Shoppers Stopped Shrugging at AI. Now They're Picking Sides.

A year ago, the most popular opinion about AI-generated marketing was no opinion at all. Back then, 42% of shoppers said they were neither likely nor unlikely to engage with AI-generated marketing content.

Now in 2026, that shrug is gone. Only 22% of shoppers are still sitting on the fence. The rest hopped off, and they didn't all land in the same place. 

The share of shoppers likely to engage with AI content jumped from 25% to 42%. The share unlikely to engage crept up too, from 33% to 36%. 

So what does this mean for your brand? And what does it mean for your emails, now that your list is split in two?

What the Data Actually Shows

Ask shoppers to describe a brand that uses AI, and you'll probably hear more positive words than you did a year ago. Compared with the results of Backstroke’s 2025 survey, more shoppers call these brands "innovative" (about 38%, up 8 points), "aspirational" (about 19%, up 10) and "admirable" (about 16%, up 6).

But the side-eye is getting sharper, too. "Suspicious" climbed to about 35%. "Lazy" barely budged, meaning it's still one of the first words shoppers reach for.

This is polarization. Positive and negative perceptions are rising at the same time, which means your subscribers are more sure of how they feel. They just don't agree with each other. 

Part of that split likely comes down to experience. People who use AI daily tend to view it differently from those who rarely or never use it. Gallup found that daily users are substantially more likely to trust its accuracy. And for the shoppers who don’t use AI, and have mostly seen it in their social feeds, their skepticism isn't coming from nowhere.

Why People Push Back (And Why They're Not Wrong)

If your experience with and outlook on AI are generally positive, you may brush off the concerns of its skeptics. Don’t. There are a lot of them and most of their objections are fair.

• They want to see what they're buying. An AI model wearing your jacket can't tell a shopper how it actually fits, drapes or feels. If an image misrepresents the product, it’s basically a return waiting to happen.
  
  

• They're tired of AI slop. Shoppers have scrolled past thousands of glossy, generic images that have that weird AI-looking sameness. There’s a reason why none of us like that look. NielsenIQ's neuroscience research found consumers tended to find AI-generated ads more "annoying," "boring" and "confusing" even when the ads were well made. Keep in mind that, even if you’re pro-AI, you’re surely also fatigued by the slop-ification of everything from marketing campaigns to restaurant menus.

• They've been fooled before. Social feeds are full of AI-generated videos, fake reviews and too-good-to-be-true posts. Once someone's been burned by a product that looked nothing like its ad, every polished image starts to look like a trick. Pew Research has found most Americans say it's important to tell AI content from human-made content, yet many aren't confident they can. That gap breeds suspicion, and that suspicion lands on brands.

• They care about the people and the planet behind the work. Some shoppers object on behalf of photographers, models and designers whose work they see AI replacing. Others worry about the energy and water it takes to run AI data centers. A 2026 Associated Press-NORC poll found that 53% of Americans are very or extremely concerned about AI’s environmental impact. These are values-focused positions. It is certain that many of your best customers may hold them.
  
  

• They've watched brands get it wrong. Coca-Cola's AI-generated holiday ad remake drew public criticism, and it landed them on a long list of brands called out for AI imagery shoppers simply don’t want to connect with.

At the end of the day, most shoppers aren’t outright rejecting AI as a concept. Rather, they’re rejecting being misled, being served crappy content and being treated like a monolith. And who would want any of those things, anyway?

A Paradigm Shift in How to Think About AI

In our recent webinar, we showed four last-chance-themed hero images and asked viewers a simple question: Which ones are AI-generated?

Most people can't reliably tell anymore. And this indicates that marketers need to change the questions they’ve been asking. Nowadays, "is it AI?" isn’t as useful as it once was. The better question to ask is, "does this image work for this person?"

Don’t take this new line of questioning to generally upend your values and AI ethics. Those are more important than ever. Instead, use this framing to deliver personalized experiences to each audience segment, be they AI-generated or not. If a shopper needs to judge fit, texture or color, show them the real product. If the people in your image are AI-generated, say so, as you’re legally required to.

The point is to transparently use AI when communicating with people who are ok with it—and we can’t say this enough—use it tastefully.

AI for Whom?

Here's where AI earns its keep. It can’t replace your judgment, but it can judge that one creative choice was never going to work for everyone. It can also predict which one will work, instead.

When we tested three hero images for the same email (a lifestyle shot, a product shot and a typography-only design), lifestyle won with 41% of audiences, product won with 40% and typography won with 19%. The winner shifted by gender, age, device, region and income.

We put that same concept to work with Perry Ellis. Half of their subscriber list received a control email, while the other half received the hero image predicted to perform best. The predicted half saw a 17.5x return, +45% revenue per recipient and a +19.5% click-rate lift. This shows that studying subscriber data—and as a result tailoring your emails’ contents toward the interests and behaviors of each segment—is one of AI’s strongest talents.

Just as consumers aren’t a monolith, neither is AI. Therefore, don't make a single use of it for your whole list. Let performance prediction tools like Surge tell you which segments respond to polished AI lifestyle shots and which ones want real product photography. 

Ultimately, you stay in charge of the brand rules, product accuracy and tastemaking, all of which shoppers see woven throughout each send. Blending these human and AI skillsets will ensure both your subscribers and inbox placements will thank you.

Use AI to Get Closer to Customers, Not Alienate Them

The time AI saves shouldn't vanish into more sends. On a recent episode of The Web Talk Show, Backstroke Founder and CEO R. J. Talyor made the case for using newfound extra time to slow down. Learn a new skill. Direct your hours toward writing or designing. Visit your customers in person. No dashboard can tell you what a real conversation does and no generative tool can truly replace your talents.

You can implement this same philosophy at scale, too:

• Ask subscribers their preferences. A one-question poll or a preference center option tells you which camp a subscriber is in. Zero-party data beats guessing, 100% of the time.
  
  

• Watch what subscribers do. Unsubscribes, spam complaints and clicks by-segment show you where AI creative is landing and where it isn't. Use those findings to inform your future sends.
  
  

• Be upfront where it matters. Flag AI-generated people, use real photos for fit and texture and let some emails be plainly, unmistakably human. 

So why pick a side, when you can pick both? Use AI where it makes sense and with subscribers who feel neutral to positive about it. Still send real photography to all, but anchor it to send to folks who are AI averse. Let each email show you which subscriber belongs to which segment, and tailor your creative accordingly.

Curious to predict how subscribers would react to different creative decisions? See how Backstroke works.