Why AI ads make consumers question product quality, according to an expert

No one outside of Coca-Cola knows what those ads actually did to the business, so I won’t pretend that I do. But the asymmetry is real.

Coke has spent a century making sure their product is not in question. People already know what’s in the can. So a bad ad gets judged as a bad ad, and the conversation stays there.

A newer brand doesn’t have the built-in legacy, and the associated resilience that brings. Most people will experience their product for the first time through their ad. So, when it looks like it was generated thoughtlessly with AI, it makes you wonder how much thought the company is putting into everything else. It feels careless, and that’s an impression that can make or break your business when you don’t have a hundred-year legacy of quality to draw on, like Coke does.

Most people probably can’t name exactly what’s wrong. There are the explicit tells that we’re all familiar with, like hands and faces morphing, lip-sync misaligning with audio, and text that’s almost a word, but not quite. But these issues are becoming less and less prevalent with model advancement.

That leaves more subtle cues that are more an indictment of the people using the technology than the technology itself. It shows that no one is putting attention and craft into the work. When the art direction, casting, and location feel default, that’s when you start to get something that “looks AI.” It’s when the lighting and wardrobe shift subtly between shots. Or when a camera move isn’t properly motivated by the narrative.

Essentially, it shows that the brand doesn’t care about craft, nor does it care about hiring people with the eye and the experience to make good work.

The artifact hunting phase has a limited shelf life. I am watching gen media models get better literally every week, and the baseline output quality is now astounding.

So, I think we’ll return to judging an ad the way we always did. Is there a powerful idea in it? Could you swap the logo and would anyone notice? Did it make you feel anything? And does the execution elevate the idea, or detract from it?

That’s a harder bar, not an easier one. Right now a lot of work is graded on a curve because the story is whether the technology works. Once that stops being interesting, the ad has to hold up on its own, and most of what I see today would not.

Put it back into the strategy and the craft. Hire the best creative people you can to conceive of incredible ideas, and make the work so stunning people can’t stop talking about it—not about whether it’s AI or not.

Budgets have always been a limiting factor in advertising, and good Creatives know how to come up with ideas that can be beautifully executed with the production techniques they can afford. If that technique is AI, then they’ll develop concepts that will be bolstered by AI’s strengths, not limited by its weaknesses.

Spend your savings on outstanding talent, and they’ll make your campaigns more impactful, regardless of production method.

I think so, and it’s really about trust more than about AI. Any product ad that looks like cheap AI looks questionably real. And unvetted. The same issue applied years ago to ads that looked obviously photoshopped. It just doesn’t feel trustworthy.

And this is most problematic for products whose promise is built on safety. So, anything you eat, swallow, or subject your body to. Categories where the whole promise is that you’re meticulous.

My husband and I actually experienced this directly when we were shopping for car seats. We looked at one online that was well-regarded, but the product imagery used a truly bizarre-looking and inappropriately sized AI baby. It was so off-putting (and honestly confusing) that we decided to go with a different option.

So, again, it’s about standards of quality applied to your advertising, not about the use of AI, per se. If your brand doesn’t appear to vet your marketing carefully, how will people trust that you’ve vetted your product?

It really runs the gamut. One one hand, there are brands that are so opposed to using AI in their final ads that they aren’t taking advantage of it in the creative process at all. They could be getting to their shoots more quickly and with more confidence if they had used AI to pre-visualize the concept and pressure test different elements in pre-pro. And gen media is invaluable now for fixing and polishing footage in post. Not to mention VFX.

On the other hand, there are brands who are fully embracing AI in their advertising, and not holding it to the same standards as their earlier work. The use of AI actors and models is a good example. To date, the performances have been frankly uncanny, and not on the same level as a human. This is rapidly changing, but the point is that people need to be asking whether the output is good, not just good for AI.

Seven things I would want answered before it ships.

1. Did someone with real creative authority own this? Not approve it. Own it.

2. Would you have made this if AI didn’t exist? If you’re only excited about it because the production was cheap, that shows.

3. Have you shown it to at least 5 people without providing any context? If their reaction is about how it was made versus the idea itself, you need to rethink this ad.

4. Have you refined the look and feel to be unmistakably yours? Swap in a competitor's logo. If it still works, you’ve made an ad for the category—and one that probably “looks AI.”

5. Is there anything in your ad that could be misinterpreted as commentary on the evolving role of AI in the workforce? Don’t add fuel to the fire.

6. Does your product look the same in every frame? Label, color, proportions. Don’t let your standards slip on this!

7. If someone screenshots one frame to make fun of you, which frame is it? There’s always one. Find it before the internet does.

If you can get through all seven of these, I don’t think anyone will spend much thought on how it was made.