Study: AI-Generated Ads Underperform Human-Made Ads in Sales Effectiveness

AI-generated advertisements are becoming harder to tell apart from human-made ones, but they still clearly fall short where it matters most: real business results. That's the conclusion of a study conducted by Ipsos and the S.I. Newhouse School of Public Communications at Syracuse University, published in May 2026. The findings are fueling a broader shift among brands toward authenticity as a selling point.

Credible but not convincing

The Ipsos and Syracuse researchers highlight a distinction they say explains the entire gap. AI-made ads can look professional and credible, yet they fail to translate into emotional engagement from viewers or measurable sales results. The researchers used identical creative briefs, varying only who produced the ad: a human in one case, an AI model in the other.

The human ads and the AI ads started from the same brief - Adam Peruta, S.I. Newhouse School of Public Communications

The gap showed up mainly in tasks that required storytelling, emotional depth, or a distinctive point of view. On simple, product-focused briefs, AI performed relatively well, but it lost ground wherever narrative mattered most.

AI is a powerful tool, but the data shows that the human capacity for storytelling and emotional connection still creates a measurable competitive edge - Ryan Barthelmes, Syracuse University

Consumers growing more wary

The Ipsos and Syracuse findings are echoed by a Kantar study published around the same time, in March 2026. It found that fewer than half of consumers respond positively to AI-generated imagery in ads, while more than half fear AI-made content could be misleading. At the same time, marketers view the technology far more enthusiastically than the audiences watching their campaigns, a gap Kantar describes as a growing perception divide between the industry and its customers.

Kantar argues the problem isn't rejection of the technology itself but a lack of context: audiences don't know what's real and what's generated in a given ad, and that uncertainty is what undermines trust. The firm recommends clearly labeling AI-generated content, especially when it involves depictions of people rather than products alone.

Brands return to authenticity

Some major brands are already shifting their messaging, positioning themselves directly against the wave of mass-produced AI content. In June 2026, Polaroid launched a campaign built around an analog summer theme, leaning on physical, one-of-a-kind photographs. Reddit has spent months running a campaign centered on real people, and in July 2026 The Economist rolled out billboards playing on the contrast between a bot and genuine intelligence.

These campaigns share a common thread: authenticity, long taken for granted, is becoming a deliberate strategic choice and a way to set a brand apart from the flood of mass-produced content. In practice, that means a return to real photographs, real customer and employee stories, and clear labeling of where AI shows up in a brand's communications.

What this means for Polish companies

For Polish marketers, these findings have direct implications for budgets and creative processes. Cutting production costs through generative AI, an appealing prospect especially for smaller companies and agencies, can hurt campaign effectiveness if AI fully replaces the creative work of building narrative and emotion rather than just handling routine production tasks like product photography.

The growing share of consumers expecting AI disclosure in ads also has legal implications. Starting August 2, 2026, chatbots in Poland will be required to tell users they aren't talking to a human, and the broader AI Act imposes new transparency obligations on the makers of the most powerful models. The push for transparency in advertising fits the same regulatory logic, even though labeling AI-generated ad content isn't yet a legal requirement in Poland.

The Ipsos and Syracuse researchers aren't calling for marketers to abandon AI tools, but rather to treat them as support rather than a replacement for human creativity. They note that the MISFITS model, which accounts for a full set of creative factors, boosted creative effectiveness by as much as 20 percent, but only when AI supplemented a team's work instead of replacing it.