Bad thinking spreads—between people, between humans and AI, and in self‑reinforcing loops. This essay introduces cognitive contagion, the proposed β_NPF coefficient, and then presents practical defences: the Binary Belief Protocol, Proportional Scrutiny Matrix, and three mechanisms (prebunking, cross‑training, dopamine rechanneling). All are offered as hypotheses to try.

Cognitive contagion formalises how entrenched reasoning patterns spread between humans and AI. This paper introduces the transmission coefficient β_NPF (exposure × susceptibility × content potency), analyses contagion dynamics (human→AI, AI→human, reinforcing loops), and explores societal vectors like algorithmic entrenchment. Case studies include vaccine misinformation and financial market fragility. The model is a hypothesis awaiting validation.