<p>This essay has three interrelated aims. First, it articulates a set of capacities I call virtues of attention—capacities for intuitive discernment, good judgment about what is worth sustained focus, and the ability to engage deeply with worthwhile things. These are eudaimonist virtues in that they help us live well, not merely act rightly. Second, the essay explores what I call poisonous persuasion: the idea that rhetorical appeals, especially those used in marketing, may not only succeed by appealing to certain desires or habits of mind but may also deepen and entrench them. Third, I bring these insights together to examine a largely neglected problem at the intersection of AI and virtue ethics: how AI-aided marketing may erode the very capacities we need to flourish.</p><p>This is not primarily an essay about how AI might itself be made virtuous, nor whether it can mimic or model virtue, though those are important questions. Rather, the focus is on what AI-driven persuasion may be doing to us. The danger is not just distraction, which has been well-documented, but a subtler debilitation: an incremental erosion of the attentional capacities on which flourishing depends. These systems appeal to what works—what clicks, what sells—without regard for what they may be nurturing in us. This paper uses virtue ethics to illuminate this harm: not a spectacular or easily measured injury, but a quiet undermining of our ability to recognize, commit to, and engage with the good. If we hope to thrive rather than merely function, we must understand and resist these corrosive dynamics.</p>