What Marketing Teams Must Record About AI Images

0
- Advertisement -

UK marketing teams use AI image generators on a weekly basis now, if not daily. Across Brighton and beyond, they’ve become the default for campaign visuals, social posts, product mockups and ad banners because they’re fast and cheap compared to booking a photographer. The problem is that teams are pumping out AI-generated or heavily edited images with absolutely no documentation about how those visuals were made, what inputs went into them, or who gave the final sign-off.

There’s no UK law requiring an “AI-generated” label on an ad, and the CAP Code has no AI-specific rule. The ASA has been clear, though, that existing rules on misleading advertising apply regardless of how an image was created, and advertisers will need to back up what their visuals show. So really, this comes down to record-keeping. What actually needs to go on file, and why are so many small teams falling behind on something that could hit them hard during a complaint investigation?

What Should Go on File for Every AI Image

Every campaign will have its own quirks, but there are records that should be standard for any team working with AI imagery. The origin matters more than anything else. Was the image fully generated from a prompt, partially edited using AI tools, or composited from multiple outputs? A fully generated landscape carries different risk to an AI-altered photograph of a real product, and that distinction becomes critical if the ASA opens an investigation.

After that, log the source material and rights. If a real photograph served as a reference or base layer, you’ll need a record of where it came from and whether the licence actually covers AI manipulation, because a surprising number of stock library terms don’t allow it. Then there’s consent, which matters whenever the output looks enough like a real person that it could be mistaken for one. Documentation proving consent was given, including how long that consent applies, will save a lot of trouble later. Keep an approval trail too, confirming that a specific person reviewed the image for accuracy and compliance before publication.

Where Brighton Teams Are Most Exposed

Brighton’s creative and digital sectors have grown faster than those in most UK cities. According to the Brighton and Hove City Council’s Economic Plan for 2024 to 2027, digital and creative jobs in the city have grown by over 35% since 2015. What that means in practice is a large concentration of small agencies, freelancers, independent brands and charities all picking up AI tools quickly, often without a dedicated compliance function or a legal team checking assets before they go live.

Inside these organisations, AI adoption has run well ahead of internal process, and it only takes one complaint to make that obvious. An approval trail won’t help much if the evidence is scattered across inboxes, a shared drive, Slack threads and four copies of the same file with no clear labelling. Platforms featured among the best DAMs are built to keep provenance, usage rights, consent records and approval history attached to the image itself, so the metadata travels with the file wherever it ends up. Teams with EU-facing audiences should also note that the EU AI Act’s transparency obligations took effect from August 2026, which added fresh disclosure requirements for AI-generated content on top of existing UK rules.

How to Build the Habit Before the Rules Catch Up

For most small marketing teams, the actual risk isn’t a deliberate attempt to mislead a customer with an AI image. It’s the missing paperwork when they need to prove the opposite. When the ASA investigates a complaint, the advertiser has to provide evidence that the image was accurate and not misleading, that it was properly approved before it went out, and that the team can account for how it was produced. Without a paper trail, that becomes extremely difficult even when the image itself is completely innocent.

This doesn’t require a massive overhaul, either. Even a basic internal log tracking what was generated, what source material fed into it, what rights were cleared and who approved the final version will put a team in a far stronger position if regulators tighten the rules. Brighton teams tend to move quickly when a change makes obvious sense, and this one does. Build the habit now, because patching together records after a complaint has already landed will cost far more in time, stress and legal fees than keeping them as you go.

- Advertisement -

LEAVE A REPLY

Please enter your comment!
Please enter your name here