The EU's new AI transparency rules will directly change how brands build and run advertising, and law firm Mishcon de Reya says UK advertisers should be preparing now, not waiting for enforcement letters.

That is the core warning in the firm's advisory on the EU AI Act's transparency provisions as they apply to advertising content. The headline message for trade readers is blunt: if your campaign pipeline uses AI to generate or manipulate content that consumers see, disclosure duties may already apply to you — even if your headquarters sits in London, not Brussels.

What do the rules actually target?

The transparency regime in the AI Act exists to make sure people know when they are looking at AI-generated or AI-manipulated material rather than a human-made product. In advertising, that touches a broad slice of modern production.

That matters because the categories of affected work are wide:

• synthetic imagery and video in creative assets;
• AI voice-overs and cloned voices;
• digital replicas of real people, including influencers and presenters;
• personalised or dynamically generated ad variants produced by models rather than designers.

The regulatory logic is simple: a consumer watching a spot cannot tell from the pixels alone whether a face, a voice or a scene is real. The rules push that knowledge back onto the brand.

Why should UK brands care?

The obvious question from any UK marketing desk is why a Brussels-written rulebook lands on a London campaign. Mishcon de Reya's answer, spelled out in the title of its advisory itself, is that UK brands cannot opt out.

The EU regime does not stop at the border of the single market. Brands that direct advertising at EU audiences — which for most consumer-facing UK businesses means most of their European media buys — fall within its reach. A campaign run from Shoreditch and served in Paris is still an EU-facing campaign.

There is also a commercial dimension. Agencies, platforms and ad-tech intermediaries increasingly ask their brand clients to certify compliance as a contractual condition. A UK brand that cannot confirm how its AI content is labelled risks friction not with regulators first, but with its own supply chain.

What should advertising teams do now?

The practical to-do list is less dramatic than the legal prose suggests. It starts with knowing what you actually have.

Teams should begin by mapping where generative AI already sits in the creative workflow: which tools, which assets, which approvals. That inventory is the foundation for everything else, because you cannot disclose what you have not catalogued.

From there, the work splits into three tracks:

• Workflow: build a disclosure step into asset approval, so labelling is a production gate rather than a retrofit.
• Contracts: check agency and production agreements for who carries responsibility for AI-generated content and its labelling.
• Evidence: keep records showing when AI was used and how disclosure was made, in case a regulator or platform asks.

Legal review of high-risk assets — anything featuring a synthetic likeness of a real person, for instance — should move earlier in the schedule, not later.

Is this a compliance burden or a competitive edge?

The skeptical read is that this is one more disclosure regime adding cost to already thin production budgets. That is partly true. Labelling, record-keeping and contract rewrites all consume hours.

But there is a second read. Consumers and platforms are already pushing back on undetectable synthetic content, and brands that can cleanly verify the provenance of their creative assets will face fewer takedowns, fewer platform disputes and faster clearances. Transparency infrastructure built for the EU rules will serve that broader purpose.

UK brands that treat this as a paperwork exercise will meet the rules late and pay for it in media friction. Those that build the inventory and the disclosure workflow now will find the obligations manageable — and their supply chains easier to keep.