The AMA is drawing a clear line between augmented intelligence (AI) as a tool to support healthcare and the practice of medicine. AI—often called artificial intelligence—can help doctors with clinical tasks, but it cannot be granted physician licensure, professional titles or certification to practice medicine, says new policy adopted at the 2026 AMA Interim Meeting.
With this recent update to extensive, existing policy on AI, the AMA House of Delegates calls for AI-enabled technologies to be integrated into physician-led teams and used at the direction of the treating physician. AI tools used for clinical functions such as diagnosis, prescribing and care management should have transparent, auditable evidence demonstrating their safety and efficacy and be subject to appropriate regulation, documentation and liability requirements.
The AMA will advocate legislation establishing a framework for physician oversight of AI clinical decision-making, including protections for physicians' independent clinical judgment and against liability for choosing not to use an AI tool or follow an AI recommendation.
“AI can help physicians analyze information, identify patterns and support clinical decision-making, but patients deserve to know who is responsible for their care,” said AMA President Willie Underwood III, MD, MSc, MPH. “Physicians also need the authority, oversight and liability protections to make independent decisions when AI is part of the care team.”
The policy also calls for model legislation to distinguish authorization of a medical technology from professional licensure and prohibit misrepresenting nonhuman entities with titles reserved for physicians. The AMA will advocate for safeguards to prevent AI developers from using patients' clinical data for proprietary or research purposes without explicit authorization.
From AI implementation to digital health adoption and EHR usability, the AMA is fighting to make technology work for physicians, ensuring that it is an asset to doctors and the patients they care for. That includes the AMA Center for Digital Health and AI, which works to ensure physicians help shape how AI is developed, implemented and regulated across the healthcare system, with patient safety and physician-led care remaining at the center of those efforts.
Hold clinical AI vendors liable for design defects
The Food and Drug Administration (FDA) has authorized more than 1,450 AI-enabled medical devices as of December 2025 and—no different from say, from an inherent problem in an MRI machine—physicians should not bear responsibility for defects.
Established product-liability law already holds that when a medical device fails because of a defect in the design, manufacturing or warnings, the manufacturer bears liability for the resulting patient harm—not the physician. Physicians who use AI tools as intended, within the scope of the tool’s marketed and cleared indication and who exercise appropriate clinical judgment in conjunction should be treated similarly.
To ensure that physicians aren’t wrongly held liable for AI tool failures, the House of Delegates directed the AMA to advocate:
- Federal and state legislation establishing that when a clinical AI product produces an erroneous output and a physician has used the product as intended while exercising appropriate clinical judgment, the liability for resulting patient harm should rest with the AI vendor, manufacturer, or distributor, not a medical malpractice claim against the physician.
- That AI vendors and manufacturers be held to the same product liability standards as manufacturers of other medical devices including but not limited to liability for design defects and failure to provide adequate instructions or warnings.
- Legislation requiring that AI vendors and manufacturers maintain product liability insurance adequate to cover claims arising from AI product failures in clinical settings, and that such insurance obligations be a condition of FDA marketing authorization required by law as a condition of employment.
- Legislation prohibiting contractual provisions, including end-user license agreements and vendor contracts, that attempt to shift liability for AI tool failures from the manufacturer to the physician or health system end user.
- Federal legislation requiring AI vendors and manufacturers to disclose known failure rates, limitations, and performance data for their clinical AI products, including performance across different patient populations, so that physicians can make informed decisions about the clinical appropriateness of AI tool outputs.
- That when a health system, hospital, payer or other entity mandates the use of a specific AI tool and that tool produces an error resulting in patient harm, the mandating entity and the AI vendor bear the liability, and the physician who was required to use the tool and exercised appropriate clinical judgment is not liable for errors originating in the mandated product.
Last month, the AMA released a framework for aligning liability with each party’s authority and ability to prevent or mitigate harm.
How AI’s used in credentialing, peer review and more
Transparency and physician-led oversight are needed as hospitals and health systems increasingly incorporate augmented intelligence (AI) into administrative and governance workflows, as well as emerging applications that support credentialing, privileging, peer review, quality review and other professional practices evaluation processes, says a separate resolution that was introduced at the 2026 AMA Interim Meeting.
Existing AMA policy says that physicians who report patient-safety concerns must be protected from retaliation. In addition, physicians must have the ability to access, understand and, when appropriate, challenge the data, methodology and rationale underlying the use of AI within administrative governance workflows. That’s especially needed when the information informs credentialing, privileging, peer review, quality review or any adverse medical staff action.
With that in mind, delegates adopted new policy to “encourage appropriate organizations to require physician-led oversight of AI when used in credentialing, privileging, peer review and quality review.”
In addition, the AMA will:
- Advocate that any use of AI in credentialing, privileging, peer review, quality review or corrective action requires physician-led oversight and transparency regarding its use, function and role in decision-making.
- Encourage the requirement of hospital and health systems to fully disclose the data sources, validation, limitations, governance, and intended use of any augmented intelligence tool used in medical staff processes.
Protecting minors from AI deepfakes
As AI is becoming more realistic with no federal guidelines to regulate its use, delegates also directed the AMA to oppose the technology’s being used to produce child and adolescent sexual exploitative material.
At least 1.2 million children across 11 countries have had their images manipulated into sexually explicit deepfakes using AI. In the U.S., 55.3% of adolescents have reported creating—and 54.4% have reported receiving—at least one AI-generated nude image. And 36.3% of American adolescents said they experienced having a nonconsensual AI-generated image created and 33.2 experienced having at least one AI-generated image nonconsensually shared.
With AI-generated deepfakes directly harming children and potentially severely impacting their mental health and jeopardizing their safety if perpetrators recognize them, the House of Delegates adopted new policy to:
- Oppose the use of AI to produce any sexual and other exploitative material of minors.
- Advocate legislation to prohibit the use of artificial intelligence for such purposes.
- Encourage state delegations to adopt similar resolutions opposing the use of artificial intelligence to produce any sexual and other exploitative material of minors and advocate on this topic at the state level.
Read about the other highlights from the 2026 AMA Interim Meeting.
