Core signal

Clinician deepfakes are becoming a healthcare identity problem.

Documented campaigns have used the faces and voices of real physicians, medical researchers, and public health figures to promote supplements, questionable treatments, and unsupported health claims. The immediate harm is deceptive advertising. The larger issue is that synthetic media can appropriate the identity signals patients use to decide whether health information is credible.

This moves the problem beyond ordinary misinformation. The false claim is no longer presented only as text from an unknown account. It can be delivered through a familiar face, a recognizable voice, professional credentials, a clinical setting, and the apparent confidence of a real medical expert.

The strongest available evidence concerns social-media advertising and product promotion. It does not yet show that deepfakes are broadly compromising telehealth systems, patient portals, or routine clinical workflows.

What has been documented

In 2024, The BMJ reported that synthetic videos were using the likenesses of recognizable United Kingdom physicians and medical presenters, including Hilary Jones, Michael Mosley, and Rangan Chatterjee. The videos promoted products claiming to address high blood pressure and diabetes and advertised products such as hemp gummies.

Later that year, ABC News documented a Facebook advertisement that appeared to show Australian diabetes specialist Jonathan Shaw criticizing metformin and promoting a supplement. Patients who saw the video contacted his clinic. Meta removed the advertisement nine days after it was reported, according to the Baker Heart and Diabetes Institute, and after the institute submitted a separate intellectual-property complaint.

In August 2025, a CBS News investigation found dozens of accounts and more than 100 social-media videos involving fictitious doctors or the identities of real physicians. Many promoted beauty, wellness, and weight-loss products. Physician Joel Bervell reported that his likeness appeared in synthetic content across TikTok, Instagram, Facebook, and YouTube.

In December 2025, Full Fact identified another group of TikTok and Instagram accounts using synthetic depictions of doctors and academics to promote health products. In one case, TikTok initially reported that no violation had been found. The platform later restricted the content and took further action after additional reporting and external scrutiny.

These investigations do not provide a complete prevalence estimate. They do establish a repeated pattern across multiple countries, clinicians, products, accounts, and platforms.

Why this is an identity problem

Traditional medical misinformation can be challenged by examining the claim and comparing it with evidence. Clinician deepfakes add a second verification problem: whether the apparent speaker is real.

The person watching the video must now evaluate at least two separate questions:

  • Is the medical claim accurate?
  • Did the clinician actually make the claim?

The second question matters because medical authority is part of the advertisement. A recognizable physician may bring an existing relationship with patients, a public reputation, institutional affiliations, professional credentials, and years of recorded media that can be repurposed into synthetic content.

That trust is being converted into an advertising asset without the clinician's participation or consent.

The Australian case shows how this can cross into an existing clinical relationship. Patients did not only encounter a false claim from an anonymous source. They believed their specialist might be recommending a new treatment and contacted his clinic for clarification.

The platform-enforcement gap

The investigations also show that published platform policies and actual enforcement do not always align cleanly.

Platforms removed or restricted some material after journalists, clinicians, institutions, or users reported it. Other examples remained available, were initially found not to violate policy, or required repeated escalation. New accounts and copies can also appear after earlier versions are removed.

This produces a structural imbalance. Synthetic advertisements can be generated and distributed quickly. The affected clinician or institution may need to discover the content, document it, identify the relevant reporting process, submit a complaint, wait for review, and repeat the process across several platforms.

The burden therefore falls heavily on the person being impersonated, even though that person did not create the content, authorize the advertisement, or benefit from the product being sold.

A healthcare identity response will need more than general content moderation. It will require reliable impersonation reporting, preservation of evidence, rapid escalation, repeat-offender detection, and continuity across accounts, products, advertising networks, and platforms.

Health advertising rules already matter

Clinician deepfakes exist inside an established health-advertising environment.

The United States Federal Trade Commission requires advertising claims for health-related products to be truthful, not misleading, and supported by appropriate evidence. The requirement applies to express and implied claims and is especially important when consumers cannot independently evaluate a medical or scientific assertion.

A fabricated physician endorsement creates several possible layers of deception. The identity may be false. The endorsement may be unauthorized. The health claim may be unsupported. The product seller, advertiser, affiliate, account operator, and platform may each occupy different parts of the distribution chain.

The FTC's current impersonation rule specifically covers government and business impersonation. The agency has separately proposed extending that rule to impersonation of individuals, including AI-enabled impersonation.

That proposal is not the same as a finalized physician-specific federal rule.

Policy is beginning to respond

In April 2026, the American Medical Association published a physician-deepfake policy framework. It calls for affirmative consent before a physician's name, image, likeness, voice, or digital replica is used; clear labeling of synthetic depictions; rapid takedown procedures; audit records; practical remedies; and shared responsibility across platforms, hospitals, and AI vendors.

The framework is a professional policy position rather than legislation, but it identifies the operational controls that medical organizations increasingly view as necessary.

California Senate Bill 1146 represents a more targeted legislative approach. The bill would require disclosures when advertisements for health-related products or services use digital replicas or synthetic performers depicted as healthcare providers. At publication, the bill remained pending in the California Assembly Appropriations Committee.

The regulatory direction is notable because it focuses on the combination of synthetic identity, medical authority, and commercial health claims rather than attempting to regulate every use of generated media through one broad rule.

