 2 August 2028. The Article 50 transparency rules were left on their original date of 2 August 2026, with one carve-out: the machine-readable marking obligation in Article 50(2) is delayed to 2 December 2026 for systems already placed on the market before 2 August 2026. The start date itself has not moved.
What to do before then
- Add a clear "you are talking to an AI" line to any chatbot or assistant you run.
- Label anything your tools generate, copy, images or audio, as AI-generated.
- Write the notice once and reuse it everywhere it applies.
You do not need a lawyer to draft the basic disclosure. Our free transparency-notice generator writes one you can paste into a chatbot, a footer or a content credit, tied to the categories Article 50 names.
Put it to work on your own case
The free tool below turns this into a result for your situation, in your browser, with no signup.
Generate an AI transparency noticeSources
More updates
The EU AI Act high-risk deadline is no longer 2 August 2026
A May 2026 political agreement pushed the EU AI Act high-risk obligations back by more than a year. Stand-alone high-risk systems now have until 2 December 2027, and AI inside regulated products until 2 August 2028.
The old ISO 27001:2013 certificate is dead: the transition window closed on 31 October 2025
The three-year window to move from ISO 27001:2013 to the 2022 edition closed on 31 October 2025, so a 2013 certificate held into 2026 is no longer valid. The 2022 standard reworked Annex A from 114 controls to 93 in four categories and added modern ones like threat intelligence, cloud security and secure coding. If your certificate lapsed, recertifying means a fresh initial audit rather than the lighter transition assessment, which changes the budget line for anyone pricing readiness this year.
Stripe now sells its own merchant of record, Managed Payments, at 3.5% on top of processing
Stripe now offers Managed Payments, its own merchant-of-record option, introduced in a 25 February 2026 release. Stripe becomes the seller of record for your digital products and takes on indirect tax (VAT, sales tax, GST) in more than 80 countries, plus fraud, disputes and buyer support, for a fee of 3.5% per successful transaction on top of the normal processing fee. That puts it in the same rough band as Paddle and Lemon Squeezy, and it can be switched on per transaction without moving to a separate platform.