EU AI Act 2026: AI labelling requirements for businesses and organizations

In August 2026, AI labelling becomes mandatory across the EU. For your business, however, this is about much more than ticking off a legal checklist: It is a critical opportunity to build genuine trust at a time when consumers have long since lost patience with unreliable AI-generated content.

Image with a computer that says hello on the screen and a yellow background

From fascination to scepticism – in just three years

Since ChatGPT launched at the end of 2022, consumer attitudes towards AI-generated content have shifted rapidly. First came the fascination, when brands such as Levi's announced AI-generated models and Sports Illustrated experimented with AI-written articles. Then came the backlash, as the term “AI slop” gained traction throughout 2024, and both Coca-Cola’s AI Christmas campaign and Toys “R” Us’s Sora-generated film were criticised for lacking authenticity.

We are now entering a third phase: the demand for authenticity. Consumers are not opposed to AI, but they are no longer impressed simply because something has been created using AI. The 2026 Edelman Trust Barometer reveals that we are navigating a landscape of 'Trust Amid Insularity', where consumers have become extremely vigilant and instinctively back away when they cannot verify the authenticity of what they read. Furthermore, the Nielsen Norman Group shows that users quickly recognise a “generic AI tone” – and lose trust in the business when they do.

It is within this landscape that the EU’s new legislation arrives.It is within this landscape that the EU’s new legislation arrives.

What does the EU AI Act say about AI labelling?

The EU AI Act is the world’s first comprehensive AI regulation. While developers of the underlying tools (such as OpenAI or Anthropic) are required to embed invisible metadata in their outputs, the legal responsibility towards your users and customers rests on your business as the deployer (the entity publishing or utilizing the content).

Starting 2 August 2026, you must be transparent about the use of AI across your digital platforms, websites, and campaigns. Labelling is required in the following scenarios:

Images, audio, and video: If you publish images, audio, or video on your digital channels that depict real people, places, or events but are actually AI-generated or manipulated, you must clearly disclose this to the audience.
AI-generated text published to inform the public on matters of public interest must be labelled, except where the content has undergone meaningful human editorial review.

For businesses operating enterprise websites and platforms, this affects far more workflows than most initially realise. Automated product descriptions, AI-generated product imagery, intelligent chatbots, automated customer service responses, newsletters, and personalized ads are all within scope.

Five steps to prepare your business for 2026

Here is how we at Bombayworks recommend preparing your business:

Audit where AI is currently used: Review your digital projects, websites, and marketing channels. From content creation and product imagery to automated emails, SEO copy and customer-facing chatbots.
Establish a clear AI labelling policy: What needs to be labelled on your platforms, how should it be designed, and where should it be placed? A well-crafted policy becomes a natural extension of your transparency – not just a compliance exercise.
Review your external supply chain: Are your digital partners, translators, or SaaS providers using AI in their deliveries? Remember, the legal responsibility ultimately remains with the business that owns the customer relationship and the website.
Turn transparency into positioning: In 2024, Dove updated its "Real Beauty Pledge" with a commitment never to use AI to represent women’s beauty in its campaigns. Being open about how you build your digital experiences will become a strong differentiator once labelling becomes the industry standard.
Measure trust, not just productivity: Many businesses measure AI ROI solely in terms of hours saved or reduced development costs. Far fewer measure the actual impact AI usage has on perceived trust. Start measuring this now – before the regulation forces the issue.

5 steps to prepare your brand for 2026

Here is where we recommend starting:

Audit where AI is currently being used. From content creation and product imagery to automated emails, chatbots and SEO copy, most brands are using AI across more workflows than they realise.
Establish a clear AI labelling policy. What should be labelled, how, and where? A policy that is easy to communicate internally and externally becomes part of your brand transparency — not just a legal compliance exercise.
Review the supply chain. Are your agency, photographer, translator or SaaS provider using AI in the materials they deliver? Responsibility ultimately remains with your brand.
Turn transparency into positioning. When Dove updated its Real Beauty Pledge with a commitment not to use AI to represent women’s beauty in its campaigns, they made a real statement. Not every brand needs to go that far, but a clear position will become a differentiator once AI labelling becomes standard practice.
Measure trust, not just productivity. Many businesses measure AI ROI in terms of hours saved. Far fewer measure the impact AI usage has on perceived brand authenticity. Start measuring now — before regulation forces the issue.

Summary and the road ahead

The EU AI Act is enforcing something consumers have already begun demanding: transparency. Brands that treat AI labelling as a box-ticking exercise will fall behind. Those that use the time leading up to August 2026 to define how, when and why they use AI will gain a clear advantage – legally, strategically and in the eyes of consumers.

Need help mapping where AI is used across your organisation, developing an AI labelling policy, or revising your brand strategy ahead of 2026? Get in touch with us.