Can people tell your ad was made with AI?

If you are thinking about an AI ad, this is probably the worry sitting under all the others. You do not want your customers to scroll past and think “cheap” or, worse, “scam”. That worry is reasonable. Here is what the research actually says, and what we do about it.

What did the research find?

NIQ (NielsenIQ), 2024. NIQ tested AI-made ads with surveys, eye tracking and EEG, which reads brain activity. Viewers picked out most of the AI ads on instinct and described them as more “annoying”, “boring” and “confusing” than conventionally made ads. Even the most polished AI ad they tested left weaker memory activity in the brain. The scenes that looked strange or unrealistic did worst of all: they took the most effort to watch and were remembered least.

There was a quieter finding too. The AI ads did reinforce what people already knew about the brand. They just risked a “negative halo” that could cancel the benefit out.

Tinuiti survey, reported in February 2025. In this US survey, 40% of Gen Z and 42% of baby boomers said brands that use AI in their ads seem like scams. Yet 61% said they would still buy from a brand whose ads include AI-made images, video or text. People are suspicious, but they have not walked away.

Washington State University, 2024. Researchers surveyed more than 1,000 US adults across eight product and service categories. Simply mentioning “artificial intelligence” in a product description lowered people’s intent to buy, because it lowered emotional trust. The effect was stronger for high-risk purchases, like expensive electronics, medical devices or financial services. The lead author’s advice was to describe the features and benefits and skip the AI buzzwords.

One honest caveat: the Tinuiti and Washington State studies surveyed people in the US, and NIQ has not published where its viewers were. Indian audiences may react differently.

So what actually gives an AI ad away?

Read the findings together and a pattern shows up. People react badly to:

• Faces and bodies that are almost real but not quite: odd hands, dead eyes, mouths that do not match the words
• A product that looks a little off: wrong label, wrong colour, a cap that changes shape between shots
• Things that pretend to be real when they are not: a “customer” who does not exist, a review nobody gave
• Ads with no idea behind them, where the visuals are the whole point

Notice what is missing from that list. Nobody is upset that a whey shaker was lit by software instead of a lighting crew. Obvious stylisation is fine. The trouble starts when an ad tries to pass for something it is not.

What do we do about it?

• The idea comes first. A filmmaker writes and directs every ad before any AI is used. An ad with a clear idea gives people something to remember other than the technique.
• Your pack has to be your pack. We rebuild it from your artwork and photos and check it shot by shot.
• No fake customers. Presenters and characters in our ads are generated, and we never present a generated person as a real customer or a real review.
• We say when a shoot is better. If your ad needs a real person on camera, filming it is the honest choice, and we will tell you that.

Should my ad say “made with AI”?

Two different things are going on here. Shouting “AI-powered!” as a selling point is usually a mistake, which is what the Washington State study suggests. But some AI content now needs a plain disclosure label in India, and that is a different matter: it is about honesty, not marketing. We explain which ads need one in Do AI ads need a label in India?

The short version

People do not hate AI ads. They hate being fooled, and they hate ads with nothing to say.

If you want to see how this looks in practice, watch the ads we have made and judge for yourself.

Frequently asked questions