Google CEO Sundar Pichai this week touted several examples of how the company has successfully used artificial intelligence approaches and machine learning to improve its products and services in recent years.
Speaking at the company’s annual Google I/O developer conference at the company’s Mountain View, Calif. headquarters, Pichai on May 8 described AI as presenting a “huge opportunity” for transforming many fields.
One area where Google has been able to already make significant progress with AI is in health care, Pichai said. A neural network that Google developed two years ago can detect evidence of diabetic retinopathy using just the images of the eye. The same images can now be used to predict an individual’s risk of stroke or heart disease with considerable accuracy, he claimed.
Google has also been able to apply its AI models to large data sets of de-identified patient data to predict with a high degree of accuracy the chances of a patient being readmitted to hospital, average hospital stays and other medical events. Google is currently working with members of the medical community to figure out how to apply its research findings in practice within the health care sector, Pichai said.
Google has been using AI to make it easier for users to interact with products including Gmail, Google Photos and its Google Assistant technology as well. One recent example is the so-called Smart Compose function in Gmail, Pichai noted.
The feature uses AI to understand email messages and to help users craft quick replies by suggesting contextual phrases they can use. Similar, smart inline suggestions in Google Photos make it easy for users to share photos instantly or to perform other tasks such as brightening a photo or colorizing old black and white pictures, Google’s CEO said.
“One of the biggest time-savers of all is the Google Assistant, which we announced two years ago at I/O,” Pichai said referring to Google’s AI-powered digital assistant.
Google has said it wants to integrate the technology into as many of its products as possible so users can communicate with the products using voice commands and natural speech. Over the past two years, Google has integrated Assistant into various products including its Google Home smart speaker. “Thanks to our progress in language understanding, you’ll soon be able to have a natural back-and-forth conversation,” with Assistant, Pichai said.
As part of planned updates to the technology Google is currently working on making Assistant more visually assistive. On mobile phones for example, Assistant will enable displays that give users a quick snapshot of their day along with location-based, time-based and interaction-based activity suggestions. Google is also working on integrating Assistant into Google Maps so drivers can get navigation information more easily.
Google is also working on making Assistant capable of handling tasks that presently require a phone call—such as booking a haircut or scheduling specific types of appointments. The technology, which the company is calling Google Duplex, will offer significant value for small businesses especially while saving people considerable amount of time, Pichai said.
Other areas where Google is using AI to improve its services include Google News. The service has been newly updated and uses artificial intelligence to highlight the best and most relevant stories on a topic from journalists around the globe, Pichai said.