Google is tweaking its search engine to emphasize more personalized results and more compelling images.
Executives said on Monday that the changes were driven in part by the company’s increasing use of artificial intelligence technologies like deep learning. The technology can better determine the kind of information related to a particular search, even if it’s not obvious from the query used, and to do tasks like more accurately identify cats in photos.
A new “Journeys” feature is intended to keep track of users’ individual searches about particular topics. People searching for “camping,” for example, will be able to see their previous searches about camping gear under a new “Activity Card” tool, Nick Fox, Google vice president of product and design for search and the Google Assistant, said at an event in San Francisco. The idea is that people tend to repeatedly search for a particular topic as opposed to only searching for something once; it’s essentially a better-indexed version of the company’s search history feature.
People will be able to opt-out of the new feature if they don’t want their searches cataloged for them.
Google will also show more detailed information about particular topics at the top of the core search query page. For example, users who search for “pugs” will see a list of topics like “health” or “where to adopt” that they can then can click on for more information. Meanwhile, users who search for “Yorkshire terrier” will see topics that are relevant to that particular breed, like “how to groom,” because that breed has particularly long and wild fur.
The search giant is also debuting a new “Discover” tool that will show up on the main Google homepage on mobile browsers, said Shashidhar Thakur, Google’s vice president of engineering for search. The “discover” feature works similar to Facebook’s News Feed by presenting articles and videos to users based on what an algorithm identifies as most likely to appeal to them. This discover tool could keep people glued onto Google’s search engine if it shows people a greater variety of information that they otherwise would not have asked for.
Google did not say when or if Discover would debut on desktop browsers.
It’s unclear what safeguards Google is taking to ensure that the new Discover tool doesn’t recommend that people click on the kind of fake news and conspiracy theories that have plagued Facebook’s News Feed. To be sure, Google’s search recommendations aren’t influenced by what friends or family members are sharing like Facebook’s News Feed, but is influenced by one’s personal searches.
Cathy Edwards, director of engineering for Google Images, also said that Google Image search has been updated to make it easier for people to access pages to buy products that they searched pictures for or to find other helpful information related to a certain picture. Edwards said people searching for images of “swings” would now see photos of swings from websites where they can buy one, see third-party tutorials about how to install a swing set, and see a link to “stock photo” sites, in case that’s what they were looking for.
The change to Images search is similar to recent changes Facebook’s Instagram made to its popular Stories feature that lets people more easily buy products like purses from third party retailers when they see something they like in a particular picture or short video.
Get Data Sheet, Coins2Day’s technology newsletter.
Edwards also said that Google would debut its own version of Stories—the collage of text, short videos, and photos that photo-sharing service Snap popularized and many companies like Facebook have imitated. Now when people use Google to search for certain celebrities or professional athletes, they will see a Google Story that shows information like the person’s date of birth, recent highlights, and other content from the web. Google already bundles information on certain topics into its “Knowledge Panel” feature, but the new feature appears to present some of the information typically shown in the knowledge panel feature in image-friendly Stories feature.
Google did not say when it would expand the new Story feature to include more than celebrities or athletes.
Google also started testing a new feature in India that will use the company’s AI technologies more accurately predict when and where flooding may occur during storms or other natural disasters, explained Google search senior product manager Nick Zakrase.
The company said in blog post that it is partnering with India’s Central Water commission for unspecified data that it will use with information like “historical events, to river level readings, to the terrain and elevation of a specific area,” to predict where flooding may take place.
