This paper challenges a familiar diagnosis of online misinformation: that citizens are misled because they are gullible, irrational, or insufficiently vigilant. Drawing on Mercier and Sperber’s argumentative theory of reason and Mercier’s account of epistemic vigilance, I argue that human beings possess real capacities for suspicion, source evaluation, coherence checking, and resistance to deception. In the age of AI-assisted disinformation, however, the problem is not the absence of vigilance, but its social and infrastructural vulnerability. Vigilance depends on socially supplied cues of trustworthiness, expertise, reputation, salience, and credibility. I call the resulting failure captured vigilance. In AI-assisted and platform-mediated environments, users may become more suspicious, investigative, and committed to “doing their own research,” while their suspicion is redirected toward reliable institutions and their trust is routed toward pseudo-experts, in-group authorities, and identity-confirming sources. Conspiracy thinking illustrates the problem: vigilance is present, but organized within self-sealing trust environments where warrant is inverted, correction is absorbed as confirmation, and standing is redistributed according to loyalty and distrust. The paper’s social-epistemological contribution is to show that vigilance is not merely an individual epistemic capacity, but an environmentally scaffolded practice vulnerable to capture under AI-assisted, platform-mediated conditions.
