BBC News:
One Bollywood star is making obscene gestures to the camera, another is posing while scantily clad.
Except neither of those things actually happened.
They are the latest in a string of deepfake videos which have gone viral in recent weeks.
Rashmika Mandanna, Priyanka Chopra Jonas and Alia Bhatt are among the stars who have been targeted by such videos, in which their faces or voices were replaced with someone else’s.
Pictures are often taken from social media profiles and used without consent.
So what is behind the rise in Bollywood deepfakes?
- How AI may affect India’s vast film industry
- India actress calls out ‘scary’ deepfake video
- Fears UK not ready for deepfake general election
Deepfakes have been around, and have targeted celebrities, for a long time.
“Hollywood has borne the brunt of it so far,” AI expert Aarti Samani told the BBC, with actresses such as Natalie Portman and Emma Watson among the high-profile victims.
But she said recent developments in artificial intelligence (AI) have made it even easier to create fake audio and video of people.
“The tools have become so much more sophisticated over the past six months to a year, which explains why we are seeing more of this content in other countries,” Ms Samani said.
“Many tools are available now, which allow you to create realistic synthetic images at little or no cost, making it very accessible.”
Ms Samani said India also has some unique factors, including a large young population, heavy use of social media, and “fascination with Bollywood and obsession with celebrity culture”.
“This results in videos spreading quickly, magnifying the problem,” she added, saying that the motivating factor for creating such videos was twofold.
“Bollywood celebrity content makes an attractive clickbait, generating large ad revenue. There is also the possibility of selling data of people who engage with the content, unknown to them.”
