Natural Unintelligence: Intellectual Arrogance, Technological Bad Faith, and the Machine We Misname

This essay takes its title from a remark by Justice Cármen Lúcia of Brazil's Supreme Federal Tribunal, who observed that she had begun to worry less about artificial intelligence than about the "natural unintelligence" of those who act without the slightest trace of intelligence. Building on her insight, the essay examines three failures of human intelligence in the presence of artificial systems. The first is the arrogance of the learned: the characteristic error by which mastery in one domain produces overconfidence in another, leading even formidable minds to mistake a statistical process for an awakening self. The second is the bad faith of the industry, in the precise Sartrean sense, an organized not-seeing by which "machine learning" is marketed as "artificial intelligence," inflating wonder for the enthusiast while insulating the seller from responsibility, so that the public cannot consent to what it is not permitted to understand. The third failure belongs to the author and to anyone who writes about these systems at all: the refusal to audit one's own certainty. Drawing on the Dunning-Kruger effect, the Nuremberg findings on the ordinariness of catastrophe, and Sartre's account of bad faith, the essay argues that natural unintelligence is not the absence of intelligence but the failure of intelligence to check itself, a failure as available to the brilliant as to the foolish. It concludes that the danger was never that the machine would think, but that we would stop insisting on understanding, and would hold a tool that flatters that ceasing.