Drawing on Daniel Dennett’s intentional stance, Christian List advances a pragmatist framework according to which any system—human or artificial—that satisfies three macro-level conditions—intentional agency, alternative possibilities, and causal control—may be ascribed free will, even in the absence of phenomenal consciousness. Employing conceptual analysis and engaging with prominent critiques in the philosophical literature—including Robert Kane’s account of the origination problem, Nancey Murphy and Warren Brown’s critiques of downward causation and emergence, and empirical findings by Eddy Nahmias on the role of P. F. Strawson’s reactive attitudes in free will ascription—this paper argues that List’s framework suffers from a structural ambiguity in distinguishing between “functional autonomy” and “responsibility-grounding free will.” Reducing free will to mere “explanatory sufficiency” not only results in conceptual inflation but also opens the door to “algorithmic responsibility evasion,” whereby developers and institutions may deflect accountability by appealing to the system’s agency. In response, the paper proposes a three-level framework that distinguishes functional autonomy (Level 1), robust free will (Level 2), and moral responsibility (Level 3), and introduces a threshold condition for robust free will: the capacity to revise one’s ultimate ends in light of evaluative commitments. This condition presupposes phenomenal consciousness and P. F. Strawson-style reactive attitudes, thereby clarifying why no contemporary artificial system—even the most advanced language models—possesses robust free will. By resolving the ambiguity in List’s account, the proposed distinction carries important normative implications for AI ethics, legal responsibility, and technology policy.