AI-generated ads achieve parity but lack emotional pull

Ipsos, in collaboration with Syracuse University's S.I. Newhouse School of Public Communications, published research in May 2026 finding that generative AI video tools can now produce ads most viewers cannot reliably distinguish from human-made work - but human-created ads consistently outperform AI-generated ads on business metrics (Ipsos). The research was subsequently summarized in Mumbrella (June 2026). The study tested 20 ads from 10 brands including Cheerios, Chewy, Febreze, Fiat, H&M, Old Navy, Herbal Essences, Ray-Ban Meta, TurboTax, and Visa, with 3,000 U.S. respondents tested via Ipsos' sales-validated Creative|Spark platform and the MISFITS Mindset framework.

Per the Ipsos report, Professor Adam Peruta of Newhouse used Google's Gemini to reverse-engineer the creative brief from each original human-made ad. That brief was used to develop a shot list, which was fed into OpenAI's Sora 2 to generate the AI counterpart - fully AI-produced spots with no human creative intervention. The study controlled for strategy (same brief) while isolating the human-versus-AI creative execution variable. Ads were shown to respondents in a distracted, high-stimulus environment mimicking real viewing conditions.

Only 25 percent of AI ad viewers were "at least somewhat confident" the spot was AI-made, while 40 percent of all viewers were uncertain either way - indicating viewers largely cannot detect AI advertising in practice. Yet the effectiveness gap materialized on business metrics: human-made ads were 14 percent stronger on short-term sales impact (Creative Effect Index) and 17 percent stronger on long-term brand health (Equity Effect Index), per Ipsos. Viewers still rated human-made work as more eye-catching and imaginative even without identifying it as AI-free.

The gap was largest in ads requiring storytelling or emotional resonance. The Fiat AI ad, briefed to be "energetic, defiant, and confident with a playful, rebellious edge," ranked among the least emotionally stirring in the set. The strongest AI performers were product-driven spots - Febreze and Herbal Essences - where functional structure is well-represented in AI training data. The best-performing pair overall was Cheerios, where a clear human-empathy proposition in the brief ("Cheerios helps keep your heart healthy, so you can be there for the ones who need you most") allowed the AI spot to close the effectiveness gap.

The study captures a snapshot of AI video capability as of mid-2026; model progress is rapid. Ipsos notes 79 percent of consumers surveyed say companies using AI should disclose that use, and the IAB has released disclosure guidelines. Observers should track replication of this framework across non-U.S. markets and higher-funnel brand categories, vendor features that surface provenance or creative intent, and whether the 14-17 percent effectiveness gap narrows as AI video models iterate on narrative and emotional arc.