<![CDATA[Accounting firms increasingly report that AI adoption improves measurable performance, even as separate evidence points to a decline in the judgment capacity that professional licensure presupposes. This article develops a conceptual explanation for why both can be true at once. Following a theory-synthesis method, it integrates five theoretical traditions, human-AI symbiosis, virtue ethics, institutional theory, governmentality, and sustainability accounting, into a single mechanism. It introduces legitimacy debt, the accumulated gap between the judgment institutions still recognize as belonging to accountants and the judgment those accountants can still exercise, and presents a revised SIGMA model specifying the conditions under which this debt forms, stays concealed, and is eventually repaid. The article traces the mechanism through developmental displacement, in which task substitution erodes the practice that builds professional judgment, and through miscalibration under algorithmic visibility, in which both excessive trust and excessive discounting of AI output produce the same failure of proportionate judgment. Six falsifiable propositions link these mechanisms across individual, organizational, and profession-wide levels, with Indonesia examined as a boundary condition that reveals the double-edged role of collective deliberation. The article offers accounting research a testable construct for a phenomenon current frameworks describe only partially, and reframes integrity as a designable organizational infrastructure rather than an individual disposition, with direct implications for accounting education, firm practice, professional standards, and regulation.]]>
