Showcased this week, the latest version of Google Assistant has distinctly human-like qualities. Are you comfortable with the advancing AI technology and would you be willing to share your data with it to make it even better?
When I watched this video from Google IO, its annual showcase for developers, I gasped and let out an expletive.
It shows Sundar Pichai, Google’s chief executive, demonstrating the next leap forward for Google’s virtual assistant.
In an apparently real conversation between an unsuspecting woman in a hair salon and Google Assistant, the AI bot books a haircut for the device’s owner. What is really startling is how the Assistant adds ’umms‘ and ’mm-hmms‘ in the right place, leading to a rash of think-pieces wondering if this is the moment that a robot passes the Turing test.
If that’s not amazing enough, a subsequent demonstration, with the Assistant trying to book a table with someone whose speech isn’t always clear, is astonishing.
And what’s really exciting is that this technology is coming to a phone near you in the not-too-distant future.
Android P operating system
Google IO’s stated purpose is to show off to developers where the Android platform is going and give them tools to develop their applications to run on the new version of the platform.
It also gives developers and tech journalists like me the chance to get our hands on a beta version of the new Android P operating system.
There are lots of interesting things thrown up by this that are worth mentioning. First, if your phone qualifies for the beta of Android P will you be installing it?
Installing a beta is always a bit of a risk, especially on a device you rely on every day: Google has warned that ’system and app performance is known to be periodically slow and janky‘ on its Pixel phones, for example.
Second, would you be comfortable delegating tasks to a virtual assistant that involves it interacting on your behalf with a third party?
I use both the Google Assistant and Amazon’s Alexa a lot. On my phone, I quickly dig up information from the web and surface content on the device itself by summoning the assistant. At home, I use Alexa to control my lights.
Doing both of these felt downright unnatural at first, but it’s second nature now. So while the thought of getting a virtual assistant to make phone calls for me is now rather unsettling, I have no doubt I won’t bat an eyelid in a few years’ time.
The bigger question to consider, though, is data sharing. We tend to think of data as personal information about ourselves, the kind of information Facebook and Google hoover up, potentially at some cost to our privacy.
However, signing up to a software beta and consenting to our calls being used to train AI to build products is also sharing data – but in a good way. If you consent to this kind of data sharing, you are contributing to building something and, hopefully, to making it better.
When I install the beta of Android P on my phone later today, I’ll be agreeing to share details of how it works with Google. And by doing that, I’ll be helping Google build the services of tomorrow.
That seems like a fair exchange: I get the geeky fun and satisfaction of an early look at the next generation of something I’m interested in, and Google gets data to help make it good when it eventually gets rolled out to consumers.
How do you feel about the kind of services that Google – and others – are building? Will you be helping tech companies by sharing data with them?