Your buyer has started checking the ad against the product. Many have already been burned.
A Talker Research study for Auros, published October 16, 2026, surveyed 4,000 adults in the U.S., U.K., and Australia. Among Americans, 43% said they have felt duped by a product advertised with AI, and 53% said they do not trust AI-advertised products to be described accurately. Forty-four percent said they go out of their way to buy from brands that do not use AI in their advertising. When shown real, AI-enhanced, and fully generated product images, only about one in five correctly spotted real products that had been heavily enhanced.
Canva's 2026 State of Marketing and AI report, fielded by The Harris Poll across 1,415 marketing leaders and 3,547 consumers in seven markets and presented October 15, shows the same gap from the other side of the table. Ninety-seven percent of marketing leaders use AI in daily creative work. Sixty-five percent of consumers say AI-generated ads are so obvious they are laughable, and 78% would rather see ads made by people. Yet 68% do not mind AI in an ad when the ad is helpful or relevant.
Read together, the two studies say the problem is not AI. It is claims nobody stands behind.
The trust gap is a revenue line
Buyers are happy to use AI themselves. In the Auros study, 53% of Americans said AI has helped them shop, mostly to find a better price, read reviews, or find alternatives. The same people distrust AI when it is pointed at them in an ad.
That split matters for every leader scaling AI creative. A polished image that oversells the product does not stay inside the ad. It shows up as returns, refund requests, sales calls that open with "that is not what I saw," and reviews that say so in public. B2B is not exempt: the same doubt attaches to generated customer photos, dressed-up product screens, and results claims on a case study page. The cost lands outside the creative budget, which is why it rarely gets counted.
Put a named human behind every claim
Speed is not the issue. Ownership is. Before AI-made or AI-edited creative ships, someone accountable should be able to answer three questions:
- Does the product in the image look like the product that arrives?
- Can we prove every claim in the copy with something we would show a customer?
- Would we be comfortable telling the buyer how this asset was made?
If the answer to any of them is no, the asset waits. That review does not have to slow everything down. Sort assets by trust risk. Product imagery, pricing, results claims, and anything showing a real person get a named reviewer. Background textures and resized variants do not.
Make the landing page the receipt
The ad makes a promise. The landing page is where the buyer checks it.
- Show the real product, in real use, close to the claim the ad made.
- Put the proof beside the promise: specifications, customer results, and reviews you did not write.
- Say plainly where AI helped when a buyer would reasonably want to know.
When the page confirms the ad, the click turns into trust. When it contradicts the ad, the click becomes a story the buyer tells someone else.
This week
Pull your ten highest-spend ads from the last 30 days. Open each one beside its landing page and the product itself. Mark every image that was generated or heavily edited and every claim without a source. Give each one an owner, fix the ones that overpromise, and add a named human approval step for new creative in those categories before the next campaign launches.
AI can make the ad in minutes. A person still has to stand behind it.
