Visible AI content is costing brands trust at better than a four to one ratio, and brands keep shipping more of it anyway. Klaviyo surveyed more than 8,000 consumers across eight countries in December 2025, with research partner Datalily, and found that 32 percent of people say spotting AI generated content in a brand’s marketing makes them trust that brand less. Only 7 percent say the opposite. The remaining 61 percent land on neutral, which sounds like a safe middle until you notice that neutral is not where trust gets built. It is where nothing happens for the brand that paid for the content.
Marketers are not ignoring the AI content question, they are mostly watching the wrong door. Integral Ad Science’s 2026 Industry Pulse Report, run with YouGov across nearly 300 US digital media experts surveyed in October 2025, found that 83 percent see the rise of AI generated material on social platforms as a significant concern, and a majority name ad adjacency to that content a top challenge for 2026. That is a real risk. It is also a different risk than the one Klaviyo measured. Adjacency is about a brand’s ad showing up next to someone else’s bad AI content. Klaviyo’s number is about a brand’s own AI generated post, email, or product description losing trust directly, with no neighboring content required.
DoubleVerify’s global numbers sit in the gap between those two findings. Its 2026 study, covering 22,000 consumers across 22 markets and more than 2,000 marketers across 21 markets, found that 42 percent of consumers say low quality AI advertising would hurt their opinion of the brand running it, not just the platform showing it. DoubleVerify frames this as quality and context moving perception more than the mere presence of AI. That lines up with Klaviyo’s split: the 7 percent who trust a brand more after noticing AI content likely saw it used carefully, and the 32 percent who trust less likely saw it used badly, or obviously.
Most brand safety spending still points at the adjacency problem, not the authorship problem. Third party verification tools, the kind IAS found 84 percent of respondents planning to lean on for social platforms, are built to keep a brand’s ad away from someone else’s flagged content. None of that tooling checks whether a brand’s own marketing team shipped an obviously synthetic product description or an AI written customer email that reads like one. The budget line for that second problem barely exists yet, even though Klaviyo’s data says it is the one actually moving trust numbers today.
The fix looks more like an editing standard than a new procurement line. Brand teams get more out of reading their last ten AI assisted pieces the way a skeptical customer would than out of adding another verification vendor. Disclosure, specificity, and a human pass before publishing are the practical gap between the 7 percent outcome and the 32 percent outcome in Klaviyo’s data, and none of the three show up on an adjacency scorecard.
Sources referenced in reporting: