It began with the cutest little desserts: chocolate acorns with nut-covered caps that popped up in my seek for Thanksgiving cookies on Pinterest, a web site I go to for inspiration and a few step-by-step directions. There had been additionally chocolate-dipped strawberries that you may flip into little turkeys utilizing pretzel sticks, with marshmallow items for the drumsticks, however the acorns appeared simpler.
“Who wouldn’t love these things?” I thought.
The recipe mentioned you may whip up a batch in 45 minutes. I splurged and purchased my favourite Lindt chocolate bars to make them.
It ended at 1 a.m., after about 5 hours of effort – together with an emergency run to the grocery retailer for a lacking ingredient – with parchment-covered baking sheets unfold throughout my kitchen, holding misshapen globs of chocolate-dunked peanut butter sandwich cookies that didn’t come near resembling acorns. Despite my many makes an attempt at adjusting the temperature of the chocolate and MacGyvering the cookies — I reduce them, stacked them and caught them along with chocolate — they only couldn’t match the acorns within the picture.
I’m normally fairly good at these items (pie crust being a irritating exception). I’m an skilled baker, and this could have been an straightforward challenge, extra meeting than baking. I wasn’t going for gourmand; I wanted a variety of treats for a church occasion and was quick on time.
I lastly needed to admit that irrespective of what number of instances I reread the recipe and tried to observe the directions, it wasn’t going to work. I was stumped.
Then it dawned on me: Could this recipe have been generated by AI? Had I been had?
Even as I had been studying via the directions, I saved questioning how the blogger received the Nutter Butter cookies — that are flat, with a peanut-like curved form — to show into the good rounded form of the acorns. The recipe recommends holding the cookies by the pointier finish “to create the more natural looking” acorn.
OK, I thought. Maybe I want to simply belief the method.
“They’re ridiculously cute, surprisingly simple, and honestly? They’ve become the most photographed dessert at every family gathering I bring them to,” gushed meals blogger Anna Kelly, who described herself as a professionally skilled pastry chef who runs the web site DessertsPro.com. “My sister-in-law now requests them specifically, and I’ve made them at least twelve times since I first stumbled upon the idea three years ago.”
I fell for it. I purchased three packages of the regular-size Nutter Butter cookies and searched a number of shops for the bite-size ones, which had been speculated to be for the caps. Then I seen an issue: The bite-size cookies had been a lot smaller than the ends of the large Nutter Butters. It didn’t look like they might work to make the overhanging acorn cap.
I nonetheless didn’t query the recipe. Instead, I thought I had purchased the mistaken measurement cookies. Was there a medium-size Nutter Butter? Oh, the recipe mentioned Kelly typically makes use of Nilla wafers. That made extra sense, but it surely was getting late, and I wasn’t going again to the grocery retailer.
Thankfully, the chocolate melted prefer it was speculated to, but it surely was fairly skinny, and I was shedding hope that it could coat the cookies thickly sufficient to appear to be the acorns within the picture. I tried popping it within the freezer to thicken it.
The web site DessertsPro.com says Kelly created it to share her favourite tried-and-true recipes with residence cooks. Her “About Us” web page guarantees “clear, simple instructions,” easy-to-find substances, “honest tips and tricks” and “recipes that actually work.”
The drawback is that neither Kelly nor her creations seems to be actual.
I tried to succeed in the creators of the location via a “contact us” e-mail in addition to via the contact type however received no response by my deadline. I additionally tried to contact Kelly via one other meals weblog arrange along with her identify and profile, MuffinIdeas.com. This web site makes use of the identical picture and syrupy textual content that talks about Kelly’s nostalgia for her grandmother’s baking, however the bio is barely completely different: It says she’s not professionally skilled however self-taught by studying numerous baking books by the likes of celebrated cooks like Dorie Greenspan, the very actual writer of “Baking with Dorie,” and David Lebovitz, writer of “Ready for Dessert.”
Emails despatched to each web sites had been returned as undeliverable.
“This site is almost certainly AI,” mentioned Adam Gallagher, who runs the meals weblog Inspired Taste, which he began together with his spouse, Joanne, in 2009.
The Gallaghers have been outspoken in regards to the rising quantity of AI-generated pictures and recipes meant to appear to be real meals blogs which have permeated social media websites like Facebook and Pinterest.
They say that not solely are these faux websites siphoning site visitors that used to go to actual bloggers, AI know-how is scraping websites like theirs and emulating their type to create the computer-generated content material it churns out.
It’s not that synthetic intelligence, as a software, is essentially a foul factor, says Tom Critchlow, government vice chairman for viewers progress at Raptive, an organization that helps bloggers and different creators monetize their content material. It can be one factor, possibly, for an actual chef or meals blogger to make use of some AI to assist refine their writing or get new concepts.
But that’s not what’s taking place right here, Critchlow says.
“These are bad actors generating hundreds of thousands of brand new, fresh domains that have never been seen before, using AI to generate hundreds and thousands of AI-generated recipes, images, et cetera, and covering the websites with ads,” he mentioned.
“Pinterest has been overrun by these spammers,” Critchlow mentioned. “Almost any kind of search term or any kind of exploration on Pinterest in the food and recipe space leads you to these websites, which are just spam top to bottom.”
Even websites with nugatory content material can earn money when individuals click on and scroll via them, simply as I did when attempting to determine what I was doing mistaken.
On common, content material creators earn about $30 to $50 for each 1,000 web page views to their websites, in response to Raptive.
One meals blogger would possibly spend hours creating, testing and capturing a recipe to earn that charge, however AI-generated websites can earn the identical cash for content material created in minutes.
“When you look at the scale and volume of these things, there are whole networks of sites, hundreds of sites, made by a single individual,” Critchlow mentioned, noting that movies on YouTube additionally present how to do that.
