Maurice Sholas, MD, PhD, a pediatric rehabilitation specialist, may have a strong online presence, but he still considers himself “the old guy in the club” where social media is concerned.
Last year, as he (@thedocmosho) was “stumbling my way through TikTok” a friend sent him a video of a man that looked exactly like him endorsing a vitamin product. His friend asked: “Is this you?”
“I was baffled,” Sholas told Medscape Medical News. “…It looks like me, but the voice wasn’t me. And it wasn’t something I had ever talked about online.”
Sholas, a pediatric rehabilitation specialist at Ochsner Health in New Orleans, reached out to TikTok to file a complaint. TikTok’s basic response was: “I don’t have standing because they’re like, ‘well, that’s not you in the video and it’s not your video, so basically, it’s not your business,’” Sholas said.
The fake video of Sholas’ likeness, which was promoting supplements specifically to Black people, started appearing on X and Instagram as well. On X, it’s been viewed more than 19,000 times.
Sholas was stunned. “I just started to see me selling vitamins all over the place,” he said.
Reporting the fraud to X and Instagram didn’t go much better than with TikTok. Essentially they either ignored him or said it was free speech, Sholas said.
They’re using a fake image of me “to sell a product that I don’t know about and didn’t authorize, and they didn’t pay me for,” Sholas said of the bad actors circulating the video.
AI deepfakes are a growing problem in medicine, said John Whyte, MD, MPH, CEO of the American Medical Association (AMA).
People in the community would believe their physician is telling them to take this supplement or take this product, Whyte said. “That’s what’s so concerning.”
Whyte first started hearing from doctors who had been deepfaked using AI over the past year. At first, Whyte figured that meant their identify was used to open a fraudulent account. But “what’s happened with AI deepfakes, it’s actually stealing your likeness, your voice, and it’s actually looking like you online or on social and sounds like you,” Whyte said.
Deepfakes are widespread across scams, fraud, political misinformation, impersonation, and other forms of online deception, Siwei Lyu, PhD, director of the Institute for Artificial Intelligence and Data Science at the State University of New York, University at Buffalo, Buffalo, New York, said in an email.
“In medicine, the problem is newer but clearly growing,” Lyu said. “We are already seeing real physicians’ faces and voices used to promote supplements and questionable treatments, and other products without their knowledge. Some of these videos have received millions of views.”
“When used maliciously or without consent, these tools can spread medical misinformation, damage physicians’ reputations and careers, and undermine patient trust in evidence-based care,” according to AMA written materials.
The AMA learned that there is no penalty if social media platforms don’t take the AI deepfake down after they have been notified about it, Whyte said.
After Sholas complained to the social media companies and the bad actors posting the fake video, his online face was aged by adding more wrinkles or his teeth would be “messed up,” he said. This way it could be argued that it wasn’t exactly him, Sholas guessed.
“What’s important here is that we be proactive, that we get ahead of this and address it,” Whyte said.
The AMA launched a policy framework in April to counter AI deepfakes, and separately, the organization joined other groups like SAG-AFTRA and YouTube in supporting the NO FAKES Act of 2026.
“The bill creates a federal intellectual property right to protect individuals from unauthorized digital replicas (eg, digital content made using generative artificial intelligence of their voice and visual likeness),” according to a summary of the bill on Congress.gov.
It also creates a notice-and-takedown process for unauthorized digital replicas.
“If we don’t criminalize the individuals with real penalties that are doing this, they won’t stop,” Whyte said. “And if we don’t require the platforms to address it in a timely manner, they’re not doing it.”
Creating a federal baseline that gives people meaningful control over highly realistic digital replicas of their own face and voice “addresses a genuine gap in current law,” Lyu said.
In mid-June, the Senate Judiciary Committee advanced an updated version of the legislation (S. 4591), but Whyte would like to see more movement on the bill.
Not everyone is on board. “…the NO FAKES Act goes much further than targeting deceptive impersonation,” Becca Branum, deputy director of Center for Democracy & Technology’s Free Expression Project, said in an email. “It would create a broad new federal right over highly realistic digital replicas of people, backed by a takedown regime that could be abused to censor lawful satire, commentary, news, and other protected speech, including information about health and medicine.”
