Serge Hercberg, MD, PhD, managing director at Université Sorbonne Paris Nord, Paris, France, has filed a criminal complaint after becoming the target of artificial intelligence (AI) generated deepfake (DF) videos that misused his name, image, and voice to spread false nutrition claims — highlighting a growing risk for health misinformation and the urgent need for stronger platform safeguards to protect scientific communication.

It took 5 days, multiple reports, and a formal complaint before YouTube shut down the channel responsible for the fabricated content. Hercberg chose to maintain his complaint to draw attention to the growing risks for health misinformation.

False Advice

Several days ago, Hercberg, known for co-developing the Nutri-Score nutritional labelling system, discovered multiple DF videos posted on a YouTube channel in which an AI-generated avatar resembling him used his name and voice. Seated at a desk in a white coat, the fabricated version of Hercberg delivered false advice, including claims about slowing dementia, avoiding certain vitamins to protect the liver and heart, and promoting the use of three plant-based proteins to repair worn-out knee cartilage.

Speaking with the Medscape French edition, he said,“I do not know the motivations of the authors of these videos, whether there is a commercial objective or an intention to cause harm.”

Created on December 31, 2025, Santé Sage published 32 videos containing fabricated health content using Hercberg’s identity without disclosing that the material was AI-generated.

Reporting Failures

Hercberg filed multiple reports with YouTube and with the French Pharos platform and lodged a formal complaint against Santé Sage.

After 5 days of unsuccessful efforts and several refusals by the platform, YouTube ultimately removed the channel on Tuesday, February 10. Hercberg intended to pursue legal action against the creators of the DFs, which damaged their reputation.

Speaking to Medscape French edition, Mathieu Molimard, MD, PhD, professor and head of the Department of Medical Pharmacology, Bordeaux University, Bordeaux, France, who recently reported on health misinformation to the French Minister of Health, criticized platform moderation failures and said, “It required escalation at the highest level for this channel to be closed.”

“This case shows that the reporting procedure does not work,” said Molimard. “Moderation is not the priority of platforms, which either do not do this work or do it poorly because it conflicts with their economic model, which relies on the number of video views. Today, people who spread disinformation or pose as experts often act without consequences and attack scientists without accountability. This has to change.”

In a LinkedIn post published immediately after the channel was removed, Molimard called it “urgent to implement more effective safeguards, relying on genuine trusted whistleblowers: public agencies, scientific societies, universities, hospitals.” Without decisive action and simplified appeal procedures, he argued that platforms should be considered responsible — or even complicit — for such misconduct.

Global Reach

The reason for YouTube’s delayed action despite repeated reports remains unclear. When contacted by Medscape French edition, Asmaa Zerkdi, strategic partner manager for health partnerships at YouTube, did not respond to questions. “I had contact with the platform the day the channel was shut down,” said Hercberg. “They indicated they would investigate why the reporting procedure had not worked.”

DF video manipulation, which uses AI to alter real footage or create realistic fake videos, has become increasingly common in recent years. Several physicians have been targeted, including television doctor Michel Cymes. After earlier deepfake clips promoted weight-loss treatment, a fabricated version of Cymes appeared in autumn 2024, advertising a so-called miracle syrup for diabetes.

In July, Marine Lorphelin, a general practitioner and former Miss France, spoke out after AI-generated videos were used to market dietary supplements. In a post on Instagram, she stated that she would never promote miracle foods or products whose effects have not been scientifically proven.

These concerns extend beyond the borders of France. Australia’s favorite doctor-scientist-engineer and science communicator Karl Kruszelnicki saw his image used in a DF video to sell “miracle” pills in Australia. In England, Michael Mosley, a television journalist and physician who died on June 5, 2024, on the Greek island of Symi, was similarly featured in AI-generated videos promoting products that claimed to treat hypertension and diabetes.

Legal Consequences

The use of AI have become a growing concern for France’s National Council of the Order of Physicians (CNOM).

“We are addressing the issue within a dedicated working group because we are seeing rapid developments in artificial intelligence that could create real challenges in the near term,” said Jean Canarelli, MD, general delegate for Health Data, Digital Technology, and Innovation at CNOM.

He urged physicians to report incidents to their local medical council. “Whatever form this identity theft takes, we strongly condemn it because it undermines physicians and the trust patients place in them. A patient who does not trust their doctor may receive inadequate care,” he said, calling for regulatory changes to allow stronger action against those who spread disinformation.

French law sanctions the unauthorized use of a person’s image or words. Article 226 8 of the penal code provides for up to 1 year in prison and a €15,000 fine for publishing a montage using someone’s image or statements without consent and without clearly stating that it is a montage, explained Maître Renan Budet, an attorney and head of the health commission, Paris Bar Association, Paris, France.

Arnaud Dimeglio, lawyer specializing in digital, communications, and intellectual property law, noted in an article published on the legal blog Village de la Justice that editing content using the image or words of a consenting person, or clearly disclosing the use of AI, is not in itself unlawful. “The offense lies in creating content without consent and in concealing that manipulation.”

Dimeglio added that France’s SREN Law (Sécurité et Régulation de l’Espace Numérique — Security and Regulation of the Digital Space), aimed at securing and regulating digital space, increased penalties to 2 years of imprisonment and a €45,000 fine when the offence was committed via an online public communication service.

The removal of videos does not end the legal action. Victims may still seek compensation because damage has already been inflicted. “The removal of the videos does not end legal proceedings insofar as they have caused harm to the individual concerned,” Dimeglio said. Platforms that fail to act promptly may also face liabilities. A key challenge remains in identifying the responsible parties.

Dimeglio said that he regularly files requests with the presiding judge to identify the identities of individuals hiding behind anonymous accounts. “It is very important to file a complaint promptly in cases of defamation because the statute of limitations is 3 months,” he said.

This story was translated from Medscape’s French edition.