Does the social penalty for AI use fade as the tool becomes ordinary?

The authors' attribution account makes the penalty depend on novelty, which turns a finding about the present into a question about time. They write that "behaviors that are not considered mainstream—or part of the consensus—are especially likely to elicit dispositional attributions," and that "emerging technologies are, by definition, new, and therefore, their use is unlikely to be perceived as customary." From this they predict that the use of such technologies "is especially likely to evoke negative dispositional inferences about their operators." The excerpt does not say whether the same logic runs in reverse once a tool is customary, though its wording invites that reading.

The excerpt also supplies history that cuts both ways. The doubt about help is old: the notion "has echoed in debates over new tools for centuries," from Plato's question about whether writing would ever yield true wisdom to educators' worries about calculators and studies of patients who "assume that physicians who use diagnostic aids are less capable." Those cases show the judgment recurring across tools, but the excerpt does not say whether any of those judgments faded once the tools became ordinary. The authors also argue that AI is a different case, because it "may be perceived as more agentic" than earlier tools that "simply performed specific operations or made predictions." If agency rather than novelty drives the judgment, familiarity may not remove it. The excerpt gives no way to choose between these readings.

The anticipated penalty in Do people fear judgment when they use AI at work? is measured at a single set of comparisons, so it cannot show the penalty's trajectory. The question matters for adoption because, if the penalty tracks evaluation norms rather than the tool itself, its duration becomes an institutional question. What makes accountable judgment scarce when AI cognition is cheap? argues that institutions, not raw capability, decide labor-market outcomes. Connecting the two is my extension; the excerpt does not make that link.

The excerpt reports no measures at more than one point in time, and it does not report how common AI use was among the people doing the judging. Any claim that the penalty is temporary, or that it will harden into a lasting barrier, needs data it does not contain. What it supports is narrower: the authors expect the penalty to attach to tools that are new and agentic, and the history they cite shows a similar doubt about earlier tools. A design that varied how customary a tool was seen to be would test the attribution account directly; the excerpt reports no such manipulation. Until one exists, the duration of the penalty stays open.