Skip to main content

Google heads into the kitchen with new recipe search feature

Google wants to be your sous chef the next time you’re planning a meal. Well, sort of. The Internet giant would at the very least like to help you select a recipe. (For the time being, you’ll still have to dice your own veggies.) This week, Google unveiled Recipe View, a new way to search for and filter recipes on the Internet.

The feature lets amateur chefs (and even professional chefs, we suppose) filter the Web to only show pages with specific recipes. Simply select “recipes” from the sidebar on the search results display page, and voila, only pages that contain instructions for making steak tartar, candied yams or whatever it is that’s prompted you to preheat your oven. If you select a more broad category, say “thanksgiving,” you’ll get a list of recipes associated with that term. Search results can be further filtered by ingredients, cooking time and even by number of calories.

What allows Google find and display pages with carefully parsed recipes? The key is in structured data — something that has been sorely lacking in the relatively brief history of World Wide Web.

Wired offered this explanation:

The entire search is built on structured data that webmasters have built into their webpages using markup code that’s invisible to humans but is extremely useful to machines. The dream of the so-called semantic web is built upon the idea that web pages will be full of such underlying tags so that search engines can parse a webpage to learn someone’s e-mail address or know exactly what a restaurant’s operating hours are by scanning underlying code invisible in the browser…The dream of a structured web has proven nearly impossible to create in practice as it requires coordination on building specs and then that web page builders take the time to mark their pages up in complicated XML

The lack of an established practice of marking up content really hinders an operation like Recipe View. And it shows. After performing a couple of searches through Recipe View, it’s clear that the feature leans heavily on bigger sites that are likely to be carefully marked-up to be more search engine-friendly. That means you’ll be seeing a lot of results from recipe aggregation sites like All Recipes, Epicurious and the Food Network, but nary a dish from someone’s personal cooking blog. It’s not a big deal, but don’t expect to find anything beyond the well-grazed options that you could likely uncover with a little scrolling through the results of Google’s standard search.

However, as Web designers and developers adapt better structure into their coding, search functions like Recipe View will become more accurate, more comprehensive and more attuned to the subtleties of human language. Who knows,  as we recently saw IBM’s Watson conquer humans in Jeopardy, maybe we’ll one day see a computer crowned Top Chef.

Topics
Aemon Malone
Former Digital Trends Contributor
A dangerous new jailbreak for AI chatbots was just discovered
the side of a Microsoft building

Microsoft has released more details about a troubling new generative AI jailbreak technique it has discovered, called "Skeleton Key." Using this prompt injection method, malicious users can effectively bypass a chatbot's safety guardrails, the security features that keeps ChatGPT from going full Taye.

Skeleton Key is an example of a prompt injection or prompt engineering attack. It's a multi-turn strategy designed to essentially convince an AI model to ignore its ingrained safety guardrails, "[causing] the system to violate its operators’ policies, make decisions unduly influenced by a user, or execute malicious instructions," Mark Russinovich, CTO of Microsoft Azure, wrote in the announcement.

Read more