Last night Google held its latest glitzy product update presentation at the start of its I/O developers conference in San Francisco. Clocking in at a punishingly long 2 hours and 17 minutes, the presentation was a detailed look at updates across the entire suite of Google products - from Android M, to wearables, to virtual reality.
Perhaps the most striking announcements were the ones that resonated with Google’s oft-repeated mission objective: To organise the world’s information. Early in the presentation a representative from the company demonstrated the power of Google’s machine learning techniques - with the new Google Now on Tap feature able to deduce what the user is looking for by considering the context in which they’re asking.
For example, if the user is listening to dance music producer Skrillex on their Android phone, they can use Google Now to voice search “what is his real name?” - and without the user even having to specify they mean Mr Skrillex, Google would know that “his” refers to Sonny John Moore himself. As the presenter remarked, from a computer science perspective, this is mind-blowing, as it means that Google has managed to not just decode a voice and turn it into text, but actually understand the meaning of the words used, and provide a meaningful answer.
The machine intelligence can also be applied to images. In the presentation, Google was able to pick a Tree Frog out of a selection of frogs by analysing the photo using “deep neural networks” to figure it out. A later part of the presentation, focusing on Google’s VR work showed similar image analysis, which was able to combine multiple photos to determine the depth of objects in a picture. Whilst it isn’t exactly surprising that Google is capable of doing stuff like this, it is a brazen demonstration of what machine learning can do and how much power the company wields.
A bit later on in the presentation Google revealed that it is spinning off photo uploads from Google+, and launching Google Photos as a separate app. It looks like a nice way to organise all of your pictures, and has some neat sharing functionality. But then came the surprise: to coincide with launch, Google is now offering unlimited, free photo and video storage for photos up to 16MP and video up to 1080p.
Cleverly, Google Photos too will use machine learning algorithms to analyse your photos - capable of facial recognition that can even transcend generations. The presenter showed how the app had grouped together photos of his eight year old niece - from birth right through to now, recognising it was the same person. Photos will automatically sort your, umm, photos, into collections - based on the people in them, the location you took them, and even other objects in them (such as if you take photo of Tree Frogs).
And that’s when it hit me: When will we finally decide this getting a little bit creepy?
Your Life on Google's Buffet Cart
Nothing that Google showed off last night is particularly revolutionary in and of itself, but taken cumulatively, seeing the full reach of Google’s abilities, we’re surely reaching a point where this becomes more weird than neat.
Over the course of one presentation, Google showed us how it can interrogate and understand photos and other complex data, and then baited us to feed its insatiable appetite for data. Every photo we upload will be good for us, as it will mean we can back-up precious memories to the cloud… But it will also be more raw meat on which Google can sharpen its algorithms and software. All of that free storage may seem like a free lunch, but it turns out that we are the ones who are on the menu.
Machine learning and related technologies are only going to get smarter - meaning that the data we feed Google today is only the start. Who knows what Google will be capable of mining out of the photos that we have already uploaded by this time next year? Or in ten years?
Even leaving aside the obvious implications for the sort of government surveillance that Edward Snowden revealed to the world, surely I’m not the only one unnerved that we’ve handed over so much of our lives to a single corporation? When is this going to get weird?
There are plenty of prior examples of Google’s all-knowing Professor Xavier-style powers. I remember when Google Now was first launched speaking to one Android user who was stunned by just how Google magically knew so much about where they were and what they were doing.
In fact, it is conceivable that perhaps one day Google will know that you’ve searched to find out the best rides at Disneyland, and will have already checked the weather and your calendar to see what day you are free to go (probably with that friend you take all of those selfies with). But rather than creep the hell out of you when you wake up by predicting your plans before you make them, it might deem it necessary to play dumb instead, so that you’ll keep feeding it data without being creeped out.
There will probably never be a single moment when we all realise what is going on, due to the drip-drip nature of technological innovation and our individual tolerances for just how much information about ourselves we’re willing to surrender. But just as natural selection occurs at an undetectable pace, one day we might wake up and wonder why all of the dinosaurs that have been protecting our privacy have disappeared (okay, this metaphor doesn’t quite work - but I bet Google’s algorithms are smart enough to find me a better one).
With companies that deal in information like Google, the trade off has always been that we surrender some of our personal information in exchange for useful services - but whether the convenience will always scale with the price remains to be seen. What is clear is that the more data we provide, the more locked in to Google’s ecosystem we become. Bing will never give results as accurate as Google, because the latter knows you so much better - so even if we do decide we’re creeped out by Google, not using the company’s products and services is going to be nearly impossible.
It isn’t an original observation, but the Stasi would be jealous of the tight grip that Google could have over our lives. In fact, perhaps we’re already past that point, in this article I’ve been referring to Google as though it is a sentient, like Terminator’s Skynet. It's almost cliche to make the reference at this point, but we're inarguably marching closer to at least elements of that frightening sci-fi future becoming a reality.
During the presentation last night, Google kept repeating the “organise the world’s information” mantra - but perhaps at some point soon we might want to start asking whether some information is better left messy?
Image Credit: Good WP