The bill exempts certain uses like news reporting and satire as they are categories likely to involve protected speech under the First Amendment, however that doesn’t go far enough, Branum said.
“If a platform leaves the content up because it believes the content is protected satire, news, or commentary and it turns out to be wrong, the platform can face significant liability,” she said.
No Way to Counter
As soon as he became aware of the AI deepfake, Sholas tried to use the tools available to him as a regular social media user to get it taken down.
However, “the automated system does not allow you to specify your issue,” Sholas said.
“It was a series of sort of automatic responses that came back and said, ‘we’re not taking it down.’” He couldn’t reach anyone on the phone either.
Eventually Sholas reached out to the local media in New Orleans — WWL-TV — which ran a piece last year. WWL-TV reached out to TikTok, and only then did it ban the account behind the deepfake video. However, deepfakes of Sholas remained up on Instagram at the time, WWL-TV reported.
Sholas also reached out to Meta, the parent company of Instagram.
“I saw the postings and got myself ‘verified’ to try and combat the misrepresentation, and used the automated reporting system,” Sholas said in an email. “Meta/IG was not responsive or helpful. I am not sure that my complaints made it past automated responses. No one ever engaged with me personally or responded to my concerns.”
WWL-TV also reached out to Meta and the user who posted the deepfake video last year and neither responded.
Medscape Medical News contacted TikTok and Meta.
TikTok is currently testing a new opt-in tool called “Likeness Detection” with a select group of US creators to help them detect AI-generated content that may be using their likeness without consent, spokesperson Patrick O’Neil said in an email. Once enrolled, creators can check flagged content and provide reports “for our trust and safety team to action,” O’Neil said.
A spokesperson from Meta directed Medscape Medical News here for more information about reporting a problem. “Identifying AI-generated content can be challenging as the technology evolves,” she said in an email. “We’re continuously working to improve our systems and ability to label this content.”
“That’s why it’s important to consider several factors when determining if content has been created by AI. This includes checking the trustworthiness of the account sharing the content and/or looking for details that might look or sound unnatural,” the spokesperson said.
She confirmed that “users can report accounts they suspect of impersonation.”
Medscape Medical News was unable to reach officials at X.
The Impact
Unable to fully stop the deepfake videos of his likeness from circulating, Sholas said he is concerned about his professional reputation, something he took decades to build.
“Now I have to defend sort of from a medical liability standpoint, something I never said,” Sholas said.
AI deepfakes also affect patient safety, AMA’s Whyte said.
Sholas said he’s concerned about a person doing something based on the advice of his fraudulent avatar — such as buying the supplements it’s peddling — and then having a bad medical outcome.
“Someone may buy a supplement, change treatment, or disregard legitimate medical advice because they believe a trusted doctor endorsed something,” Lyu said.
Not able to negotiate effectively with the social media companies on his own, Sholas spoke with an attorney and some public relations experts.
A friend’s husband bought the supplements because he thought what he saw on social media was the real Sholas endorsing them. And because of that, Sholas was able to find out that the product shipped from a distribution center in Houston. However, he was unable to pin down the name of the parent company or even determine whether it is a US-based company or one based overseas.
“It was very difficult to figure out who was even responsible,” Sholas said.
Ultimately, Sholas got a recommendation to hire a PR firm “to basically go online and sanitize what the public sees” essentially through search engine optimization. But at a cost of $20,000, it was too expensive to proceed.
Don’t Fight It Alone
If you find yourself in a situation where you think you’ve been deepfaked, stay aware, Sholas said. Work with your local medical society. “This is something that’s very hard to do by yourself.”
Today Sholas monitors his social media presence and if he sees something else come up, he will report it through whatever channels are available to him.
“That’s all I got,” Sholas said. “There’s no sort of systemic fix, and I don’t know how to protect myself from someone else doing the same thing later….”
Lyu reported his research being sponsored by the National Science Foundation.
Sholas reported being on the board of the National Medical Association and has a Robert Wood Johnson Foundation fellowship that he’s taken within the last 5years. No other disclosures.
No comments:
Post a Comment
Note: Only a member of this blog may post a comment.