The policy framework, developed by the AMA Center for Digital Health and AI, establishes enforceable protections against unauthorized AI-generated deepfakes impersonating physicians, AMA officials said. The framework is designed to modernize identity protections for physicians and close existing legal gaps that have allowed deceptive AI content to proliferate, according to the association. It specifically addresses the escalating risk posed by manipulated images, videos, and audio that can be used to impersonate medical professionals.

The framework mandates a prohibition on deceptive medical impersonation without clear, informed consent.

According to AMA representatives, deepfakes have been maliciously employed to impersonate doctors and disseminate false endorsements of unproven treatments, posing significant threats to individual patients and the healthcare system at large. The association noted that such impersonation scams undermine patient-physician relationships and erode public confidence in evidence-based care, increasing the risk of medical harm through deception.

AMA officials said that any AI-generated or altered content impersonating physicians must be explicitly prohibited when used to mislead patients by falsely conveying endorsement or authorship. The framework further establishes that failure to provide clear disclosure of synthetic content will be considered evidence of deception. These provisions set enforceable legal standards to deter unauthorized impersonation.

Consent requirements are a key component of the policy. The AMA stipulates that the use of a physician’s identity in AI-created or manipulated content requires separate, explicit opt-in consent that is never implied or bundled within general agreements. Consent must specify the intended use, audience, purpose, and duration, and must be revocable if circumstances change, according to AMA officials. This informed authorization is intended to protect physician identity rights while allowing legitimate uses of synthetic content.

To ensure transparency, the framework requires mandatory labeling of all AI-generated or altered depictions of physicians in plain language. Digital watermarks must be embedded on all synthetic physician content, and patients must be proactively notified before any interaction with synthetic professionals, the AMA said. These labeling and disclosure requirements aim to prevent patient deception and facilitate informed decision-making.

The AMA also calls for shared responsibility among platforms, hospitals, and AI vendors to prevent impersonation. The framework requires the implementation of rapid takedown mechanisms for unauthorized deepfakes and conspicuous labeling across all distribution platforms. It prohibits the use of health professional titles in AI content without authorization. The association emphasizes a multi-stakeholder approach to address deepfake threats throughout the healthcare ecosystem.

Enforcement provisions include designation of a federal agency with explicit authority to investigate violations of deepfake impersonation rules. The framework mandates preservation of relevant records and audit logs to support enforcement actions. It also calls for mechanisms to compel cooperation from violators, according to AMA officials. The association is collaborating with lawmakers, regulators, and industry partners to implement the framework and ensure regulatory oversight.

The AMA’s announcement follows growing concerns over the misuse of AI technologies in healthcare communications. The association said the framework represents a proactive step to safeguard patient safety, uphold professional integrity, and maintain public trust amid rapid advances in synthetic media. The AMA plans to continue working with federal authorities and industry stakeholders to refine and enforce these policies as AI technology evolves.

The policy framework, developed by the AMA Center for Digital Health and AI, seeks to modernize protections for physician identity amid rising concerns over AI-generated deepfake impersonations, AMA officials said. The document, dated March 26, 2026, and copyrighted by the American Medical Association, outlines comprehensive measures to close legal gaps and safeguard patient safety, professional integrity, and public trust, according to the AMA’s press release from Chicago.

The framework prohibits AI-generated or altered content that impersonates physicians without clear, informed consent if such content falsely conveys endorsement, authorship, or medical judgment in a manner likely to mislead patients, AMA representatives said.

This prohibition targets deceptive uses of physician identity that could influence health-related decisions, with a focus on protecting reasonable patients from being misled, the policy states.

A key component requires separate, explicit opt-in consent for any use of a physician’s identity in AI-created or manipulated material. Consent must specify the intended use, audience, purpose, and duration, and be revocable if circumstances change, AMA officials explained. The consent process is designed to be standardized, reusable, and institutionally supported to minimize administrative burden on physicians, ensuring that consent is affirmative and specific rather than implied or bundled with other agreements.

The policy also mandates transparency through clear labeling of all AI-generated or altered depictions of physicians. According to the AMA, such content must include plain-language disclosures and digital watermarks, and patients must be proactively notified before interacting with synthetic professionals. These measures aim to prevent deception and support patient safety by ensuring full disclosure of synthetic content, the AMA said.

Responsibility for preventing impersonation is shared among platforms, hospitals, and AI vendors, the AMA framework states. These entities are required to implement safeguards such as rapid takedown mechanisms, prohibitions on the use of health professional titles in AI-generated content, preservation of audit logs, and cooperation with investigations. The policy centers on seven key areas to establish comprehensive protections for physician identity, including names, images, voices, and digital replicas, according to AMA documents.

Physicians are granted access to robust enforcement remedies, including procedures to document misuse, trigger content takedowns, and seek appropriate remedies. The framework empowers designated federal agencies to enforce relevant laws, require corrective disclosures, and issue annual public reports on incidents involving health-related impersonations, takedown performance, and enforcement actions. These agencies are authorized to investigate violations, preserve records, compel cooperation, seek injunctive relief, impose civil penalties, and mandate disclosures, the AMA said.

The AMA emphasized that the policy is designed to minimize administrative burdens on physicians by establishing identity protection as the default and supporting standardized, reusable consent processes. The framework aims to modernize protections without adding to physicians’ workloads while addressing gaps in existing privacy, employment, and intellectual property laws, according to AMA officials.

This policy announcement comes amid escalating risks posed by AI-manipulated images, videos, and audio impersonations of physicians, which have raised concerns about patient safety and professional integrity. The AMA’s framework represents a coordinated effort to address these challenges through legal, technological, and procedural safeguards, the association said.