Thousands of websites are presenting AI-generated recipes as the work of real chefs and home cooks. The sites look like ordinary food blogs, but their stories, photographs, authors, and recipes can all lead back to invented identities.
Decrypt Ads identified more than two thousand websites that appear to operate as AI slop farms. These sites are pushing fake recipes into Google search results and social media platforms such as Pinterest, where people often look for meal ideas and cooking instructions.
The findings raise a simple question for anyone searching for a recipe: who actually created it, and did anyone cook it?
A fake chef named Elena
The most aggressive site uncovered was AddCakes, which presents a character named Elena as its chef. Elena is an AI-generated version of Emily Blunt, the British actress. The image gives the site a familiar face, but the identity behind the cooking persona is fictional.
The person behind AddCakes is based in Morocco. Blunt does not live in Morocco, and her representatives did not respond to a request for comment. The site uses the AI-generated image as part of a broader story designed to make Elena look like a real expert.
Most of the AI recipe websites identified in the investigation have fake “about” pages. Those pages contain fabricated backstories about expert chefs, including claims about professional experience and the places where the chefs supposedly worked.
That backstory matters because it changes how readers judge the recipes. A visitor may believe a dish came from a skilled cook with years of experience, not from a website built around generated content and a fictional author.
How the network grew
Zach Edwards, a security expert and Decrypt Ads researcher, initially identified about 400 suspicious websites. Technical fingerprints then connected hundreds more sites to suspicious domains, bringing the total to roughly 2,100 domains.
Decrypt Ads is a startup that tracks online advertisement supply chains. Its findings show how the same technical clues can connect sites that appear separate on the surface. A visitor may see different names and designs, while the underlying network points to related operations.
These sites are not limited to one corner of the internet. They are pumping content into Google and Pinterest, giving AI-generated recipes a path toward people who may trust search results or cooking posts without checking who created them.
“This report is about publishers who are being deceptive. Their audiences could be impacted, advertisers are wasting money,” Edwards said.
Recipes that may never have reached a kitchen
The central problem is not only that the chefs are fictional. None of the recipes on these sites have probably been cooked by the claimed owners or authors.
Edwards described the kind of story these pages tell: “This lady was a Michelin star chef and she worked at this fancy hotel or this place, and they’re basically saying that this expert not only wrote these recipes, but tested themselves. And it’s giving this air of, ‘Wow, you’re about to get a recipe from an expert. This has all been carefully vetted.’”
He then explained what makes the claim so troubling: “But none of these recipes have probably been cooked — zero of them by the actual owners or authors, and to me that’s the scary part.”
Maggie Harrison Dupré summed up the deception in four words: “It’s all a lie.”
For readers, the result can be confusing at best. A recipe may look polished, carry a convincing photograph, and sit beside an emotional chef biography, yet have no real cook behind it. The pages are built to look like food publishing, even when the claimed authors and their kitchen experience do not exist.
The advertising and privacy risks
The websites also create problems beyond unreliable recipes. Publishers in certain niches are being deceptive about who owns their sites and who writes their content, while advertisers may spend money reaching audiences through those properties.
Websites working with Chinese and Russian adtech partners carry a privacy risk for consumers. Authoritarian regimes may look to hoover up browsing data, which makes the advertising systems connected to these sites a concern for people who visit them.
That gives the fake recipe network two layers of risk. Readers may receive content from an invented chef, and their browsing data may move through advertising systems tied to partners that create privacy concerns.
The scale is the clearest warning. What began as about 400 suspicious websites grew to roughly 2,100 connected domains after technical fingerprints revealed hundreds more. AddCakes stands out because it uses an AI-generated Emily Blunt as Elena, but the larger issue is the system around it: fake identities, fabricated histories, generated recipes, search distribution, social media reach, and online advertising all working together.
The investigation was published on October 20, 2026 at 8:20 AM EDT. Its message is hard to miss: a polished food blog is not proof that a real chef created the recipes, tested them, or even exists.
