When Everything Becomes Visible: Privacy, Power and Digital Identity

Much of what we now spend enormous effort protecting exists because the systems around us require it to exist. Identity records, financial histories, location data, medical files, employment records and administrative identifiers are artefacts of a world that needs people to remain identifiable across time. Cybersecurity then protects the informational machinery required to make that machinery work. There is nothing necessarily sinister about this. Complex societies require persistent distinctions. A hospital really does need to know which patient is which, a bank should probably distinguish your account from mine, and a government cannot administer rights or obligations while continually forgetting who everybody is. The problem is not that the machinery exists. The interesting question is what happens when we become dependent upon representations of ourselves that the machinery requires in order to see us.

Much of “our data”, in this sense, was never entirely ours. Credit histories, browser fingerprints, government identifiers and institutional profiles are generated through our relations with other systems. Increasingly, these representations are also the versions of us upon which those systems act. Credit is extended to the record. Access is granted to the credential. Risk attaches to the profile. The informational artefact begins as a description and quietly acquires causal force. If it is wrong, the human being standing in front of the system can become strangely irrelevant because the system can recognise them only through another representation. Anyone who has ever attempted to explain to a database that the database is mistaken will recognise the philosophical depth of the problem.

Something else is happening, however. Legibility is becoming reciprocal, although emphatically not symmetrical. Institutions acquired extraordinary power by making populations visible to themselves: identifying, recording, categorising and correlating people across time. Their advantage was never simply that they could see more. It was that they could make what they saw consequential. A database entry can affect employment, credit, insurance, taxation, mobility or access while the person represented may know almost nothing about the process producing the decision. But observation leaves residue. Decisions generate records; organisations generate communications; transactions expose relationships; policies can be compared with outcomes; statements persist long enough to be compared with later statements. Connect enough of these traces and institutions acquire informational doubles of their own. Governments, corporations and universities become increasingly reconstructable from the residue of what they actually do. The architecture built to make the world legible is discovering the mildly inconvenient fact that the world can look back.

This does not abolish the asymmetry of power. Seeing an institution is not the same thing as possessing institutional power over it. Institutions possess money, law, technical infrastructure, organisational continuity and armies of people paid to manage precisely these problems. Individuals generally possess a password they have forgotten and a privacy policy nobody has read. But the relation has changed. The observed can increasingly observe the observer, while the observer retains far greater capacity to convert observation into consequence. At the same time, privacy becomes more valuable precisely as it becomes more difficult to possess. Visibility creates demand for concealment; concealment requires systems of protection; those systems produce boundaries, records and observable behaviour that can themselves reveal what is being protected. Near-universal legibility combined with radically unequal capacities to exploit, resist or escape it does not look like a stable destination.

The problem, then, runs deeper than privacy. Complex systems cannot encounter whole human beings, organisations or societies. They encounter differences that can persist: identifiers, categories, records, relationships and patterns that can be stored, transmitted, compared and acted upon. Power lies partly in deciding which of these representations counts and what consequences follow. But there is something about a person that should remain unavailable to this process: not a missing data point waiting for a better sensor, but the possibility of remaining partly unknown, contradictory, unfinished and beyond administrative resolution. A civilisation organised around ever greater legibility will tend to interpret whatever it cannot classify as a problem to be solved. That may eventually be the problem. The observer is becoming observable, the machinery is becoming visible, and everyone is discovering that complete transparency is considerably less attractive when it works both ways. Privacy, in the final sense, may not be about hiding information at all. It may be the name we give to the remaining distance between a human being and everything the world believes it knows about them.