Synopsis
Palantir did not build a surveillance system. It built a mathematical integration layer that turns every existing surveillance system into a single one. The welfare claimant, the suspected gang associate, the immigration risk, and the benefits fraudster are not different people tracked by different agencies; they are increasingly the same person whose data flows across a common identifier into a unified risk profile. What Palantir sells is not data but the arithmetic of connection — entity resolution, graph traversal, shared scoring. The result is the most complete realisation of this book’s central argument: mathematical representation of a person by state and capital is now total, cross-domain, and effectively beyond appeal. Palantir was seeded by In-Q-Tel, the CIA’s venture capital arm, and its entire design philosophy — connect everything, surface hidden relationships, treat all data as potentially relevant — was forged in counterterrorism and sold to city police forces, welfare agencies, and NHS trusts as though the context were transferable. The Metropolitan Police’s Gangs Matrix, 80% Black by documented count, flows into housing decisions, school exclusions, and welfare risk flags. Palantir’s £330 million NHS contract deploys an identical architecture across patient records. The panopticon Bentham imagined was architectural. This one is algebraic. And it connects the benefit claimant who was late submitting a digital report to the policing database of the officer who stopped them at the bus stop.
In This Chapter
- How Palantir’s entity resolution and graph traversal architecture enables the collapse of administrative boundaries between welfare, policing, immigration, and health into a single mathematical space where any flag in any domain becomes visible in all others
- How the CIA’s counterterrorism logic — all data potentially relevant, all connections presumptively suspicious — was carried wholesale into civilian welfare and health contracts, importing the epistemology of the intelligence apparatus into the administration of poverty
- How the Metropolitan Police’s Gangs Matrix demonstrates Harcourt’s ratchet effect and Richardson’s dirty data thesis operating simultaneously: biased surveillance data generates biased designations which justify further biased surveillance, with the mathematical apparatus presenting each step as objective
- How Palantir’s NHS Foundry contract extends the same architecture into patient records, creating an integrated data infrastructure whose governance gap is acute and whose clinical justification does not prevent its eventual use for non-clinical purposes
- How the welfare-to-policing data pipeline realises Eubanks’s “digital poorhouse” thesis most completely: the same populations targeted by algorithmic welfare are targeted by predictive policing, their data completing an automated loop of marginalisation
Connection Forward
Chapter 12 steps back from Palantir’s specific architecture to examine the credit score — the longest-running and most pervasive algorithmic poverty measure in operation — as the private-sector parallel to the public systems of Part III, and as the infrastructure on which financial exclusion compounds every other form of state-administered deprivation.