Public concern about ‘big data’ frequently comes down to a vague and ill-defined sense of ‘ickiness.’ I’d like to briefly suggest a way to provides structure to this vague sentiment — let’s call it data dread. Provisionally, I would argue that public distrust of ‘big data’ comes down to major tension between two promises of the digital age. On the one hand, as Floridi notes, the advent of social media represents an “unprecedented opportunity to be more in charge of our social selves, to chose more flexible who the other people are whose thoughts and interactions create our social personality” (The Fourth Revolution, p. 64). In other words, the modern internet allows us not only to more carefully craft our identities, but also to more carefully curate our communities so that our self-representations are more likely to be recognized and accepted. We now have an unparalleled ability to ‘make ourselves’ in a way that resonates, not just with the existentialist philosophies of the 20th century, but also with Renaissance conceptions of man as infinitely fluid and self-determining. This is the first promise of the digital age.
On the other hand, however, the very technologies that allow us to make ourselves also — and necessarily — produce digital traces, on the basis of which it becomes possible for individuals to be tracked, and identities co-opted, The digital traces that people leave behind as a result of their efforts to forge their digital identities can also be used by algorithms to produce identities for which individuals are not themselves responsible, but that nevertheless have real effects. This is not necessarily a bad thing. In fact, this amounts to a second promise of the digital age: the promise of personalization. People want their experience of the world to be personalized. Who doesn’t want the world to revolve around them? But the problem with the algorithms that personalize experience is that they don’t actually care who you are or how you wish to be recognized. They don’t care about the identity you wish to construct. In personalizing our experiences of the world, algorithms also have a very real effect on our self-perceptions. In order to personalize our experiences, they must first personalize us.
The two promises of the digital age, then, are these: (1) a promise that individuals are free to construct themselves in whatever was they choose, and (2) a promise that digital traces will be used to personalize individual experiences, not just online, but in the ‘real world’ as well. Our experience of dread comes from the fact that we are simultaneously promised an ability to make ourselves at the same time as we are promised that our selves will be algorithmically made on our behalf. At the same time as we represent ourselves in such a way as to be recognized and acknowledged by others in the ‘real world,’ algorithms are representing us to ourselves and others, making judgements about who we are and what we want, and intervening in our lives through nudges, recommendations, and other automated events.
For Hegel, people come to recognize themselves as selves through two basic sources: labor and others. I know that I am a self because I can recognize myself in the work of my hands, and I know that I am a self because of the fact that others relate to me as such. In the former case, I am in complete control over the self I create. In the latter, the self that I create is a function of the a negotiation. Now, for the first time, it has become possible for people to be entirely — and voluntary — excluded from the processes by which their identities are constituted. In the social world, our identities and the consequences of our behaviors may not be our own, but they are nonetheless a function of a kind of ongoing negotiation with other people. When machines start to make judgements about us, any sense of negotiation is lost.
Perhaps our sense of ‘ickiness’ in the face of big data — our sense of data dread — is not a function of data itself. Perhaps it is a function of the fact that there is a contradiction at the heart of so-called ‘social media’ — what Floridi calls Information and Communication Technologies (ICT). It is not just that digital traces from ICTs make it possible for identities to be algorithmically co-opted. The algorithmic co-option of personal identity is a necessary feature of modern-day social media technologies in the absence of which those technologies would cease to function.