AI policy proposals being considered in Washington state
The AMA recently sent a letter (PDF) to the Washington Medical Commission (Commission) regarding two proposals on augmented intelligence (AI): an Interpretive Statement on the eligibility for AI to receive a license to practice medicine and a Policy Statement on the use of AI tools in patient care by physicians and other licensees. Both will be considered at an upcoming Commission meeting.
The comment letter from AMA CEO John Whyte, MD, MPH, expressed the AMA’s support of the Commission’s Interpretive Statement clarifying that AI cannot hold a license to practice medicine, noting that this finding is a “straightforward and natural reading of the law and the current Medical Practice Act.” In addition, the AMA commended several patient protections found in the Policy Statement, such as language declaring that “AI can inform but cannot independently and autonomously assess, diagnose, treat, or prescribe medications to patients” and that physicians or other licensees must retain the final authority on patient care, including the ability of physicians to determine when and how to use AI in clinical care.
However, the AMA also expressed strong concern with several provisions in the proposed Policy Statement, most notably language centered on accountability of physicians using AI in clinical care. As drafted, the policy places full responsibility on the physician or other licensee for any error or harm caused by use of an AI tool. This accountability extends to physicians even if the AI tool was the source of the failure, and even in those cases where the tool was embedded in the clinical workflow or mandated by a physician’s employer, health system, or payer.
In the letter, Dr. Whyte points out that this “would treat AI differently than any other medical product used in care delivery.” In addition, the AMA warns that shifting 100% of the responsibility of AI tools on the physician will likely suppress adoption of these tools. As indicated by Dr. Whyte in the letter, “[a] physician who is told that they will bear full responsibility for a diagnostic miss produced by a flawed algorithm they did not know or have reason to know was experiencing a performance issue has a straightforward incentive: use AI tools as little as possible, or not at all.”
AMA policy on the issue of accountability and AI unequivocally states that accountability should be aligned with whatever party is in the best position to know a tool’s risk and to prevent or mitigate harm through design, development, validation and implementation. AMA’s comment letter encouraged the Commission to amend the Policy Statement to align with this framework and to acknowledge that a licensee’s accountability is not absolute where use of the tool was mandated, or where the tool itself malfunctioned and produced an erroneous output despite reasonable clinical review and assessment by the physician or other licensee.
