You may have seen an ad recently about the “Pink Salt Diet.” This is the latest weight loss gimmick with sparkling support from celebrities and doctors.
Lisa Rozner of CBS News New York has discovered that some of these ads are AI-generated and known as deepfakes, and bad actors use the portraits of one Manhattan psychologist to assert false support.
“Pink Salt Trick” controversy
Dr. Rachel Goldman regularly educated the public on obesity and reached the pinnacle of her career when she appeared with Oprah Winfrey on “The Oprah Daily,” which began in 2023. She later appeared in several episodes.
But her global exposure has made her a target for deepfake videos that have been digitally manipulated and generated by AI. She said she was worried that some people might be receiving harmful medical advice, thinking it was coming from her.
Goldman, who is also an assistant professor at the NYU Grossman School of Medicine, said he is warning the public because it is difficult to delete videos.
In one episode in 2023, Goldman spoke to Winfrey about talking to people suffering from obesity.
“We don't want to define an individual by our medical condition,” she said.
CBS New York 
From that same episode, an AI-generated version was somehow created, and she says she saw it on Facebook and YouTube to her followers.
“It looks like me. It's kind of like talking a bit, but it's not me at the same time,” Goldman said. “It's so weird and scary.”
But even more scary is Goldman, that people are falling into it. She showed CBS New York the text of a relative who read, “Does it work? Send me the recipe.”
Goldman says at least two people sent an email saying they had purchased a product that includes an Australian woman.
One customer messaged her that she was being charged for triple and asked, “Do you have the product or did they have everything?”
“I get messages every day,” Goldman said. “I want them to know, don't buy this. It's not the truth that I stand for, it's fake.
“I'm really trying to go back to as many people as possible and try to spread more knowledge or try something that could potentially be dangerous before buying more people. “Supplements are not FDA regulations. I tend to say that anything that's a quick fix or hack is not usually true.”
Deepfake targets Oprah Winfrey
Goldman says the fake AI version actually shares false medical qualifications such as when she attended Stanford when she studied at Farry Dickinson University.
She said she was worried about the dangers of such false medical support.
“Some people even share their medical history with me via email and phone calls and ask if it's safe for them,” Goldman said.
CBS New York 
Even media moguls like Winfrey can't overcome the challenges of telling the truth and defeating deepfakes. The person in charge said she has a digital security team that works in false advertising and channel takedowns every day.
CBS News New York came across ads generated by AI on YouTube. For one, Winfrey's video was manipulated and says, “I saw the exact spot this morning to hold a pink salt recipe.”
In August, Winfrey told Oprah.com, “Let me say this clearly. If you see an ad with my face on the 'product', it's a fake. ”
Meta and YouTube told CBS News New York that they have removed accounts and ads related to Goldman and Winfrey for violating the policy.
“That was very unsettling for me,” Gale King says.
In 2023, CBS Morning co-host Gail King took her to social media to denounce a deepfake AI video that misprinted her support for weight loss products.
CBS New York 
CBS New York asked King, “What was that like for you?”
“Well, think about this, Lisa, I didn't even know about it. People started calling me and asked me, where can I get it?” the king said. “Then I looked at it and said, 'Well, that's the outfit I was wearing. I mean, it sounds like my voice. It looks like me.
“But if you look closely, if you really have to get close to make sure your lips are a little off, then the way I was talking is not the word I used, but that was very unsettling to me,” King added. “I don't think people would buy something because I think they supported it when I had nothing to do with it.
Laws designed to take on deep fakes
Rep. Yvette Clark, DN.Y., is a senior member of the Energy and Commerce Committee. Since 2019, she has introduced and reintroduced the “Deepfake Accountability Act.”
“We don't protect American people the way they should,” Clark said. “What my law simply does is that it requires a digital watermark of manufactured Deepfark so that people can quickly recognize that this was generated by artificial intelligence.”
CBS New York 
Next May, The bipartisan “Take it Down Act” is now enabledand will provide 48 hours to websites and online platforms to remove fake content.
Attorney Susan Jacobsen specializes in law firm Brennan Manna Diamond's honor and data privacy cases.
“The Act requires public websites, target platforms, and mobile applications such as Facebook, Instagram and Twitter to establish notification and takedown processes,” Jacobsen said.
Jacobsen advises victims to screenshot or screen record deepfake videos and record the harm they cause. She says that if the victim sues for damages and can identify the perpetrator, she can seek a takedown court order.
Tech experts show how easy it is to make deep furk
Rozner spoke with three different technical experts. She showed me how easy it is to create an AI Deepfake video.
Matt Swider, editor-in-chief of the Tech Review website TheshortCut.com, has shown Rozner to the platform that offers “instant voice clones” and the platform that synchronizes the lips.
He took her live report from Times Square and entered the new script using the version generated by her ready AI within an hour.
Rosner noticed her voice was deep in the manipulated video.
Swider replied, “If you upload more of your videos, you can get an exact match, but that was just a 20-second video.”
The Lionsgate Network supports victims of AI-duped financial fraud. Founder and CEO Bezalel Aitan Rabib says cases have skyrocketed over the past six months.
“So [bad actors] Raviv learns from all the interactions online. This is not a computer hack or code hack. This is to hack your consciousness. ”
“Daily” CEO Kwindla Hultman Kramer creates virtual avatars for education tutors and customer service programs.
He agrees that consumers are targeted in an environment where they least expect it.
“The very large social media companies that happen with these videos may not be looking for signs of deepfark in casual surveillance,” Kramer said.
Fighting against deepfakes “Not a priority”
The FBI says the Internet Crime Complaint Center has received more than twice the amount of complaints about Deepfark AI Video so far compared to the same period last year this year.
Agents also have statistics showing that their economic losses have almost tripled.
“This is an issue that can be addressed through the FCC and can be addressed through the Department of Homeland Security, but it's not a priority,” Rep. Clark said.
When it comes to deciphering medical advice online, Goldman said, “Do I always talk to people, pause and pause before responding? I like to pause, ask questions, ask people, ask doctors. Be honest, don't email me.
As for how to find deepfake AI, experts say you should look for troublesome facial movements, lip rotation issues, human hand movements, and unnatural audio patterns.
If the validity of medical advocacy is unknown, experts will look at the doctor's official website and social media to see if they have announced approval for the page.
