Google bets on AI-first as computer vision, voice recognition, machine learning improve

At Google I/O, CEO Sundar Pichai said that all of the company and its products are being revamped to be AI-first. The shift may be bigger than mobile computing.

cloud-tpus.png

Google is revamping its infrastructure to be AI first. Credit: Google

Whether it's search, Google Assistant, Android, Gmail, Google Photos, or Google Cloud Platform and its data centers, the path to success for Google flows through two words: Artificial Intelligence (AI).

If there's one takeaway from Google I/O it's that CEO Sundar Pichai is pivoting the company to an AI-first orientation. Last year, Pichai outlined the AI theme, but this year Google's unifying theme is that artificial intelligence and machine learning has hit an inflection point.

"Computing is evolving again. We're moving from mobile first to AI-first. In an AI-first world we are thinking through all our products," explained Pichai. "We are building AI-first data centers. We are focused on applying AI to solving problems."

And, yes, AI is going to be learning how to build more AI applications.

Read also: Google's strides in computer vision leads to Google Lens feature | Google unveils next-gen TPUs to both train and run machine learning models | Google brings Smart Reply to Gmail for Android and iOS | CNET: Google Assistant makes its way to your large home appliances | Google Home can now make phone calls for free

The most common way this AI will become mainstream will be Google Assistant, which is now available on Apple's iPhone. Meanwhile, developers are utilizing more tools for AI. For enterprises, AI will be delivered via the Google Cloud Platform. Google is also using machine learning to power job searches for FedEx and Johnson & Johnson.

Google is building in more AI into Android to identify addresses and phone numbers to enable easier cut and pasting and bring in related apps like phone and maps. Google is also adding TensorFlow Lite to enable more machine learning on mobile devices.

But Google's vision and voice recognition will also be the most common way to utilize AI and machine learning no matter the device.

Indeed, the error rates for Google's computer vision and voice recognition technologies in the US are impressive.

google-speech-recognition.png
google-vision-recognition.png


Here's why Google's approach to AI is notable:

  • Google realizes that the best AI is when you don't notice it as a separate technology silo.
  • The company understands scale and has a broad portfolio to embed AI.
  • And AI needs to be everywhere.
  • AI needs to solve problems like connecting to your Wi-Fi router by using computer vision to read the bar codes and numbers on the device.
  • Meanwhile, AI is going to largely be delivered through the cloud. In the enterprise, this reality means companies such as SAP are integrating with Google Cloud Platform.

Pichai plugged Google.ai, which highlights the technologies and tools to develop machine learning and AI applications.

Read also: Google Assistant integrates with GE's connected appliances | Google Cloud IoT Core lands NXP support, guns for smart city use | NXP launches Google Android Things modules as voice AI proliferates in smart home | SAP plots more integrations with Google Cloud Platform

What's next? More applications and use cases based on Google's AutoML, which uses neural networks to create more neural networks. The AI is learning to learn and Google is setting itself up for an AI-first architecture. "The results are promising," Pichai said.

special feature

How to Implement AI and Machine Learning

The next wave of IT innovation will be powered by artificial intelligence and machine learning. We look at the ways companies can take advantage of it and how to get started.

Newsletters

You have been successfully signed up. To sign up for more newsletters or to manage your account, visit the Newsletter Subscription Center.
See All
See All