Dropbox has hired a pair of computer vision/machine learning experts as the beginning of a push to aid the organisation, curation and understanding of all “the memories we’re accumulating”. Essentially, there’s a lot of data, but not much information for a machine in a photograph, and providing a machine with access to that information (faces, settings, time of day, etc) means a lot more power for Dropbox.
But what does this mean for users? It means having every photographed moment (or, in Dropbox’s parlance, ‘memory’) analysed and logged – a deep recording of your actions. Not only for Dropbox users, but for those around any Dropbox users wielding a camera, whether you know them or not, being snapped in a crowd by someone who then places that image in Dropbox is going to be enough to place you somewhere, by virtue of your identifiable face. This isn’t necessarily any different to Facebook’s capabilities (and actually, doesn’t necessarily involve the same access to identity) but moving from what one chooses to put on a social network, to, potentially, all of a user’s files and folders is a solid increase in the reach of surveillance.
The mobile phone has been described by Jacob Appelbaum as “a tracking device that lets you make calls”, but it is at least (somewhat) optional (you can choose not to own one, you can turn it off, or place them in a signal blocking pouch), as is your own use of ‘cloud’ services or social networks. Your face, however, is not.
In conjunction with our inability to control what others do with their images and data, and the creation of, essentially, a global network of handheld, networked, surveillance cameras, it becomes difficult to retain anonymity and privacy. We cannot turn off our faces, or choose not to own them, but we can do the equivalent of blocking their signal: cvdazzle is a project looking to ways of fooling facial recognition technology through fashion – with face obscuring hair and makeup, the creation of an “anti-face”.
The face is just one part of this, however; it’s becoming possible to identify people through “intrinsic biometrics”, or how they move. In order to counteract mobile phone based surveillance, cheap pay as you go phones, unlinked to an identity and thrown away after use, are used. These are called burner phones. We need to use this approach for other identity signifiers, we need burner phones and burner faces, burner limbs*, burner voices, in short: whole burner identities.
* Perhaps something like this could help here.