AI Coverage Decisions Need Physician Review, New AMA Policies Assert
By Alisa Pierce

To protect physicians and patients from unchecked artificial intelligence (AI), the American Medical Association’s House of Delegates adopted new policies this summer that underscore physician oversight in insurer and clinical processes.

As AMA asserts health plans increasingly use AI in coverage decisions, the association’s new policies oppose the use of “autonomous or semiautonomous AI systems” as substitutes for physician review. The policies also call for regulations requiring any use of AI technology in health care to be physician-led.

“Patients deserve care decisions that are informed by the latest medical evidence and guided by a physician who understands their individual needs,” said AMA CEO John Whyte, MD, in a press release. “Whether AI is helping a physician make a clinical decision or assisting with an insurance review, there must always be transparency, accountability, and meaningful physician oversight. Technology should support better care – not stand between patients and the care they need.”

In 2025, the Texas Medical Association successfully backed Senate Bill 815, authored by Sen. Charles Schwertner, MD (R-Georgetown), which prevents payers and their agents from using algorithms or AI in the utilization review process to “make, wholly or partly, an adverse determination.”

In other words, a payer’s determination that care is not medically necessary, appropriate, or is experimental or investigational, must be made by an appropriately trained and qualified human reviewer. SB 815 also grants new authority to the Texas Department of Insurance to audit health plans’ use of algorithms or AI systems in utilization review. 

TMA approved its own AI policy at TexMed this year, advocating against the use of AI as an independent diagnostic or prescriptive tool, or as a care management substitute for a physician.

TMA’s policy further asked the Texas Delegation to the AMA to bring forward a resolution promoting similar initiatives at the federal level – which led AMA to adopt policy calling to advocate for legislation and regulation of AI in health care, including for diagnostics, prescriptions, care management, or other functions. The AMA policy states such tools must: 

• Integrate with a physician-led team and be used at the direction of the treating physician;
• Respect continuity of care and best practices related to transitions of care;
• Have transparent, auditable data that demonstrate the tool’s safety and efficacy;
• Be subject to relevant and appropriate regulations, including but not limited to those related to liability and documentation; and
• Adhere to AMA’s policy on augmented intelligence in health care.

“The long-term ramifications of having autonomous agents without a human in the loop – to replace the decision-making of a physician with AI – is the breakdown of the patient-physician relationship,” said Philip Bernard, MD, chair of TMA’s Committee on Health Information Technology and Augmented Intelligence. “That’s why TMA leadership in this area is exceptional. We have to make sure that there is a physician who is actively engaged in any kind of clinical decision making.”

AMA says it will call for heightened transparency when AI is used in prior authorization decisions, including advocating for health plans to disclose clinical logic, data sources, and guidelines used to reach adverse determinations. 

AMA also adopted policy calling on the association to establish standards for AI clinical decision support tools by working with medical specialty societies, regulators, AI developers, and other stakeholders. The AMA says these standards, such as requiring AI tools to feature evidence attribution and explainability, could help physicians better understand, evaluate, and trust AI.

“While AI-enabled tools may improve efficiency and help synthesize large amounts of information, the AMA notes that important concerns remain regarding transparency, bias, explainability, and long-term impact on physician practice and patient outcomes,” AMA said in the press release.

For more information about AI and TMA’s related resources, see the association’s AI webpage