Pre-Inference Governance in Medical Robotics: Preventing Unauthorized Autonomous Action in Surgical and Assistive Systems

The rapid advancement of medical robotics—particularly in autonomous and semi-autonomous surgical systems—has shifted clinical risk from mechanical failure to epistemic misalignment. While robotic systems increasingly demonstrate high procedural precision, they also introduce a critical governance gap: the translation of probabilistic inference into irreversible physical action without sufficient epistemic authorization. This paper introduces Pre-Inference Governance in Medical Robotics as a domain-specific extension of Pre-Inference Responsibility, building on the Akarkach 7-Stage Emergence Model (A7SEM) and Action-Specific Oversight (ASiSO). The central argument is that most critical incidents in medical robotics do not arise from hardware malfunction or algorithmic error, but from epistemically immature situational assessments being allowed to authorize physical intervention. Using concrete scenarios from robotic surgery, assistive medical devices, and adaptive intra-operative systems, this paper demonstrates how stage-aware epistemic gating, action-specific authorization, and veto architectures can structurally prevent unintended harm, preserve human authority, and enable audit-ready, insurable robotic systems. The proposed framework reframes medical robotics governance from post-incident safety controls toward ex-ante authorization architecture, enabling safer deployment, clearer accountability, and legally defensible autonomy boundaries.