Brands Are All Using AI in Ads — So Why Can Almost No One Bring Themselves to Ship It?

A while back, I was chatting with a friend who works in branding.

He told me his company's entire marketing workflow had gone fully AI. Copywriting, pitch decks, churning out creative assets at scale — all handed off to AI. Efficiency was genuinely up, and they were saving massive amounts of time.

But toward the end of the conversation, he sighed.

He said, "When it comes to the final ad — the one we actually publish — I still hand it to a real person."

That made me pause. AI can generate practically anything, so why, at the most critical moment — the finished product — does everyone suddenly pull back?

I started looking into what the industry's been doing lately, and it turns out he's not alone. A whole wave of brands are doing the exact same thing: using AI to boost efficiency, but too afraid to ship AI creative to the public.

What Does "Afraid to Ship" Actually Mean?

Let me walk you through a few brands and how they handle it.

Liquid Death sells water, and their brand style is wild. Their marketing lead put it bluntly: AI writing code, running workflows, handling back-office automation — no problem at all. But when it comes to consumer-facing creative content, their stance is: made by hand, first.

Why?

The marketing lead at Columbia Sportswear offered an answer. They're also using AI to automate workflows, but they're highly skeptical of AI ads that consumers see directly. The gist of what he said: judging by social listening data, Gen Z pushes back harder against AI-generated content than any other demographic.

And then he dropped a term — AI slop — which translates to exactly what it sounds like: AI-generated junk. The flood of machine-made images and videos plaguing the internet, the kind you can spot from a mile away.

Their reasoning is simple: consumers want the real thing, not an artificial, plasticky feel.

The founder of Lupine Creative was even blunter. She's worked with clients like Spotify, LG, and Colgate-Palmolive. Her take: AI-generated human figures have a certain "glossy sheen" that immediately looks off. That sense of unreality creates cognitive dissonance in the viewer, and that's a direct hit to brand trust.

She said, "That missing layer of authenticity creates dissonance. And that's a risk not worth taking."

Cautionary Tales, One After Another

You might think these brands are being too conservative. Is AI-generated imagery really that easy to spot?

It really does get spotted. Good lord, consumers' eyes are far sharper than brands give them credit for.

Last August, Guess ran an ad in a fashion magazine. Social media users quickly pointed out that the model looked AI-generated. The controversy blew up fast.

Around the same time, J. Crew's promotional images for their Vans footwear line were also called out for apparently using AI-generated figures. On their Instagram, those images were labeled "digital art."

A year earlier, Coca-Cola's holiday ad sparked its own round of debate over its use of generative AI.

Notice the pattern?

Consumers react far more intensely to "AI pretending to be real" than brands ever expected. It's not about whether your tech looks impressive — it's about whether you're trying to pass off something fake as the real deal.

But What Really Keeps Brands Up at Night Isn't PR

A PR blowup? Issue an apology statement, ride it out a few days, and the heat dies down.

What genuinely keeps brand folks awake is the law.

Here's what's been happening lately.

Disney, Warner Bros. Discovery, and NBCUniversal — three media giants — jointly sued a Chinese AI company called MiniMax, accusing it of copyright infringement. Before that, they had also sued Midjourney, one of the most prominent names in AI image generation.

On another front, The New York Times and a number of other publishers and news organizations are also suing OpenAI.

What does all this mean?

Does the data used to train AI models carry copyright protections? Who owns the copyright on AI-generated content? If a brand uses an AI tool to make an ad, and that tool's underlying data sources turn out to be problematic — does the brand bear liability?

None of these questions have clear answers right now.

Someone who works in agency legal put it well. He said the current situation is basically a game of hot potato among brands, agencies, and AI companies. When something goes wrong, who pays?

Right now there are roughly two approaches: either have the client sign an indemnification agreement and let them shoulder the risk, or make the production house or post-production house hold the bag.

In plain terms: whoever's weakest eats the risk.

So Has Anyone Actually Used AI Successfully?

Yes. But how they did it matters.

Over the summer, Popeyes dropped an AI-generated diss track — a rap callout — on social media. They partnered with a popular AI creator, and the whole thing was clearly labeled "this is AI-made."

But notice one crucial detail: Popeyes ran a full copyright clearance sweep throughout the process, making sure every single element — from the lyrics to the character likenesses — was original and completely clean.

They dared to use it because they cleared every landmine, one by one.

So How Should Brands Actually Think About This?

Here's how I see it.

The value of AI in the marketing creative process doesn't need proving anymore. Boosting efficiency, cutting costs, scaling content generation — those benefits are real and concrete.

But when it comes to human likenesses, when it comes to photorealistic visuals designed to pass as genuine — human craft and authenticity are still the baseline.

The tech is good enough. But consumer trust hasn't caught up.

Run AI in the back office to crunch data and optimize workflows — consumers can't see it and don't care. But slap an AI-generated face on a billboard and pretend it's a real person? Consumers will feel deceived.

Brand trust is built up spoonful by spoonful, and one spectacular faceplant can shatter it overnight.

So the picture you're seeing now is this: brands are still using AI, but they're getting smarter about it. What can stay behind the scenes, stays behind the scenes. What does go public gets clearly labeled.

In other words, AI is a tool, not a body double. Use the tool well and it makes you faster. Get the double wrong, and it takes your trust down with it.

That's a calculation every brand person should be making — especially when you've spent ten or twenty years building trust that a single AI face could evaporate overnight.