You Think, Therefore I Am — Summary
When Margaret applied for a credit card, a system had already decided who she was — from her zip code, her age, her browser, her search three days earlier for “best credit cards for seniors.” It constructed a probability distribution, a predicted behavior, a risk score. Not Margaret as she knows herself: a woman who has never missed a payment in forty years, who keeps a handwritten expense ledger, who treats her obligations with something close to reverence. That Margaret is invisible. The interest rate she received reflected the system’s Margaret. She will never know the difference.
Descartes: I think, therefore I am. The foundation of identity is the self’s own consciousness authenticating itself from the inside. Something has been quietly inverted. The systems mediating Margaret’s access to credit, healthcare, and insurance do not care what Margaret thinks. They think about Margaret. And what they think becomes, functionally, what she is. You think, therefore I am — where “you” is the algorithm.
Algorithmic identity is actuarial logic generalized across the entire economic landscape. Hiring algorithms score candidates based on statistical predictors from previous hires. Lending algorithms assess risk based on variables correlating with historical default. Your economic identity is not earned through your actions but assigned through your resemblance to a population. For James, twenty-three, Black, from a zip code the model associates with high claim rates — a man who has never missed a payment, works two jobs, saves — the model does not see James. It sees the zip code, the demographic category. His value is determined before he acts, by a system that has already decided what his actions will be. This is the “I AM NOT AVERAGE” problem writ large.
The loop compounds: Margaret’s credit score affects her insurance rate, which affects her healthcare options, which shapes her medication adherence, which generates the data informing her next risk assessment. Each turn narrows the institutional version of her, making it harder for the actual Margaret to be seen. The habitus has been automated — removed from the domain of human negotiation where it might, slowly, be transformed.
The deepest distortion is simple: no system that models Margaret has ever asked her the irreducibly open question every genuine human encounter begins with. Who are you? Not “what category do you belong to?” but the question that invites a response no model could predict. The threat of algorithmic identity is not that it is worse than the messy human process of recognition. It is that it is faster, cheaper, and scalable — and in being all of these things, it renders the original process obsolete. Not because the original was inferior. Because no one can afford it anymore.