Sunday, March 3, 2013

Why Google Glass-like recording technologies will NOT lead to Big Brother (my response to Mark Hurst's post)

Google Glass has certainly attracted the...um, eye of a number of people who are thinking about the ramifications of this technology.

Jim Tipping recently shared an item written by Mark Hurst at Creative Good entitled The Google Glass feature no one is talking about. As I began reading Hurst's piece, I ran across this line:

The Google Glass feature that (almost) no one is talking about is the experience – not of the user, but of everyone other than the user.

My first inclination was to yawn. The reaction of others to Google Glass has been discussed ad nauseum. Loren Feldman has talked about the device's effect on interpersonal relationships. A bunch of people, including me, have talked about the Shotwell's Bar incident.

But it turns out that Mark Hurst was talking about something far deeper.

What makes Glass so unique is that it’s a Google project. And Google has the capacity to combine Glass with other technologies it owns.

First, take the video feeds from every Google Glass headset, worn by users worldwide. Regardless of whether video is only recorded temporarily, as in the first version of Glass, or always-on, as is certainly possible in future versions, the video all streams into Google’s own cloud of servers. Now add in facial recognition and the identity database that Google is building within Google Plus (with an emphasis on people’s accurate, real-world names): Google’s servers can process video files, at their leisure, to attempt identification on every person appearing in every video.


It's appropriate to disclose at this point that I work in the biometrics industry. While facial recognition has not advanced to the point where you can get a grainy picture of anyone from a distance and identify the person in milliseconds, it is certainly conceivable that the technology could move in that direction eventually. At a minimum, if you have a select list of people (for example, known terrorists) and you point a camera at a place where those people congregate, you may be able to identify when a person on the list enters that place. From there, it's merely a computational and algorithmic challenge to expand that list to a list of all Google+ users, or all Facebook users.

Oh, but Hurst isn't done yet. He makes several other points (I encourage you to read his entire piece), including this one:

Let me paint a picture. Ten years from now, someone, some company, or some organization, takes an interest in you, wants to know if you’ve ever said anything they consider offensive, or threatening, or just includes a mention of a certain word or phrase they find interesting. A single search query within Google’s cloud – whether initiated by a publicly available search, or a federal subpoena, or anything in between – will instantly bring up documentation of every word you’ve ever spoken within earshot of a Google Glass device.

Now consider that Google's competitors are not going to let Google have this market all to itself. Perhaps Apple is already working on iRecord. A slew of other companies - Facebook, Microsoft, Amazon, Oracle, IBM, Lenovo - might have their own reasons for pursuing this technology, either from the hardware perspective (the recording device), the software perspective (the integration of the device's data with other data), or both.

Certainly, in the worst case, a chilling scenario of the denial of civil liberties in the Untied States, China, and throughout the world.

But that's the worst case - which probably won't happen.

For one, it is highly unlikely that Google's data will be shared with Facebook's data or the CIA's data or whatever. Organizations have a natural tendency to keep their power to themselves, and since data is power, that means that Google has precious little incentive to let Facebook run through Google's data.

However, it's conceivable that any one particular silo - the CIA's silo, Google's silo, the DMV's silo, whatever - could, in and of itself, yield a treasure trove of data for those who are willing to find it.

Let me repeat that - for those who are willing to find it.

Even if you're willing to pay the cost to store all these petabytes or yottabytes or sortabytes of data, there is also a cost of searching through all this data. And you're only going to pay the search costs if you can realize a positive return on investment.

Even if you perform a targeted search - say, a search of all of Google's data showing Julian Assange's activities in the United Kingdom since his arrival - you're going to have to plow through a slew of unrelated data. Video of Assange sitting on the toilet. (Hey, the UK has a lot of cameras.) Video of Assange sleeping. Video of Stoke on Trent football club Assange FC.

Maybe the powers that be will pay that search costs for a Julian Assange search. But will they pay those costs for anyone else? Will the powers that be search through all of Google's archives for a few dozen people who showed up at a protest rally? Will the powers that be search the people who were watching the protest rally?

Just because you have a bunch of data lying around doesn't mean that people are going to spend the money to turn that data into information...much less knowledge or wisdom.
blog comments powered by Disqus