If Google is in the middle of reinventing itself around AI, you wouldn’t know it from the latest Google Maps updates.
Yes, there are the obligatory new Gemini AI integrations in the form of review summaries and more granular search queries, and those are what dominate Google’s press release about the news. But for the most part, the update is about meat-and-potatoes improvements to the core mapping and navigation experience, most of which don’t involve generative AI.
“We’re not just doing it to do it,” Chris Phillips, Google’s general manager of Geo, says of Google Maps’ AI features. “We’ve been really, really careful to make sure it’s used for helping us build the cutting-edge with Maps in a responsible way. That’s been a really important priority for us, versus just being able to say every feature is powered by Gemini.”
What’s new in Google Maps
One of the most promising new Maps features simply adds more detail to turn-by-turn navigation, including crosswalks, road signs, and—most importantly—clearer instructions on which lanes to turn into. While Google Maps has offered lane guidance in its top navigation banner for more than a decade now, the map itself will now start showing which lane to be in as you move through the turn. Google says these “enhanced navigation” features will arrive in 30 cities next month and will expand over time.
Google’s also expanding its “Immersive View” feature, which stitches aerial and Street View images into a 3D perspective that includes simulated weather and traffic. The feature will soon be available in 150 cities after debuting in just a handful last year, and it’ll add college campuses to its list of supported places. You can also use “Immersive View for Routes” for a 3D preview of what your drive will look like, which will now include icons that highlight parking options and difficult turns.
Once you’ve arrived, Google Maps will try to fill in some extra details. After parking down the street or in an offsite parking garage, for instance, you’ll see quick launch buttons to remember your parking spot, get walking directions the rest of the way, or bring up an augmented reality “Live View” of your current location. For buildings, the map will also highlight where the entrance doors are.
Even before you start driving, Google can also suggest notable places to stop along the way, including popular restaurants and points of interest. These will appear under a new “Add Stops” button, next to option to start navigation. Users will now be able report weather disruptions along the route as well.
The point of all these updates, Phillips says, is to give people more confidence during all phases of the trip.
“That ability to actually roll forward two days and see what the weather is going be like, and the temperature overlay with where you’re going to park and walk, this is a level of insight that is just really the next phase of innovation and mapping,” he says.
Where generative AI comes in
Google does see a role for Gemini in the trip planning process, mainly as a way to sift through information in user-contributed reviews.
When searching for restaurants, for instance, Gemini will allow for more detailed natural language queries. While basic searches such as “coffee shops near me” won’t invoke Gemini, the assistant will kick in for more complex requests such as “restaurants with a hip vibe and outdoor seating.” Gemini will also generate summaries of user reviews and help answer follow-up questions, such as “how noisy is it” and “is there free parking nearby.”
“It’s not just indexing places with certain attributes, but it’s digging in through what could be thousands of reviews where people have talked about how great the outdoor seating is,” Phillips says.
Meanwhile, Google’s Waze app is tapping Gemini to help users report road incidents by voice. Tapping the hazard button will soon bring up a voice prompt, so you can say things like “we’re in bumper-to-bumper traffic” or “there’s a double-parked car in front of me.” The app will then ask for follow-up information if needed before generating the appropriate report. (This feature is launching in beta for “trusted testers” this week.)
But the most intriguing use for Gemini doesn’t involve the actual Google Maps or Waze apps at all. Instead, Google is offering Gemini as a tool for third-party developers and for organizations that want to analyze data in Google Earth. Phillips gives the examples of an apartment rental app letting users ask about nearby dog parks, or a city planner figuring out which highly-populated areas have the lowest density of EV chargers and land available to install them.
“What is literally taking city planners and urban developers months of surveying that’s super expensive, we can give them in minutes,” he says.
While Google’s Geo unit is dabbling in generative AI, Phillips says there’s no company directive or incentive to cram in more Gemini features, and he’s wary of introducing it in a way that would erode users’ trust in the data. Even for the Gemini features that Google has announced in Maps, he describes a “grounding” process that effectively-fact checks the answers it produces against the data that Google has.
“You don’t want a model that’s going to hallucinate and . . . makes up a place to go,” he says.
Ultimately, Phillips says generative AI is just one more tool Google can take advantage of as it tries to build new mapping experiences.
“If we know people trust us and have confidence, and we’re helpful in our product, then we have the permission to build really cool next-generation mapping experiences,” Phillips says. “That may be powered by generative AI, that may just be powered by a different AI technique, or just powered by more interesting data.”
Autentifică-te pentru a adăuga comentarii
Alte posturi din acest grup
An online spat between factions of Donald Trump’s suppo
U.S. tech investor Cathie Wood is calling on
Visiting adult and gambling websites doubles the risk of inadvertently installing malware onto work devices, according to a new study.
There are certain social media rules we can all agree on: Ghosting a conversation is impolite, and replying “k” to a text is the equivalent of a backhand slap (violent, wrong, and rude). But what
For Makenzie Gilkison, spelling is such a struggle that a word like rhinoceros might come out as “rineanswsaurs” or sarcastic as “srkastik.”
The 14-year-old from
Japan Airlines said it was hit by a cyberattack Thursday, causing delays to
