When the EU AI Act’s enforcement timeline was revised in May 2026, most businesses breathed a sigh of relief. The high-risk AI system obligations moved to December 2027. Surely there was time to prepare?
There was one obligation that did not move. Article 50 — the AI content transparency requirement — remains on schedule for 2 August 2026. That is weeks away. And most UK and EU businesses are not ready.
What Does Article 50 Actually Require?
Article 50 of the EU AI Act requires that when an AI system interacts with a person, or generates content that a person will see or hear, the organisation deploying that system must disclose that AI is involved. This is not optional. It is not limited to high-risk systems. It applies to any AI system that generates or manipulates content — text, images, audio, or video — that reaches individuals in the EU.
The four key obligations under Article 50 are:
First, AI-generated text published for the purpose of informing the public must be labelled as machine-generated. This covers blog posts, news articles, social media content, and marketing materials produced using AI tools.
Second, AI systems designed to interact with people — including chatbots, virtual assistants, and AI customer service agents — must disclose to users that they are communicating with an AI, unless it is obvious from context.
Third, deep fake video and audio — content that depicts real people saying or doing things they did not say or do — must be clearly labelled as artificially generated or manipulated.
Fourth, AI-generated or AI-manipulated images, video, and audio that is not clearly artistic or satirical must carry a machine-readable disclosure marking.
Who Does This Affect?
If your organisation does any of the following, Article 50 applies to you:
You use AI tools such as ChatGPT, Claude, Gemini, or Jasper to write blog posts, social media updates, newsletters, or website content that reaches EU individuals. You operate an AI chatbot or virtual assistant on your website. Your marketing team uses AI image generation tools such as Midjourney or DALL-E to create promotional images. You produce video content using AI voice synthesis or face-swapping tools.
The obligation applies regardless of where your business is based. If the content reaches individuals in the EU, Article 50 applies.
What Does Compliance Look Like in Practice?
For most businesses, Article 50 compliance requires three things.
The first is a disclosure policy. You need a written internal policy that defines when and how AI-generated content must be labelled. This policy should specify which tools trigger the obligation, what disclosure language to use, and who is responsible for ensuring disclosures appear before content is published.
The second is visible disclosure on content. For AI-generated text and images published publicly, you need a clear, visible statement that the content was produced with AI assistance. This does not need to be prominent — a footer note or an “AI-assisted” label is sufficient — but it must be present.
The third is chatbot disclosure. If you operate any AI-powered chat function on your website or customer service platform, the opening message must identify the system as AI. Something as simple as “You are chatting with an AI assistant” at the start of every conversation is sufficient.
What Are the Penalties?
Enforcement of Article 50 sits with national competent authorities in each EU member state. The UK’s position on Article 50 equivalence is still developing, but UK businesses serving EU customers face EU enforcement jurisdiction directly.
Fines for Article 50 violations are set at up to 1.5% of global annual turnover or €7.5 million, whichever is higher. For a business with £10 million annual revenue, that is a potential fine of £150,000 for what amounts to a missing label on a blog post.
The Three Steps to Take Before August 2026
Step one: Audit your AI content use. List every AI tool your organisation uses to produce content that reaches customers or the public. This includes writing assistants, image generators, chatbots, and video tools.
Step two: Implement disclosure language. Agree on standard disclosure wording and add it to your content templates, website footer, and chatbot opening messages. Your disclosure policy does not need to be long — one page is sufficient.
Step three: Document what you have done. Create a simple record showing that you identified your AI content tools, assessed your Article 50 obligations, and implemented disclosures. This documentation is your evidence of compliance if a regulator asks.
How Wishory Can Help
Wishory’s Article 50 Content Transparency Toolkit (£147) gives you everything you need: a disclosure policy template, standard disclosure language for six content types, a chatbot compliance checklist, and a documentation log to evidence your compliance position.
If you would like a senior practitioner to review your specific situation, book a free 30-minute AI Compliance Review. We will tell you exactly what you need to do before August.