This AI-generated quilt has flooded a KY history group. It has the markings of a scam | Opinion
Kelly Woodham of Georgetown and I are both members of a Kentucky-focused Facebook group, and we’ve both seen the same AI-generated image of a quilt.
The group, History of Kentucky, is your run of the mill Facebook group with posts focusing on Kentucky history. Recently, however, it has become overrun with posts promoting an AI-generated image of a quilt, along with links encouraging members to buy it from questionable online store fronts.
This quilt displays images and iconography of Kentucky and is shown being held by a white woman with blonde hair.
I’ve seen countless iterations of this Kentucky quilt, however, all of them share some overlapping themes: horses, cardinals, bourbon barrels, a grainy image of our state seal which is supposed to depict two men shaking hands but you could be excused for thinking it looks like two bland rectangles.
There’s also the text of the Facebook post which accompanies the quilt.
The exact message has changed over time, but the general contours have stayed the same: The woman (perhaps AI-generated herself) shown in the photo says she has an autistic son, who is a graphic designer/who learned quilt design, he has set up an online store/he has started an online business and you can order from here.
Kelly is one of the few people in the History of Kentucky Facebook group who has publicly stated in the group that she bought the quilt, but hasn’t received it.
Tracking information Kelly shared with me shows the quilt left its country of origin, China, in early December, and arrived in the United States Dec. 13, 2024. Since then there’s been no movement.
“I don’t really know what’s going on,” Kelly tells me. “They continue to advertise this quilt, now I’m the only one that seems to be saying I didn’t receive it.”
Is this a scam? Potentially, yes. There is one other person in this group I’ve been able to publicly identify as having actually received the quilt, but messages to them have not gone through Facebook and comments I left tagging the person have gone unanswered.
More than just the likelihood of a scam, this is also a microcosm of how Facebook, as a platform, is putting all their chips in on AI-generated content and how it’s likely going to confuse and take advantage of its real human users.
What is artificial intelligence?
Artificial intelligence encompasses a growing number of facets of our life from language to robotics to games and finance. The way AI is used differs drastically among who is using the tool. In this instance, I’m referring to an AI-generated image (i.e. this quilt).
What does it mean when I say this image is AI-generated? An AI-generated image is made from computer programs that use learning algorithms to make digital images from scratch or modify existing images, according to New York University Libraries. So it’s either real in some sense, using an existing image as a base, or it’s created out of thin air by a prompt in an attempt to generate an image from that.
How do I know this quilt in question is an AI-generated image? Well there are a few tell-tale signs.
For example, there are multiple quilts being posted all across Facebook, often with the same caption text about an autistic son, and in questionable online storefronts. The website USA Gift Mart, for instance, has a number of these quilts for different states in the United States all accompanied by the same woman holding it. I find it hard to believe one person could’ve made so many quilts for so many states.
Then there are plenty of other mistakes you can see in the quilt images. In one image, you can see text on a barrel and I assume the AI meant to put “bourbon” on the barrel but what came out was an amalgamation of Kentucky and bourbon that ultimately reads like an eye-exam test, “KOURBOY.”
In our conversation, Kelly tells me she’s heard of AI but admittedly doesn’t understand a lot of what it means. I certainly don’t blame her as advances in AI in the last five years have skyrocketed.
People becoming aware of AI, on a more mainstream level, and perhaps incorporating it into their everyday lives, in the form of AI images or ChatGPT, feels to me like more of a recent phenomenon that came onto the scene in 2023 and really went into hyperdrive last year, even though the work that’s gone into the AI has been happening behind the scenes for longer.
Facebook and AI
Facebook, and its parent company Meta, have been expanding and experimenting with AI-everything on the platform in very public ways. Facebook has long used machine learning, a sort of AI subset, on things like ad targeting and their Facebook feed algorithms.
But now you can use Meta AI to chat with, or create, an AI character, you can ask the AI to make you a to-do list, you can even make a movie. You know, all the things you could do yourself or with other people.
However, with the Pandora’s Box of AI fully open, spammers and scammers are taking advantage of this more widespread technology on Facebook.
An August 2024 study from the Shorenstein Center on Media, Politics and Public Policy at the Harvard Kennedy School studied 125 Facebook pages that posted at least 50 AI-generated images and classified those into spam, scam and other creator categories.
The study found these images received hundreds of millions of views and that Facebook users were shown AI-generated images in their feed even if they don’t follow the pages posting the images.
Spam pages use clickbait tactics to direct users off-platform and Scam Pages attempt to sell non-existent products or obtain users’ personal details.
Additionally, and this mirrors with what I’ve seen, the study found that pages sharing AI-generated images often used “repeated caption text across Pages–even Pages that were seemingly unconnected.”
If you search “My son is autistic, he learned quilt design,” one of the variations for this quilt caption, on Facebook then you will be greeted with quilts for different states in different history-themed Facebook groups, and elsewhere, across the platform.
Meta, by its own policy, should be labeling these images as being AI-generated, but none of them are. I could endlessly scroll quilts from sea to shining sea across America and not come across a single disclosure about these being AI-generated images.
“If I thought it was AI created, I would have had suspicions from the beginning,” Kelly said.
And those off-platform sites people are encouraged to visit to buy this quilt? There are at least three, from what I’ve found, that all share the same web domain, the same physical address in San Francisco, California and the same phone number.
If you Google that address you find a listing for the Better Business Bureau for a business named Shopbase which carries an F rating for failure to respond to 57 complaints filed against them.
And if you call the phone number you’re directed to a robotic voice that lets you know they’re experiencing high call volume and will email you a form to fill out. I filled out the form, but I doubt I’ll ever hear back, and I doubt Kelly will ever get much of an answer or that quilt.
Social media without the people
You may be thinking, “This isn’t anything new, online scams have been happening since the advent of the Internet.”
That’s true, but at least in the earlier days of online scams there was a real person on the other end of the screen trying to scam you; with the rise of AI that’s probably going to become more and more unlikely.
After the quilt she ordered still hadn’t arrived, Kelly reached out to the woman who posted the quilt in the Facebook group, Ashley Elliot, and another woman who often reshares the quilt, Maria Castillo Ruiz.
Both of them are admins of the History of Kentucky group and both of them are, quite likely, not real individuals. They’re either fake accounts, automated bots who keep promoting this quilt, some kind of AI or maybe these were real people with these accounts that have since been taken over.
There’s a lot of possibilities in what these accounts could be, but here’s what I can say for certain: both show signs of being bots and/or fake accounts. First, and foremost, both of them are set up on Facebook as pages and not profiles, Ashley can be found pitching different versions of the quilt across Facebook and Maria largely reshares what Ashley posts.
The last time Maria posted in the group was a picture of a Kentucky Fried Chicken restaurant that she says, and the text in the photo claims, was in Laurel County, Kentucky. As commenters noted that wasn’t true and the actual location, it turns out, was in Austin, Texas.
The original photograph was taken in 1960 by Neal Douglass and is part of a Texas history collection hosted by the University of North Texas.
To cover all my bases, I have reached out to both Ashley and Maria for comment.
All of this leaves me wondering: who asked for this? Setting the scammy portion aside, what is the purpose of bots, of AI, when interacting with people on this level?
Sure they can find an image online that real people will react to, engage with and comment their memories of, but these bots, this AI, it has no shared history with us, it has no idea the meaning these images carry to certain people.
It’s just empty.
This story was originally published January 30, 2025 at 10:17 AM.
