Sponsors (3)
The bill strengthens protections for clinicians to override and report unsafe AI and creates enforcement and transparency mechanisms to improve patient safety, but it also imposes substantial compliance, litigation, and privacy risks and costs for providers, developers, and state actors.
Health-care workers (including non‑employee clinicians) are explicitly protected from materially adverse employment actions when they override or challenge AI/CDSS recommendations and can report safety concerns without retaliation, preserving clinical judgment and encouraging reporting of unsafe AI behavior.
Providers and developers get federally defined terms for AI clinical decision support (e.g., AI/CDSS output, override, override data), which clarifies compliance boundaries and reduces regulatory uncertainty about what systems and behaviors the law covers.
Clinicians retain final authority to override AI recommendations to protect patients and comply with law, affirming clinical judgment and patient safety as primary drivers of care decisions.
Hospitals, clinics, health plans, and other covered entities will face substantial new administrative and operational costs (tracking and storing override data, policies, committees, training, reporting and recordkeeping) that could strain budgets and shift resources from care.
Covered entities face increased litigation exposure and potentially large financial liabilities (civil monetary penalties, statutory damages, treble damages, and attorneys' fees), which may raise insurance costs and deter adoption of AI tools.
Override logs and 'override data' can be disclosed in malpractice, negligence, or law‑enforcement actions, exposing clinicians to legal risk despite statutory workplace protections.
Based on analysis of 9 sections of legislative text.
Requires entities using AI clinical decision systems to preserve clinician judgment, allow timely overrides, protect clinicians from retaliation, limit identifying override data sharing, and set enforcement and training.
Introduced October 9, 2025 by Edward John Markey · Last progress October 9, 2025
Requires health-care employers and other entities that use AI clinical decision support tools to preserve clinicians’ independent judgment, allow timely clinician overrides of AI recommendations, and protect clinicians from retaliation for appropriate overrides or for reporting concerns. It also limits sharing of clinician-identifying override data, requires training and oversight (including a multidisciplinary committee and reporting), assigns federal enforcement responsibilities to HHS (for AI-use rules) and the Department of Labor (for workplace retaliation), and gives State attorneys general limited authority to sue to enforce compliance.