38%: The AI Trust Ceiling Creators Face This Peak Season

Only 38% of consumers say they'd be comfortable with a creator using AI to generate an entire product review — the exact format peak-season shoppers want most. That tension sits at the center of exclusive Rakuten Advertising research, run with Suzy across the US, UK, Australia, France and Brazil in September 2026, and it lands at the worst possible moment for partner programs squeezed to scale content through Singles' Day and Cyber Week.

Do shoppers already know creators use AI?

Yes, and by a wide margin. The survey found 83% of consumers are somewhat or very aware that creators use AI tools, with 46% noticing it frequently. Awareness peaks in Brazil, at 63% very aware, and bottoms out in Australia at 30%. Gen Z is the most alert cohort: 92% are aware, and more than half spot AI frequently.

The commercial stakes are concrete. Rakuten Advertising reports 23% of consumers have bought a product directly through a creator's link or code. For anyone running CPA (cost-per-action) or rev-share (a percentage of referred revenue) creator deals, that conversion path depends on the audience trusting the recommendation enough to click. Undisclosed AI risks the trust that drives the click.

Where do shoppers draw the line?

Consumers accept AI that supports a creator's work and reject AI that replaces the creator's judgment:

• Translation tops acceptance at 54% comfortable.
• Photo and video editing follows at 46%.
• Fully AI-generated reviews and recommendations drop to 38%.
• AI-generated images of products that don't exist in real life test as the least acceptable use.

The pattern: shoppers don't object to AI as a tool. They object to AI as the voice. A translated caption still carries the creator's judgment. A generated review doesn't.

Does disclosure build trust or destroy it?

The numbers favor disclosure. Across the five markets, 45% of shoppers say AI disclosure would increase their trust in a creator, versus 26% who say it would decrease it. Another 29% report no difference.

The split varies. Trust gains run strongest in Brazil and the US; Australia is nearly even. Millennials are the most receptive group at 52% saying disclosure lifts trust, while boomers are the only cohort where more respondents say disclosure lowers trust than raises it. Even among skeptics, the dominant demand is transparency: half still want to know when AI was used.

Why the double standard on AI?

The asymmetry is control. Some 66% of shoppers have used an AI tool to help them shop, and they trust AI recommendations almost as much as creator recommendations. But when they use AI themselves, they choose the question and judge the answer. Undisclosed creator AI removes that control without the audience noticing. Disclosure hands it back, letting viewers decide how much weight to give the content — which is why it tends to build trust rather than erode it.

What should brands put in creator briefs?

Rakuten Advertising's guidance, drawn from the survey data:

• Keep reviews and recommendations in the creator's own words — the formats shoppers want most and least want automated.
• Use AI for production, not persuasion; translation and editing are the accepted zones.
• Avoid AI-generated products that don't exist, the lowest-testing use case.
• Adapt disclosure by market — the US and Brazil reward clear labels; Australia and older audiences may need an explanation of what AI did and did not do.
• Pair AI labels with familiar formats: 60% say clear disclosure of paid partnerships builds trust, so AI labels can sit naturally alongside #ad — relevant for FTC disclosure compliance in the US market.

Peak season puts brands in a bind: pressure to produce more content, faster, precisely when shoppers scrutinize authenticity hardest. The survey suggests shoppers don't want less AI everywhere — they want clarity about where it fits and confidence the opinion is still the creator's own. Brands that bake disclosure into creator briefs now will enter Cyber Week with audiences that trust what their partners recommend, which the report argues is worth more than extra content.