AI will soon be your personal tour guide on Google Maps

Indoor playgrounds? Obscure flea markets? Hipster burger joints that use donuts or grilled cheeses as buns? There could be a lot of cool — or unusual, or just helpful, or downright ridiculous — stuff hiding right on your block, and you’ll soon get help from Google Maps in sniffing out the best in your neighborhood.

According to a report today from The Verge, Google is working on a new AI-driven component that will comb Google Maps’ network of registered places and community contributions to make speedy recommendations, whether you’re inquiring after shopping centers or restaurants or tire shops. The feature will (hopefully) be more interactive and provide a more customized search result than just using Google Maps’ current search bar.

The generative AI feature will reportedly be able to pick up on keywords in search queries — like trendy, vintage, upscale — and cross-reference them against hundreds of millions of business listings, reviews, and photos to deliver recommendations. How the considered listings will be honed in geographically is unclear; the Verge quoted Google as saying that including a city or region by name, like Austin or Texas, will instruct the AI system to curate recommendations within those bounds. It’s likely that using prompts such as “nearby” or “within 20 miles” will also be accepted.

Related

Google’s most important AI advancements of 2023

A relentless quest to see what sticks

Google Maps’ AI makeover, which is powered through large language models (LLM), will see a slow roll-out. It will start in the United States, but there’s no word yet on an international expansion. And only a small group of U.S.-based users, specifically Google’s Local Guides, will have access to the new AI feature at first.

Google’s been busy with Maps-related changes, in an effort to take the application beyond being a simple GPS into a more informational hub. To streamline dining and restaurant listings, Google Maps recently rolled out a feature that automatically assigns information to random user-uploaded food pictures on a restaurant’s map page. That feature will analyze the pictured food, identify the menu item it most likely reflects, and even provide dietary information like whether the dish is vegetarian. It’s easy to imagine how a feature like this would often be glaringly imperfect; for that reason, Google Maps allows users to suggest edits for misidentified pictures.

But it’s not all good news — Google Assistant Driving Mode, which helps minimize distracted driving by accepting simple voice commands for things like reading messages and making calls, will be laid to rest this month.