What: A study of 3,000 U.S. consumers found AI-generated video ads underperformed human-made ones by 14% on short-term sales potential and 17% on long-term brand equity, even though consumers could not reliably tell them apart.

Why it is important: It gives retail marketing leaders an evidence-based framework for where AI creative works (conventional, product-driven work) and where human judgment is still required (emotional nuance, brand storytelling).

A study co-led with Ipsos tested 20 video ads for major brands, half made by humans and half entirely by AI from the same creative briefs. Consumers could not reliably distinguish the two, but the AI versions performed measurably worse commercially. The Chewy example illustrates why: the human-made spot built a precise emotional device around a dog aging over the years, while the AI version told a more generic, less specific story.

AI performed best on conventional, product-driven briefs with clear structures, as seen with Febreze and Herbal Essences, where AI versions matched human ones on sales potential. It performed worst on briefs requiring emotional specificity or narrative surprise, such as a Fiat spot that lost the human version's humor. The recommendation is to stop measuring AI creative by cost or speed and instead track its actual effect on persuasion, memory, and brand equity, scaling only what performs as well or better than human-made work.

IADS Notes: Earlier coverage has tracked AI's expanding but uneven role across the marketing function. Journal du Net's July 2026 reporting framed AI-generated product visuals as a measurable commercial lever tied directly to conversion and return rates, treating output quality as a performance question rather than a production shortcut — a framing that anticipates this study's insistence on measuring efficacy over cost. The regulatory dimension has also been building: Reuters reported in June 2026 that Eurocommerce lobbied to exempt non-misleading AI-generated retail advertising from the EU AI Act's disclosure requirements, while Inside Retail's coverage the same month noted that Korea moved in the opposite direction, mandating disclosure of AI-generated content and endorsers to protect consumer trust. Separately, BCG's June 2026 analysis described a parallel shift at the delivery layer, with AI agents increasingly assembling personalized offers and micro-journeys in real time rather than following fixed campaign structures — a reminder that the effectiveness question this study raises about *what* AI creates sits alongside a broader transformation in *how* that creative gets deployed to customers.

AI-generated ads perform worse than human-made ones, even when customers can't tell them apart