Apparently Apple might be launching a new device that promises to lure people into Apple Stores around the world in mobs, waiting in line to get their very own. Call me crazy, I’ve just never understood that … level of zeal. What’s caught my eye about the rumored tablet, though, is that it has supposedly been designed to be “intuitive to share,” including the ability to recognize users vis-à-vis a camera and a face print.
Almost 8 years ago I started out on a year-long journey of researching the intersection of biometric and surveillance technologies (pdf). Facial recognition was the main focus of the work – we had recently seen the first mass deployments of the technology in surveillance contexts in the name of security. But the technology just wasn’t that good at the time, at least not good enough to be rolled out in consumer devices.
I wonder if the rumored Apple tablet’s rumored capability will follow the same path of the fingerprint scanners that were deployed in laptops years ago. To this day, I’ve yet to see anyone actually use one of those things. I suspect two reasons why: first, it wasn’t seamless to use; second, there was an easier alternative. I can imagine that if Apple has actually done a nice job with face authentication, it might actually take off.
It’s hard for me to believe that 8 years has passed, but clearly it’s been enough time that the technology has transformed dramatically. There are a number of companies offering face recognition services that run on top of social networks like Facebook. Many commentators have speculated about the use of face recognition in Google Goggles, after the product was launched explicitly without the technology in recognition of the privacy challenges ahead.
When I wrote about the technology so long ago, I had a lot of discomfort with the implications for individual privacy and our social fabric. In general, I still do. If one thing is certain, though, it is that technology marches on. Innovation will continue, and efforts to hamper it out of fear of the unknown tend to be unsuccessful. Instead, we can influence innovation such that it moves in a direction that acknowledges and adapts to a wide range of individual discomforts and desires, even the ones traditionally underrepresented among a technologically savvy crowd.
The challenge with face recognition of course is that it’s hard to keep my face private. I walk around and meet people all day, I show my face in public locations, I subject myself to photographs with friends, some of which later make their way onto the Web – I simply do not live in isolation, as it would not be much of a life to do so. But the prospect of a tablet device that automatically recognizes me has given me a new way of framing this challenge.
Computers are acquiring senses and learning how to use them. They can see (center stage: built-in camera), they can hear (center stage: built-in microphone), and they can even touch (center stage: touch screens). And computers are also learning how to use those senses. Google Voice can transcribe my voicemails, albeit with some errors, but it’s probably doing about as good a job as a young child might. It not only hears, it listens too. Similarly, image recognition programs usher in the era of not only seeing, but interpreting sight.
So perhaps the science fiction of yesterday is closer to reality than most of us think. Soon, maybe in a matter of weeks if the rumors are true, we may sit down to check our email on a computer that knows us personally, in the way a friend does. That computer will in all likelihood be networked, and might very well be easily integrated into a social Web. You can imagine the computer being able to introduce you to its friends out there on the network, giving them the ability to recognize you at some later point in time on the network, and some of these introductions might happen without you realizing it. In the world of human interaction, we generally have some control over who we meet, although since few of us live in isolation that control is limited. Still, we are generally present for and thus aware of situations in which we might be introduced to others. I wonder, how do we give individuals similar presence and awareness with respect to the computers to which they get introduced?
Privacy on the social web is going to largely come down to how well we can translate our human-to-human social cues and build them into the digital system. There are many privacy-related initiatives underfoot that attempt to narrow in on this vision, many of them related to the idea of a “Creative Commons for Privacy.” I’m somewhat skeptical of privacy solutions for the social Web that begin from legal or contractual frameworks, but am a supporter of the general signaling idea. Not that long ago I saw my friend Jonathan Zittrain give a talk in which he pitched the idea of wearing a tee-shirt that asks others to “please not photograph me.” A humorous idea, but perhaps a best effort signaling solution? (It certainly might help me, I instead just dash off to the restroom when cameras appear at dinner parties!) I’ve also written about the nuanced ways in which social signals influence our understanding of privacy in the human-to-human world. To get privacy right in the next decade, we have to do more than simply offer users a choice of “privacy settings” which require constant maintenance and management.
No one seems to have quite figured out how to digitize social signals, and I wonder if that’s because the digital world expects standards. I maintain that privacy is inherently a subjective quality, one that cannot easily be standardized into a set of discrete options. The discrete approach has worked relatively well so far, but as we move forward into a world in which the components of our identity that are beyond our control – for example, our face instead of our chosen username – have been digitized and are generally available for others to recognize, we will need far more intelligent digital signals to communicate and be aware of our privacy choices. I suspect those signals are being developed in research labs somewhere, in all likelihood with a variety of use cases in mind. I just hope the folks developing them have a broad understanding of privacy in the back of their mind as they march, unstoppably, forward.