A recent study by Raptive, the digital media firm behind ad sales for publishers like Stereogum, shows that readers don’t just side-eye AI-written content—they flat-out trust it less. In a survey of 3,000 U.S. adults, Raptive found that when people suspected an article was generated by AI, their trust in the piece dropped by half, and their opinion of nearby ads took a hit too. Even more concerning for brands: consumers were 14% less likely to consider buying a product shown alongside an article they believed was AI-made, regardless of whether it actually was.
Participants were shown a mix of AI and human-written articles about everyday topics like food and travel, each paired with an ad. When told the content was AI-generated, readers docked it across metrics such as trustworthiness and emotional draw—even if the writing was technically fine. The buzzkill effect extended to the ads themselves, which were called less “inspiring,” more “artificial,” and—you guessed it—less “premium.” This psychological spillover, dubbed the “AI stink,” means even good ads can lose their luster when stuck next to suspicious content.
For publishers banking on faster output through machine-generated content, these results serve up a tough reality check. The study spotlighted how vague, unlabeled AI efforts can quietly chip away at a brand’s perception and media ROI—one pixelated paragraph at a time. As AI continues to flood the feed, brands might want to ask not just what their content says, but who (or what) it sounds like.
Full story at Adweek.