If you have an iPhone, you’ll notice that some of the photos it takes have a little watermark in the top left-hand corner. This indicates they are ‘live-photos’ placed in a hyper-short video sandwich (1.5 secs before and after the photo) to help bring the photos to life.

Could marketing tech take a leaf from Apple’s tree and add an ‘AI-Trustmark’ to display ads that have been optimised, versioned, placed or generated using artificial intelligence?
An AI-Trustmark/watermark could help improve advertising transparency at a glance, by helping audiences become more aware of the techniques being used to market to them.
Same goes for chatbots; a simple AI-Trustmark could alert people to the fact that they are interacting with an AI agent as opposed to a human agent.
If the AI-Trustmark included a simple hyperlink, then this could alsoallow audiences to know more about the algorithms used to generate and place the ad or recommendation.

Of course, this is not a full solution to the issues surrounding AI in marketing, but using an AI Trustmark might help foster trust in a marketing system suffering from an unprecedented trust-deficit.








The link behind the trustmark seems more useful than the badge alone: it could explain what AI actually changed. For an ad made with an image editor such as Grok Image 2.0, “AI-assisted” could mean a new background, a retouched product or a fully invented scene—quite different things for a buyer. A short note identifying the generated elements and whether the product is a real photograph would make that disclosure easier to interpret. The label should describe the intervention rather than imply that the advertised claim has been verified.
The Live Photos analogy makes the idea concrete, and the existing comment already names the weakness: a badge without an explanation risks repairing a trust deficit with a label nobody can interpret. We run a small text-to-video product that returns 5, 10 or 15 second clips, and the disclosure buyers actually ask about is not whether AI was involved but which elements are synthetic, because a photographed product placed in a generated scene is a different claim from a wholly invented one. If the trustmark resolved to that breakdown instead of a generic logo, I suspect small vendors would adopt it well before the agencies it is aimed at, since for us the label is a way to answer a question the landing page already has to answer.