With Facebooks recent policy flaws in terms of privacy and the trust this company has gambled on, the question of identity and identity providers has reached new importance.
Jeff Jarvis has nice piece on that, where he basically pinpoints the two biggest issues right now: how can I manage information that is available about myself, and how can I prove that I am indeed the one I say I am?
Managing identity is crucial, but it not only happens around the web. It happens with information we gather about ourselves, it happens with data that is gathered about ourselves, identity is much bigger than the web. Identity is my risk-profile while driving, information that can be gathered utilising my GPS and the accelerometers built into basically every modern car. Now with platforms like , this information can be discerned and analysed. This information is part of my identity, this information is way more valuable than the menial status updates we currently consider when we are talking about facebook and privacy. But when we are talking about identity providers, or identity providing systems for that matter, we regularly fail at taking into consideration these aspects of our identity. Numbers, that seem to say nothing to us – yet.
When we are talking privacy controls, when we are talking identity, we need to realise that we are not only talking about our profile pictures (which to remove in protest indeed creates powerful imagery, but actually helps nothing at all) or our social graph (now this gets interesting) but our everyday behaviour that shows in the numbers.
In Eastern Standard Tribe, a very interesting read indeed, Cory Doctorow approaches this issue in two or three paragraphs when he let’s his main character fantasize over possible user experience improvements of mental health institutions. In this scenario, the main character thinks of public prescription stats of mental health professionals, and more to the point, tracking of patients by means of their badges, measuring proximity to other patients, mobility, cross-referencing that with their therapy and group plans and subsequently tracking their progress. We are doing this already, albeit in a different context. Visa recently came forward with a press release in which it disclosed that the company was now able to tell whether a couple was about to get a divorce, just based on their purchase history and some statistical modelling.
It’s this data that we really ought to keep an eye on. it’s this data that in the long run really matters. My social graph probably tells a whole lot about me, after all, the saying is that if you tell me your friends I indeed know who you are. But there is way more data available about who I am if we use standardly available statistics. Efficiency 2.0, one of the hottest start-ups in the energy efficiency field, manages to provide custom tailored recommendations on how to improve energy efficiency in the most economical way using not only my usage profile which is derived from a smart meter but by additionally taking into account my demographic data: which neighbourhood do I live in (rich or poor, inner city or suburbs?), my age (a student might have a shorter time span for ROI) etc. Using these kinds of statistical modelling, Efficiency 2.0 is able to provide surprisingly valid feedback on how to curb energy consumption.
Similar data is used in methods called score modelling. These methods are used by rating agencies trying to provide credit ratings for individual customers. These models not only take into account my personal credit history, the amount of money I have or owe, but further use aggregate demographic data: how likely are people that live in my area to default? how likely are people that moved as often as I did to default? in short: they are creating risk profiles based on statistical models which include personal and aggregate data.
On the other hand, in opening up this personal data lie tremendous opportunities. By using anonymised aggregate data, banking service mint.com is able to give recommendations for financial planning based on comparisons with other users, basically opening a pool of experience. Similarly, Efficiency 2.0 might learn to give better recommendations if certain projections just don’t work out.
So I think what I want to say is this: we need to discuss Facebook’s privacy issues, because they matter to most users. But we may not limit ourselves to thinking that Facebook’s privacy issues are all that matters, because identity is not only what is portrayed by our status updates there. We need to think about systems that allow us to take some amount of control over the data that comprise our identity by capturing what we really do.