Decolonizing AI at the U.S. border

Lead with the systems being deployed at the US southern border — Anduril sensor towers, CBP's Mobile Surveillance Capability, Palantir-built Investigative Case Management — and which biometric and behavioural classifiers run on whose bodies. The angle: AI 'decision support' at borders is not neutral tooling; the training data, error rates and consequences of false positives all fall asymmetrically on Latin American and Caribbean migrants. Pair with the legal vacuum: there is no federal AI law governing border use, and DHS exempts itself from most algorithmic accountability rules. Reader cares about mechanism, not slogans.