AAA-02- Behavior Is Not Agency. Contemporary algorithmic systems increasingly exhibit behavior that resembles human judgment: they recommend, classify, prioritize, optimize, and intervene across social, economic, and institutional domains. In public discourse, this behavioral similarity is frequently treated as evidence of agency or moral standing. This paper argues that such an inference constitutes a fundamental category error. The core claim is simple but structurally decisive: behavior is not agency, and performance is not responsibility. Historically, observable behavior functioned as a reliable proxy for agency because actions were tethered to identifiable human authors who could be questioned, sanctioned, and held accountable. Algorithmic systems sever this tether. They act pervasively while lacking a locus that can stand behind outcomes. This paper develops a structural account of responsibility under algorithmic mediation. It shows how decisions increasingly emerge from distributed pipelines spanning data collection, model design, institutional incentives, and interface constraints—while responsibility remains socially anchored to individuals who no longer control the decisive variables. Attempts to resolve this mismatch by attributing moral status or agency to systems themselves are shown to be a decoy: they displace attention from the institutional reconfiguration of agency rather than addressing it. The central risk diagnosed is not the rise of artificial moral agents, nor the arrival of machine consciousness. It is the normalization of action without ownership—a condition in which decisions persist, consequences bind, yet no agent remains structurally positioned to bear responsibility. The paper situates this failure as a foundational problem for ethics, governance, and political legitimacy in algorithmically mediated societies. AAA-02 establishes the core analytical distinction upon which the Agency in the Age of Algorithms series is built and provides the conceptual groundwork for subsequent analyses of delegation, intent, punishment, and the illusion of choice. Upcoming: PFS-03 – Capital Concentration and the Geography of Innovation (03 Feb) Upcoming: Delegation Is Not Distribution - Agency After Interruptibility - Essay 1 (05 Feb) Upcoming: AAA-03 – Human-AI Hybridity Is Not New. Whats New Is the Disappearance of Accountability (12 Feb)