Food bloggers say Google’s AI summaries are ruining their recipes and their traffic

Google

Food bloggers allege Google’s AI summaries are starting to hurt both their recipes and their livelihoods. Many creators who have spent years testing and perfecting dishes now find their work pushed aside by quick AI-generated answers that often get basic steps wrong, according to a report by Bloomberg

Eb Gargano, who runs the blog Easy Peasy Foodie,told the publication that the holiday season looks very different from the past. Normally, her turkey and Christmas cake recipes rise in search results as people start preparing for the holidays. This year, Google is showing AI summaries instead of linking to her site. One AI version of her cake recipe suggested baking a small cake for nearly four hours — a mistake she says would leave home cooks with “charcoal.” Her traffic is already down 40 percent compared to last year.

Other food bloggers report similar issues. They say AI summaries pull pieces from many recipes and mix them together, leading to instructions that are confusing, unsafe or simply impossible to follow. Some bloggers have even found AI-generated copies of their recipes and photos on random websites and social media pages.

Platforms like Pinterest and Facebook are also filled with AI-made food photos that look impressive but don’t match the attached recipes. This makes it harder for readers to find real, tested dishes and easier for inaccurate content to spread.

For many creators, the drop in traffic has become a serious financial problem. One blogger says she has lost 80 percent of her views in two years and has had to shut down parts of her business.

Bloggers worry that if this trend continues, fewer people will see real recipes created and tested by humans. Instead, they fear cooks will rely on AI-generated instructions that haven’t been tried in any kitchen.

As one creator put it, the promise of a recipe used to be simple: “Someone has cooked this before you.” Now, many fear that promise is disappearing.