On December 6, 2025, McDonald’s Netherlands posted a 45-second holiday spot built entirely from generative AI clips. A Christmas tree topples in a living room, carolers struggle through a blizzard, Santa causes a traffic jam. By December 9, the video was gone. Viewers had called it creepy, cynical and, in the phrase that has become the internet’s verdict on an entire genre, “AI slop.” The production company defended the work, pointing out that ten people had spent five weeks on it. McDonald’s quietly pulled the spot anyway and described the episode as a learning experience.
That would be a footnote if it were only about one burger chain’s bad week. It isn’t. Nearly a year later, the reaction to AI in advertising has hardened from scattered complaints into something measurable. The 2026 Bentley University-Gallup Business in Society survey found that 79% of Americans had seen ads in the previous month that looked AI-generated. Only 19% view businesses’ use of AI in advertising positively, while 49% view it negatively. Among adults under 30, the negative number climbs to 66%. Those are the customers brands will be courting for the next forty years.
Why AI in Advertising Feels Fake to Consumers
The backlash is not limited to global brands with holiday budgets. It is landing hardest on the businesses that adopted these tools for the most understandable reason: they couldn’t afford anything else.
This spring, Myer’s Bagels posted promotional images that blended real photos with AI-generated ones. Regular customers noticed, the criticism spread across social media and Reddit, and the shop removed the posts and apologized. In September, a Michigan orchard took down AI-generated signage on its own farm after customer complaints. Restaurants have become the front line. As Fortune observed, many customers would rather see an imperfect photo of a meal that actually existed than a flawless image of one that never did.
That distinction explains a great deal. A generated photo of a cheeseburger is not merely an aesthetic choice; it is a promise the kitchen never made. When the plate arrives looking nothing like the picture, the ad hasn’t just failed. It has told the customer something untrue. Gallup’s data points in the same direction. Sixty-two percent of Americans consider it unacceptable for a business to use AI to create people or voices, which suggests the line consumers draw is less about technology and more about deception.
The Hidden Cost of AI in Advertising: Ads That Look Fine and Sell Less
The more unsettling research comes from a place where nobody was complaining at all. Adam Peruta, an associate professor at Syracuse University’s Newhouse School, worked with the market research firm Ipsos to stage a head-to-head contest. The team took ten real human-made commercials from brands including Cheerios, Febreze, Chewy and Visa, all produced before 2021 so no AI could have touched them. Peruta used AI to reverse-engineer the strategic brief behind each one, then had AI build a fully generated counterpart from that same brief. Three thousand U.S. consumers watched the results.
Most viewers couldn’t reliably tell which ads were which. Yet according to the Syracuse summary of the study, the human-made ads outperformed anyway. Using Ipsos’s sales-validated scoring, human work was 14% stronger on predicted short-term sales impact and 17% stronger on long-term brand equity. It’s worth noting these are predictive measures that Ipsos has validated against real-world outcomes, not cash register receipts from a live campaign. Still, they are the same metrics major advertisers use to decide which spots to air.
Writing in the Harvard Business Review, Peruta framed the real danger plainly. The problem isn’t creative that is obviously bad. It is creative that looks good enough to approve in a conference room and then quietly underperforms in the market. Because AI makes production cheap, the bar for “acceptable” drops, and nobody notices the gap until the quarterly numbers come in. Ipsos’s own published report put the reason simply: AI draws on what already exists, so it reproduces the conventions of advertising but struggles to move beyond them. Memorable ads tend to break a rule or two. Averages rarely do.
AI Pros and Consof AI in advertising: What the Research Actually Rewards
None of this means AI is useless in a marketing department, and the evidence is more nuanced than the headlines. In the Ipsos study, AI held its own on functional ads that simply explained product benefits. It fell behind on the emotional work of making someone feel something about a brand.
A separate experiment from researchers at NYU Stern and Emory found that fully AI-generated visual ads actually lifted click-through rates by as much as 19% over ads built by human experts. Then the researchers labeled those same ads as AI-made, and click-through fell 31.5%. The creative was identical. The only thing that changed was whether people knew.
Put the studies side by side and the AI pros and cons come into focus. The advantages are real: speed, lower cost and the ability to produce clear, functional messaging at scale. The drawbacks are just as real: weaker emotional pull, softer brand building and a steep trust penalty the moment audiences discover the shortcut. That last point matters because discovery is increasingly likely. Platforms are adding filters and labels, the IAB’s updated disclosure framework is pushing advertisers toward formal policies, and customers are getting better at spotting the telltale gloss. We’ve written before about how AI fatigue is reshaping what audiences respond to, and the advertising data tells the same story with sharper edges.
Using AI in Advertising Without Losing Authenticity
So what does a sensible marketer do with a tool that saves time but can cost trust? The answer emerging from the research is less about whether to use AI and more about where in the process it belongs.
Gallup found that 75% of Americans are comfortable with businesses using AI to brainstorm ideas or create early drafts, as long as it’s disclosed. That is a meaningful permission slip. AI is very good at the front end of creative work: generating twenty headline variations, roughing out a storyboard, resizing a layout for six platforms or stress-testing a concept before anyone books a photographer. Those are tasks where speed genuinely helps and where nobody outside the building ever sees the machine’s fingerprints.
The finished work is a different matter. The faces, the food, the storefront and the voice should belong to real people and real places. A restaurant can use AI to plan its menu layout and still photograph every dish on an actual plate. A law firm can use it to draft a dozen ad concepts and still put its actual attorneys in the final spot. The judgment about which idea is worth making, and why it will resonate, is still a human job. That judgment is the difference between good design and strategic creative, and it’s rooted in how the human brain actually responds to advertising: emotion first, logic second.
Testing matters too. Peruta’s own advice is to measure whether AI-assisted creative actually delivers on brand strategy rather than assuming that work which looks polished will perform. If an AI-assisted version wins a fair test, use it. If it merely clears the approval meeting, that’s exactly the trap his research describes.
There’s a telling postscript to the holiday ad wars. Coca-Cola faced its own backlash over an AI-generated Christmas ad in 2024. For 2025, it released another one, this time accompanied by a documentary about the people who worked on it. That ad stayed up. The lesson for AI in advertising may be that audiences don’t demand the absence of machines. They want proof that someone with taste, accountability and a reason for being there was holding the wheel. Want to dive into this topic further? Let’s connect.