Eugenic Logic of IQ Testing Resurfaces in AI-Driven Education Systems, Reproducing Racial Hierarchies Under Algorithmic Guise
Original framing: “Why the ugly history of IQ testing matters for AI and education” — The Conversation - Global
Eshu (Yoruba trickster) stands at the crossroads where IQ testing meets AI: he whispers that the same universities selling 'bias audits' for ed-tech algorithms built their endowments on eugenics research. Erasmus's 'Praise of Folly' applies: the folly is believing quantification *is* understanding. Anansi (Akan spider-god) would note the delicious irony — AI trained on human-generated IQ data faithfully reproduces human prejudice, then gets marketed as 'objective.' The trickster doesn't mock; he reveals that the emperor's new algorithm is wearing the emperor's old eugenic robes.
The genealogy from Galton's anthropometric laboratory to Khan Academy's mastery metrics reveals a continuous project: rendering human complexity legible to state and capital through statistical reduction.
IQ testing didn't merely 'influence' AI education — it provided the ontological categories (gifted/remedial, normal/deviant) that make algorithmic sorting intelligible as 'personalization.' The trickster insight is that AI doesn't amplify bias; it *institutionalizes* the epistemic violence of quantification itself. Resistance requires not 'better algorithms' but epistemological disobedience: refusing the premise that intelligence is a scalar property measurable across cultures, histories, and bodies. The 1979 Larry P. decision proved courts can reject psychometric tyranny; the 2024 EU AI Act may prove legislatures can regulate algorithmic classification — but only if movements like the Chicago mothers' coalition and Māori kura networks define what intelligence *is* before technocrats encode what it *isn't*.