The AI Mistake REI Never Meant to Make

A conversation with my dad became a reminder that AI isn’t just changing marketing, it’s changing who we trust to make decisions.

This past weekend I had the pleasure of visiting my parents at our family lake house. It was lovely to sit outside on the beautiful July day, watching bumblebees fly by and water ripple across the lake. My parents and I got to talking about life, school, and work, which in marketing always leads to questions about AI these days. Usually I’m answering questions about what an LLM is or how we use AI at our company as my retired parents have no need to think about AI in the corporate world any more.

Surprising to me, my dad (a big outdoors guy) asked if I saw the trouble that REI had gotten into using AI images in their ads last month. I hadn’t, so naturally went into a deep dive to find out. Right away I was thinking, “Really? REI? An outdoor company using AI? That’s a real low.” But what I found out happened was a lot more complicated.

Here’s What Happened:

In mid-June, an REI Instagram ad promoting a Van Rysel bicycle started gaining traction, but not for the reason REI hoped: the bike in the photo had two sets of handlebars, a misrouted chain, and a warped human face. What you might not be expecting to hear is that the original photo was real, taken during an official Van Rysel shoot with cyclist Amity Rockwell. REI later confirmed that Meta had auto-enrolled the company into an AI “personalization” tool called Advantage+ without its knowledge or consent, and that tool altered the photo before it ever reached the public.

REI ran this Instagram ad which Meta’s AI tool altered.

Here’s the part that matters maybe more than the deformed bike: the ad ran for roughly a week before anyone at REI caught it. Meta’s own terms state that advertisers, not Meta, are responsible for reviewing AI output; enrollment happens by default, and turning it off requires active effort most teams don’t know to make preemptively. So while REI didn’t choose to publish an AI generated image, it also didn’t build a system to catch it, and for a brand whose entire value proposition rests on paying close attention, that gap is its own kind of confession.

The Confrontation

As AI permeates more of our marketing, I think it’s important for companies to be diligent and transparent. Because this REI example is not just about an embarrassing week that called for a PR response, it is a real warning sign to others.

First, this cuts directly against the standard REI has publicly set for itself. REI’s own Chief Merchandising Officer, Kristin Shane, in a recent interview, said that trust at REI is “literally all the micro decisions that we make and making sure they’re aligned with our values.” An unreviewed AI alteration of an ad is exactly the kind of small decision Shane was describing. For a co-op with over 26 million members who reportedly hold it to an unusually high standard, not meeting the bar that they set for themselves is unfortunate at best.

Cyclist and photoshoot model Amity Rockwell responds to the ad.

Second, the damage landed on real people and partners, not an abstraction. The altered image didn’t just look strange; it modified a paid vendor’s product and a hired model’s face without their consent. The cyclist in the photoshoot, Amity Rockwell said, “Why are they Al deep frying the images? … And my face along with it?” Plus, Van Rysel had to publicly confirm its original photo was accurate. This oversight created a vendor-trust problem as much as a brand one, and winning back trust is a lot easier said than done.

Third, this isn’t really a story about REI’s judgment; it’s a story about where the AI ad trend is heading industry-wide. REI likely wasn’t the first brand caught off guard by Meta’s push toward default AI enhancements, and we’re getting pushed toward more automation, not less. That matters because a brand’s reputational cost doesn’t wait for intent. Fractl’s 2026 research found that the share of consumers who’d trust a brand less over heavy AI use doubled in one year, climbing from 20% to 40%, and Canva’s 2026 report found that 78% of people would still rather see ads made by humans, even ones that perform worse. If an accidental AI slip carries nearly the same cost as a deliberate one, the real question isn’t whether REI meant to do this. It’s whether any brand can actually opt out of this trend before the trend opts them in first.

Where I’ve Landed (Sort Of)

Part of me wants to defend REI a little; they didn’t choose to share an altered image, and the pile-on, cancel culture environment ignores the larger problem of how quietly these tools get switched on. But I can’t shake Shane’s line about trust being built from micro-decisions; not noticing an AI tool edit your ad for a week is itself a decision.

As I sat back by the lake later that day, I kept thinking about how quickly a simple conversation with my dad had turned into a discussion about AI, brand responsibility, and trust. Maybe that’s the bigger story here. AI is becoming normal enough that everyone is paying attention, but still new enough that everyone is figuring it out together.

What I do know is if a trend can enroll your brand without asking first, that trend probably deserves more scrutiny than most marketing teams are giving it.