AI to Empower Google

In this year’s Google I/O, annual developer conference held by Google, new features and announcements were released, and they were just as good as those from previous years, if not better. Some of the key features include Android P, updated Google Home and Google Duplex.

My favorite feature, and I’m sure it’s also many people’s favorite, is Google Duplex. The 2 demos Google showed are booking appointment for hair cut and restaurant reservation (In case you haven’t watched it yet, see link[1] in the References). The way Google Duplex could understand the context and handle the nuance of conversations is just epic. It is easily the most remarkable human-computer interaction conversation I’ve never heard. In the hair salon conversation example, if not known ahead, I wouldn’t be able to tell which one is Google Duplex between the caller and receiver. I would also show no surprise if Google Duplex soon offers artificial intelligence (AI) phone answering to replace business’ call center. Even when the call didn’t go quite as expected, as in the restaurant reservation case, the thinking time for Google Duplex isn’t too long for people to raise concerns. Together with the WaveNet technology (a text-to-speech technology synthesize voices based on learning from a sample of personal speech) showed in the earlier part of Google I/O, personalized AI conversation is only a step away.

It is mainly AI, machine learning (ML) and deep learning (DL) that enabled the above feature possible. In fact, most of this year’s Google I/O releases are backed up by AI/ML/DL technology. For example adaptive battery, an on device machine learning feature on Andriod P, and ‘For You’, a more personalized recommendations on Google Map.

Like Cloud Computing is the backbone of innovations on the application/service level, AL/ML/DL is also boosting the ongoing technology wave. Although ML functionality has been made available for everyone by different products like Google ML Kit, researchers are still working on the next leap of algorithms and models behind. And obviously, Google is taking up the role. What’s the reason behind? Of course it’s a hotspot that nobody wants to let go of its potential market share. But another incentive I feel is it aligns well with Google’s core value.

As mentioned by Jen Fitzpatrick, VP of Engineering and Product Management at Google in the conference, ‘building technology to help people in the real world everyday’ has been the core of the company from the very start. If you take a closer look at the Google’s releases, all the focuses are on freeing people from mundane chores, bringing the world closer for everyone on the planet, and sparkling comprehensive understanding of the surroundings.

In contrast, this year’s Apple WWDC, corresponding developer conference by Apple, emphasizes Augmented Reality (AR), which is also a hot topic on today’s techy wave, but the manifests of AR are more on the artsy, entertaining side. Of course Google is not letting AR go, but the only relevant feature I saw from this year’s Google I/O is the cute dog jumping around guiding you in Google Map’s walking navigation. Ever since Sony PlayStation or Microsoft Xbox time, Google is not heavily investing in gaming.

In sum, guided by its core value of building products that have deep impact on people’s lives, Google is committed to pushing AI further, and AI is seeding the next technology transformation.




[1] –

[2] –

[3] –





2 comments on “AI to Empower Google”

  1. The demo of Google Duplex was very exciting for me as well. They were able to perceive the nuances of phone conversations and adapt to the different ways people express themselves. For example, I was surprised by how AI handled the call perfectly, even though it went in a very unexpected direction with the restaurant’s employee. If machines can interact so naturally with humans, the benefits and potential applications of this technology are limitless.

  2. Hello Hailun,
    The technology you shared on the post above is great and it seems hard for anyone to be able to compete with Google on this side. One of the aspects that fascinates me the most about Google is how they complement their Software / IT development business with their hardware / Data Center business. Currently, 60% of my finished goods go straight to them as they are building data centers all over the world. I basically can’t build enough for them, they are being aggressive, they are targeting specific areas where they anticipate grow and they are building enough infrastructure for a life time. It is hard to see how anyone will be able to pair with them in the long term.


Comments are closed.