Raptive says it kicks creators and web sites out of its community if it discovers that they’re utilizing AI instruments to make spam recipes. This yr alone, the corporate says, it’s blocked almost 600 websites – of about 6,000 in its community – for AI-related points.
When I learn again via my apparently faux recipe, I discover clues I initially missed.
In the introduction, “Anna Kelly” writes that her nephew calls her Aunt Sarah, moderately than Anna. The textual content says the location was created in 2019, but it surely – like MuffinIdeas.com – was copyrighted in 2025. Had I clicked round on a number of the different recipes, I may need seen that a number of the photographs had been suspect: The blueberries in a loaf of sourdough bread are impossibly round purple craters, and the cupcakes or muffins (cuffins?) in Kelly’s profile pic have frosting that magically fades into the cake.
The recipe I tried didn’t have any scores or feedback – one other purple flag, since I usually depend on the cooks who’ve gone earlier than me for suggestions.
“People have a shiny look. It’s almost like everyone’s a little plasticky. That’s a telltale sign,” Gallagher mentioned. “It’s like, everything’s in a little bit of Saran Wrap.”
I discovered the recipe on Pinterest, however the web site hyperlinks to accounts on Instagram, Facebook and X, although these profiles have little to no followers.
“Consumers are being duped by all the platforms,” Gallagher mentioned.
He says platforms like Pinterest aren’t actually motivated to filter out AI spam as a result of they earn money irrespective of the place the recipe comes from, so long as it retains individuals scrolling.
Meta, the mum or dad firm of Facebook and Instagram, didn’t reply to a request for remark by deadline.
Pinterest didn’t supply a direct remark however mentioned it disagrees with the concept it isn’t motivated to assist customers who need to keep away from AI-generated content material. It says it can definitely damage its enterprise if individuals get pissed off by the content material that’s on the location and cease coming.
It additionally says consumer suggestions led the corporate to create an AI filter that customers can activate of their settings, which might help hold computer-generated content material out of their feeds. It added the filter to the Food and Drinks class in mid-December. AI pictures are additionally speculated to be labeled whenever you enlarge them. The chocolate acorn picture I clicked on doesn’t have such a label, although.
Gallagher and Critchlow mentioned they’ve tried the filter and didn’t discover a distinction of their feeds. The firm says the software is new and may get higher over time.
Pinterest additionally pointed me to an audit, carried out by the digital safety reporting web site The Indicator, testing the guarantees of platforms to label AI-generated content material, which was posted by the reporters. The audit discovered that 5 main platforms repeatedly did not label AI-generated content material on their websites, together with content material created utilizing their very own AI instruments.
Pinterest was probably the most profitable at labeling AI pictures, the audit discovered, however nonetheless had successful charge of 55%.
“I want to give Pinterest a little bit of credit, because they were early in introducing a way for users to filter AI content out of their feeds, which I think is something that all platforms should be doing,” mentioned Alexios Mantzarlis, a former Google government who’s co-founder of the Indicator and did the AI testing for the story.
“So Pinterest, I think, is understanding from its users that the stuff sometimes is very annoying, and it appears to be reacting,” Mantzarlis mentioned, although its instruments aren’t nice at flagging AI simply but.
The firm says it doesn’t need to ban AI fully as a result of some individuals discover that it conjures up them.
The Gallaghers suppose there are extra pragmatic causes some social media websites haven’t been aggressive in tackling AI content material.
The extra content material a platform has and the longer a consumer continues to scroll, the extra advertisements it could actually present, Joanne Gallagher mentioned.
Spam websites aren’t the one method AI makes up recipes. Adam Gallagher factors out that should you Google a selected factor you need to cook dinner, the AI-generated abstract that seems on the prime of the web page could also be an entire recipe that was created with components of recipes taken from the Inspired Taste weblog and others which are smushed collectively and in the end won’t make something price consuming.
“We call them Frankenstein recipes,” he mentioned. Although the summaries hyperlink again to their sources, Gallagher mentioned, few individuals click on via to it, relying as a substitute on the abstract outcomes.
Google says it cares deeply in regards to the high quality of the data on the web and factors to its policies that prohibit producing content material at scale for the first goal of manipulating search rankings. Last yr, the corporate says, it launched a coverage against scaled content abuse that targets the creation of many low-value pages for the first goal of manipulating search and information rankings.
As for the AI summaries in its search, the corporate says customers had been searching for these sorts of instruments, which are supposed to spotlight and assist floor content material moderately than stop site visitors to it.
“AI Overviews are often a helpful starting point to learn about a dish, but we see that people still want to go and read original recipes from creators. We’re focused on making it easy for people to discover and visit useful sites that have a good user experience.”
The Gallaghers mentioned the AI spam, or slop, doesn’t damage established websites like theirs as a lot as a result of they’ve a loyal viewers that involves them immediately for his or her recipes.
Instead, they fear about meals bloggers who’re simply getting began, who rely on social media and search instruments to generate site visitors.
“Unfortunately, on all of these platforms, you have high-quality, trusted creators intermingled with all of this AI slop,” Adam Gallagher mentioned.
When it’s onerous to inform them aside, it erodes belief.
“When they keep chiseling away that trust, then you’re going to trust us less too, because you’re not going to find us,” Joanne Gallagher mentioned. “You’re not going to be using Pinterest or Facebook or Instagram or whatever to actually find recipes anymore.”
As for me, although I reside and work on-line and Googling has develop into my default option to discover nearly every thing, I discover myself popping out of this episode with a brand new appreciation for my cookbooks.
Who is aware of, possibly AI will find yourself boosting guide gross sales, which wouldn’t be such a foul factor.