Google is taking generative ai to… where else? — Google Maps, which promises to help users find interesting places by using large language models (LLM).
The feature will answer queries about restaurant or shopping recommendations, for example, using its LLM to “analyze detailed information from Maps about more than 250 million places and trusted information from our community of more than 300 million contributors to quickly make suggestions about where go. ”
Google says the feature will be available in the US first, but there's no word yet on when other countries will get it.
It's a natural place for Google to flex its ai muscles
It's a natural place for Google to flex its ai muscles, as the company has been working for years to turn its flagship navigation product into something that's more like a search tool for discovering new places, rather than just a place to get directions. Using its expertise in generative ai to accelerate that transition seems like the next logical step.
The examples Google provides of its generative ai search feature in use are very Google:
Let's say you're visiting San Francisco and want to plan a few hours of searching for unique antique finds. Ask Maps what you're looking for, such as “places with an old vibe in San Francisco.” Our ai models will analyze Maps' rich information about nearby businesses and places along with photos, ratings and reviews from the Maps community to give you trusted suggestions.
Google says it's starting small with this first integration, working collaboratively with the company's Local Guides contributor community to ensure generative ai is carefully employed. That means only these contributors will initially have access to the generative ai feature in Google Maps. The company will then open it to other users at a later date.
Local Guides, which have been around since at least 2019, describe themselves as “a global community of explorers who write reviews, share photos, answer questions, add or edit places, and check data on Google Maps.”
It's unclear how different this will be from, say, simply searching for “SF vintage store” in the normal Google Maps search bar. I'm guessing the response will be more conversational in that chatbot style we're slowly becoming familiar with, rather than simply providing a list of old stores in order of proximity.
