AMA Publishes Guidance on AI Development, Deployment, Use

Yesterday, November 28, 2023, the American Medical Association (AMA) announced its Principles for Augmented Intelligence Development, Deployment, and Use. The document provides a roadmap for the ethical, equitable, responsible, and transparent integration of augmented intelligence — also called artificial intelligence (AI) — in healthcare.

According to the AMA, regulatory limitations exist for monitoring AI integrations in healthcare. Although the United States Food and Drug Administration (FDA) regulates AI-enabled medical devices, many AI tools in healthcare fall outside of the administration’s purview.

AI tools with clinical applications that may guide clinical decision-making are often loosely regulated by the Federal Trade Commission (FTC) and the Department of Health and Human Services (HHS) Office for Civil Rights. However, the regulatory authorities of these organizations are limited and cannot acknowledge the nuances of AI applications in healthcare.

Understanding these limitations, the AMA principles aim to address the gaps left by existing regulatory policies. The administration hopes that standardizing AI practices in healthcare will lead to more structured governance and responsible AI integrations.

“The AMA recognizes the immense potential of healthcare AI in enhancing diagnostic accuracy, treatment outcomes, and patient care,” said AMA President Jesse M. Ehrenfeld, MD, MPH, in the organization’s press release. “However, this transformative power comes with ethical considerations and potential risks that demand a proactive and principled approach to the oversight and governance of healthcare AI. The new AMA principles will guide the organization’s engagement with the administration, Congress, and industry stakeholders in discussions on the future of governance policies to regulate the development, deployment, and use of healthcare AI.”

The AMA press release highlights seven critical components of the document, including the following:

• Oversight
• Transparency
• Disclosures and documentation
• Generative AI
• Privacy and security
• Bias mitigation
• Liability

Although the AMA emphasizes the need for national governance focused on AI in healthcare, through the document, it also acknowledges the critical importance of AI oversight by non-government organizations.

The emphasis on transparency, disclosures, and documentation go hand in hand, noting that AI characteristics should be made available whenever possible and disclosed when applied to patient care.

Privacy and security are among the most significant components highlighted in the document. Discussions of privacy and security measures are paramount in AI. When AI accesses sensitive patient information, the technology must have safeguards to prevent security breaches.

Other advocacy efforts supported by these principles include policies to manage the risks of Generative AI, mitigate bias in AI algorithms, and limit physician liabilities.

Although these principles are not exhaustive, they begin to outline a more structured guidance for AI applications in healthcare.