What is not established

The evidence does not currently support a claim that clinician deepfakes are widespread across telehealth encounters, patient portals, electronic health records, or internal clinical communications.

Those are plausible threat surfaces, particularly as voice cloning and realistic video generation improve. They should be treated as prospective security and governance questions, not documented prevalence claims.

The documented problem is already significant without extending it beyond the evidence:

  • real clinicians have been impersonated
  • synthetic endorsements have promoted health products
  • patients and followers have encountered the content
  • platform enforcement has sometimes required external escalation
  • the same pattern has appeared across countries and platforms

Keeping that boundary clear matters. A narrow, verified problem is more useful than a larger claim that cannot yet be supported.

What healthcare organizations may need

The first response is unlikely to be a single deepfake detector.

Detection tools can contribute, but synthetic media changes quickly and individual users cannot be expected to perform forensic analysis before deciding whether a familiar clinician is genuine.

A more durable response may combine:

  • verified clinician and institutional communication channels
  • clear policies governing authorized synthetic representations
  • revocable consent records for the use of clinician identity
  • rapid impersonation-reporting and escalation procedures
  • preservation of original media, account, product, and advertising evidence
  • tracking of repeat products, affiliates, accounts, and campaigns
  • platform-level provenance and synthetic-content disclosures
  • patient guidance for independently verifying unexpected medical endorsements

These controls do not prove that every piece of media is true. They help establish whether a communication came through an authorized channel and whether the apparent speaker consented to the representation.

What people can do if they are unsure

A convincing video is not proof that a clinician recorded it, approved it, or recommends the product being promoted.

If a video, voice message, advertisement, or social-media post appears to show a healthcare professional making an unexpected recommendation, pause before acting on it.

  • Do not rely on the link, phone number, account, or contact information contained in the questionable post.
  • Find the clinician or healthcare organization's official website independently.
  • Contact the clinic through a phone number, patient portal, or email address you already know or can verify.
  • Check whether the same recommendation appears on the clinician's verified professional or institutional channels.
  • Be cautious when a supposed clinician endorsement includes a discount code, limited-time offer, affiliate link, or request for payment.
  • Do not stop medication, change treatment, or begin a new health product based only on a social-media video or advertisement.
  • Ask a licensed healthcare professional or pharmacist to review any medical claim that could affect your care.
  • Save the page URL, account name, screenshots, date, product name, and advertisement details before reporting the content.
  • Report the content using the platform's impersonation, fraud, misleading health information, or unauthorized-likeness process.
  • If money or financial information was provided, contact the payment provider or financial institution promptly.

A useful verification question is simple:

Can this recommendation be confirmed through a communication channel that the clinician or healthcare organization controls?

If the answer is unclear, treat the content as unverified.

Visual warning signs can help, but they are not reliable enough on their own. Unnatural blinking, unusual lip movement, mismatched audio, abrupt edits, or distorted facial details may indicate synthetic media, but increasingly realistic content may show none of these signs.

The safer approach is not to become an expert deepfake detector. It is to verify the identity, recommendation, and communication channel independently.

Where people can seek help

There is not currently one service that can authenticate every suspected clinician deepfake. Consumer-protection and health-product authorities can still help people assess possible deception, preserve evidence, and identify the appropriate reporting route.

Use the organization's official website to confirm its current contact details. Do not call a telephone number, follow a link, or use contact information supplied inside the suspicious advertisement, video, message, or account.

North America

  • United States: The AARP Fraud Watch Network provides guidance when someone is unsure whether a health-related endorsement or advertisement may be part of a scam.
  • Canada: Health Canada accepts complaints involving suspected illegal or misleading advertising of health products.

Europe

These organizations generally do not perform forensic video authentication. The clinician or healthcare organization should still be contacted independently through an official website, clinic number, or patient portal.

What to watch

The next signal is not simply whether more deepfake videos appear. It is whether healthcare identity protection becomes operational.

Relevant developments include:

  • standardized reporting channels for impersonated clinicians
  • faster platform response times
  • public incident reporting by healthcare institutions
  • repeat-campaign detection across products and platforms
  • legislation addressing synthetic health endorsements
  • identity and provenance controls for official patient communications
  • clear distinctions between authorized synthetic clinicians and unauthorized replicas

The recurring entities may be as important as the individual videos. The same product, affiliate network, account operator, or advertising infrastructure may use several clinician identities and reappear after takedowns.

That continuity is difficult to see when each post is treated as an isolated moderation event.

Working thesis

Clinician deepfakes are already a documented health-advertising and identity-enforcement problem.

The immediate challenge is not that patients can no longer trust any digital healthcare interaction. It is that existing systems place too much responsibility on patients and clinicians to detect, report, and disprove synthetic endorsements after they have already circulated.

The healthcare response will need to protect identity as infrastructure: consent, authorized channels, evidence preservation, rapid takedown, provenance, and continuity across incidents.

The wider risk to telehealth and clinical workflow remains important to examine. It should not be presented as established before the evidence exists.

Editorial boundary

This article reviews documented clinician-impersonation incidents, platform responses, advertising requirements, and emerging policy measures.

It does not conclude that deepfakes are currently compromising telehealth systems or patient portals at scale. It does not evaluate the safety or effectiveness of any product shown in the referenced synthetic advertisements.