Google injects AI and machine learning into even more products

By on
Google injects AI and machine learning into even more products

Artificial intelligence was the star of the show at the 2018 Google I/O 2018 developer conference on Tuesday as a driving force behind all of the internet giant's product updates.

Google is injecting AI and machine learning into some of its biggest products, including Google Assistant and in Android P, as well as within some of its popular applications, such as Google Maps and Google News.

The intended result? AI will enable Google to solve problems for users all over the world, CEO Sundar Pichai told the audience of 7000 during Google I/O's opening keynote.

"We are at an important inflection point in computing… AI is going to impact many fields," Pichai said. "It's been great to be able to reinvent products with AI."

For one, Google has been on a mission to make its personal assistant technology, Google Assistant, more naturally conversational. The company has been using AI under the hood to help Google Assistant, which can be used in over 5000 devices from dishwashers to doorbells, understand the social dynamics of conversations and become more visually helpful.

Google has tweaked the technology so that users won't have to say "Hey, Google," every time they'd like to convey a request. Google Assistant will also be able to understand if a user is talking to the technology or not, and additionally, users will be able to ask more than one request at the same time. This new feature, Continued Conversation, will be available later this year.

Google is also currently working on a new feature that will let Google Assistant place phone calls on behalf of a user. Using natural language processing technology, this new feature, Google Duplex, will be able to book appointments or make reservations for users.

The technology, which sounds convincingly human by including filler words like "um" and 'hmm" during transactions, can even handle complex interactions or back-and-forth conversations with answers that may not be anticipated, Pichai demonstrated during his keynote. The feature will be especially useful for users to contact local, small business that may not have automated systems, he added.

"We are working hard to get this right and working hard to give users back their time," he said.

Pichai said that news is core to Google's mission, acknowledging that many rely on the Search giant's platform for its news summaries. To that end, Google is updating its Google News app using AI to bring forward high-quality, trusted sources.

The new Google News app, which will be rolling out across 27 countries next week, is designed to help users keep up with the stories they care about. The new Full Coverage feature invites users to learn more about a topic by giving a bigger picture of how the same story is being reported by a variety of other unfiltered sources. The updated app will open with a briefing of the top five news stories, as well as a personalized briefing for each user with both global and local content that improves and becomes more targeted each time it's used. 

"Above all, we want to make sure we are giving [users] deeper insight and a fuller perspective on any topic they are interested in," Pichai said.

Along similar lines, Google Maps is also being bolstered by machine learning. Google Maps will now include a "For You" tab that will offer a rundown of new places and personal recommendations for users on places such as new restaurants to try.

Google took to its annual developer conference to share updates on perhaps the most widely speculated topic ahead of the show, the Android P operating system.

Android P will also be using AI to adjust to user behavior and environment. For example, the latest feature, Adaptive Battery, uses machine learning to adapt to application usage patterns to reduce CPU app "wake ups." The OS will also feature an Adaptive Brightness setting that can adjust the devices' brightness based on user preferences, or room or outdoor settings, meaning less manual adjustments will need to be made by the user.

Google Search is also getting some machine learning assistance. When a topic is searched, Google will be able to predict actions based on user behavior and automatically prompt a user to act. For example, if a user searches for Uber, Google can highlight how long a ride would be to a particular location, like home, display the estimated price of the ride, and offer the user the ability to order the ride.

In an effort to better balance technology with being present, Android P can show a user how much time they've spent on their device, set time limits on specific applications, or even switch the interface to a less stimulating grayscale after a set time to help users sleep better, according to Google.

Ahead of the Google I/O conference, the company revealed Tuesday morning that its research division has been rebranded to Google AI, a segment that will now house all of Google's research efforts going forward because AI and machine learning have become core to the development of all its services, according to the company's official blog post.

This article originally appeared at

Got a news tip for our journalists? Share it with us anonymously here.
Copyright © 2018 The Channel Company, LLC. All rights reserved.

Most Read Articles

Log In

  |  Forgot your password?