The short answer

People don't reject ads because AI helped make them. They reject ads that look cheap or use synthetic humans pretending to be real. Only about a quarter of viewers can tell AI work apart, and well-made AI ads perform like well-made traditional ones.

Every brand we talk to asks some version of the same question: will people hate it if they know it's AI? The research from this year gives a clearer answer than the debate on LinkedIn does.

The honest numbers first. In Ipsos and Newhouse's May 2026 study, human-made ads still beat AI ads by 14 to 17 percent on measured effect. Gallup found in August that half of American adults feel negative about AI in advertising, and 62 percent reject synthetic people or voices. Marketers, meanwhile, overestimate how positive young audiences feel about AI ads by nearly two to one, according to the IAB.

Now the part that matters. In the same Ipsos work, only about a quarter of viewers could tell which ads were AI. Kantar reviewed 356 AI-assisted ads and found the same spread of performance as traditional work: good creative principles hold either way, and obvious AI hurts. Coca-Cola's AI holiday films drew two years of loud backlash and still performed in the top tier on measured effect.

The line

Where is the line, exactly?

Look at what got punished. McDonald's Netherlands pulled its AI Christmas ad after four days. Fashion brands were criticised not for using AI, but for hiding synthetic models behind tiny disclosures. Ads that were openly, playfully AI, where the look was the joke, were tolerated.

The pattern is simple. Audiences push back on work that looks cheap and on synthetic humans pretending to be real. They don't push back on a good film. Worlds, products, animals and stylised characters are safe ground. A photoreal fake presenter is not.

For brands

What should a brand do with this?

Judge the film, not the tool. If nobody can tell how it was made, the performance gap closes.

Keep real people real. When a human face carries the message, cast a human. Use new tools for the world around them.

Label honestly. India now requires labels and provenance on realistic synthetic content, Google Ads requires AI labels in India, the EU and New York, and the EU AI Act's disclosure duty took effect in August. Disclosure done properly isn't a risk. Hiding it is.

The line we hear most from clients is "nobody believes these are AI". It isn't a marketing line. According to the evidence, it's the whole effectiveness argument.

Sources: Ipsos with Syracuse University's Newhouse School: AI ads are good enough, and that's the problem (May 2026); Bentley University-Gallup: Americans aren't sold on businesses using AI in advertising (August 2026); IAB and Sonata Insights: The AI Ad Gap Widens (January 2026); Kantar's review of 356 AI-assisted ads, via Research Live (December 2025); India IT Rules amendment on synthetic content (February 2026); Google Ads policy on AI labels; EU AI Act, Article 50: transparency obligations.