This past March, when Google began rolling out its AI Mode search capability, it began offering AI-generated recipes. The recipes were not all that intelligent. The AI had taken elements of similar recipes from multiple creators and Frankensteined them into something barely recognizable. In one memorable case, the Google AI failed to distinguish the satirical website the Onion from legitimate recipe sites and advised users to cook with non-toxic glue.
Over the past few years, bloggers who have not secured their sites behind a paywall have seen their carefully developed and tested recipes show up, often without attribution and in a bastardized form, in ChatGPT replies. They have seen dumbed-down versions of their recipes in AI-assembled cookbooks available for digital downloads on Etsy or on AI-built websites that bear a superficial resemblance to an old-school human-written blog. Their photos and videos, meanwhile, are repurposed in Facebook posts and Pinterest pins that link back to this digital slop.
Recipe writers have no legal recourse because recipes generally are not copyrightable. Although copyright protects published or recorded work, they do not cover sets of instructions (although it can apply to the particular wording of those instructions).
Without this essential IP, many food bloggers earn their living by offering their work for free while using ads to make money. But now they fear that casual users who rely on search engines or social media to find a recipe for dinner will conflate their work with AI slop and stop trusting online recipe sites altogether.
“There are a lot of people that are scared to even talk about what’s going on because it is their livelihood,” says Jim Delmage who, with his wife, Tara, runs the blog and YouTube channel Sip and Feast.
Matt Rodbard, the founder and editor-in-chief of the website Taste, is even more pessimistic. Taste used to publish recipes more frequently, but now it mostly focuses on journalism and a podcast (which Rodbard hosts). “For websites that depend on the advertising model,” he says, “I think this is an extinction event in many ways.”
The holiday season is traditionally when food bloggers earn most of their ad revenue. For many, this year has been slower than usual. One blogger, Carrie Forrest of Clean Eating Kitchen, told Bloomberg that in the past two years, she has lost 80% of her traffic.
Others, like Delmage and Karen Tedesco, the author of the blog Familystyle Food, say their numbers, and ad revenue, have remained steady – so far. They attribute this to focusing their energies less on trying to game the search engines than on the long-term goal of attracting regular followers – and, in Delmage’s case, viewers.
Tedesco’s strategy has been to create recipes that rely on her experience and technical knowhow honed by years in restaurant kitchens and as a personal chef. Her Italian meatball recipe, for example, based on her mother’s, includes advice about which meat to use, an explanation of why milk-soaked breadcrumbs are essential for texture, and a dozen process photos and a video.
But she is still worried about the potential impact of AI. When she recently did a Google search for “Italian meatballs”, Familystyle Food appeared as the top result. Then she switched to AI Mode. There, she found the recipe had been Frankensteined – or “synthesized” as Gemini put it – into a new recipe with nine other sources (including Sip and Feast and a Washington Post recipe for Greek meatballs). The AI-generated recipe was little more than a list of ingredients and six basic steps with none of the details that make Tedesco’s recipe unique.
AI Mode linked to all 10 recipes, including Tedesco’s, but, she says, “I don’t think many people are actually clicking on the source links. At this point, they’re absolutely trusting in the results that are getting thrown in their faces.”
Other bloggers have seen a more definite impact on their viewership. Adam Gallagher, who runs Inspired Taste with his wife, Joanne, and who has become an outspoken critic of AI on social media, told the podcast Marketing O’Clock that since spring, he has noticed that while the number of times viewers saw links to the site on Google has increased, the number of actual site visitors has decreased. This indicates, to him, that users are satisfied with the search engine’s AI interpretation of Inspired Taste’s recipes.
After the Gallaghers posted about the discrepancy on X and Instagram, a number of readers replied to say they had not realized there was a difference between the recipes on the blog and the version that showed up in Google searches. They had just appreciated the convenience of not having to click on another website, especially when Google’s page design was so clean and uncluttered.
Rodbard acknowledges that many food blogs have gotten ugly and overloaded with ads, which has exacerbated the problem. “Ad tech on these recipe blogs has gotten so bad, so many pop-up windows and so much crashing, we kind of lost as publishers,” he says.
According to Tom Critchlow, the EVP of audience growth at Raptive, a media company that works with many food bloggers to find advertisers, it isn’t ads that are driving viewers away. It’s Google itself, with its changes to the algorithm and now with AI Mode, that’s making the sites harder to find.
There is some hope though: a survey of 3,000 US adults commissioned by Raptive showed that the more interaction people had with AI, the less they wanted to engage with it, and nearly half the respondents rated AI content less trustworthy than content made by a human.
But unless the public rebels against AI Mode, there is only so much bloggers can do. They can block OpenAI’s training crawler, which gathers information that ChatGPT uses to create content, including its own recipe generator, but theyare not necessarily willing to make themselves invisible to web searches; as Delmage puts it: “You can’t bite the hand that feeds you.”
There is also the option of moving over to a subscription model, such as Substack or Patreon, and keeping the recipes behind a paywall, but both Tedesco and Delmage point out that the most successful Substackers, like Caroline Chambers or David Lebovitz, came to the platform with much more substantial followings than they have. “If I were to give up my website or even try to go over to Substack, I would be broke,” Tedesco says.
Rodbard suggests that the analog version of the recipe blog, the cookbook, might be due for a comeback. Cookbooks, after all, offer the same experience of spending time and learning from a trusted source, and it’s likely the recipes have been tested. As a bonus, unlike phones or laptops, they don’t go dark when you neglect them for too long and you can splash tomato sauce on them without inflicting permanent damage. According to the market research firm Circana (formerly BookScan), sales of baking cookbooks are up 80% this year, but other areas have been relatively flat.
But AI bots are stealing from published cookbooks, too. When Meta was training its own AI, it compiled thousands of books into a dataset called Library Genesis (LibGen). Now unscrupulous publishers have raided LibGen and repackaged some of the books into dupes, which they are selling on Amazon.
As more people become aware of the amount of AI slop on the internet and how to identify it, Critchlow believes they will develop a greater appreciation for content produced by humans. “People will ultimately place a higher premium on being able to know that these recipes have been tested and made by somebody that I follow or somebody I respect or somebody that I like,” he says.
The recipe creators themselves are not so sure. “I’m putting my faith in that there’s always going to be a segment of people who really want to learn something,” Tedesco says. But as for the business of blogging itself, “it’s like a rolling tide. It’s always up and down and you have to roll with it and adapt.”
