Celebrity deepfakes used in scam surge affecting millions
Scammers are exploiting generative AI tools to create deepfakes of celebrities and use them in fraud schemes involving fake endorsements and products. Reports from McAfee show millions of people globally have been exposed to these scams, with measurable financial losses and increasing success rates among younger demographics.
What happened
Scammers are using generative AI tools to create deepfake videos, images and audio clones of celebrities to promote fake endorsements, giveaways, and fraudulent products. According to McAfee, Taylor Swift's likeness is exploited most frequently in US scams, followed by Scarlett Johansson, Jenna Ortega and Sydney Sweeney; in India, Shah Rukh Khan, Alia Bhatt and Elon Musk are most abused.
The scammers create social media posts with fake celebrity endorsements convincing enough to deceive consumers into clicking bogus links, applying for fraudulent giveaways or purchasing fake products. McAfee reported that 72% of Americans have seen a fake celebrity or influencer endorsement, with 39% clicking on false endorsements and 10% putting personal information at risk.
Reported impact
In the United States, 10% of people who encountered these scams lost personal information and an average of $525 per person. In India, approximately 90% of the population is exposed to AI-generated celebrity endorsement scams, with victims losing an average of Rs 34,500.
Younger age groups are most at risk, with 62% of those aged 35 to 44 and 60% of 25 to 34-year-olds admitting to clicking on fake celebrity ads, compared to only 17% of those over 65.
Index analysis
Generative AI tools like image, video and audio generators have lowered the barrier for scammers to create convincing fake celebrity content without consent. AI companies have implemented systems to prevent misuse, but these safeguards have been shown to be imperfect and can be worked around.
Classification
Sources and evidence
This record was researched and written by the Index, based on the following original reporting.