Nearly six months after launching its “Vibes” feature that allows users to create and explore AI-generated video feeds, the Meta AI app is starting to become a cesspool of AI-generated sexual video clips of children and celebrity deepfakes.
Videos circulating on the platform include AI-generated videos of children morphing the bodies of adult women, real babies singing inappropriate movie songs and reciting lines from adults, and videos of women resembling celebrities being undressed.
Some users shared prompts they may have entered to generate the images in their video captions. They revealed that many users were attempting to create deepfake content of sexually explicit celebrities. Creation and distribution of such content featuring real people is a violation of India’s IT rules.
Some video captions included male and female names of Bollywood or pan-Indian actors, while others featured sexualized AI versions that bore some resemblance to real-life women. The video showed an AI version of a female actor being kissed, groped and undressed by men.
One of the AI-generated videos showed a person resembling Mahatma Gandhi groping a woman’s breasts. Another photo with a caption with a casteist slur showed the couple touching each other inappropriately.
Read | How Grok brought deepfake ‘nudification’ to the mainstream
To generate videos of school girls wearing school uniforms in explicit scenarios while circumventing child protection guardrails, some users’ AI prompts/captions described the girls as being over 18 years old, but were described as wearing school uniforms.
Other prompts showed users attempting to generate partially or fully nude images of women by inserting phrases such as “transparent short kurti,” “transparent red-and-blue miniskirt,” “tight wet saree,” “navel-baring blouse,” and “camera focus” on women’s breasts. These captions were often reused by others.
Read |IT ministry mandates labeling of AI-generated content, reducing removal timeline to 2-3 hours
One such video depicted a “romantic atmosphere” with a “girl” in a “school dress” and showed her removing her shirt and revealing her underwear. In some AI videos, the faces of young children and teenagers were morphed into the bodies of adult women wearing dresses with low necklines.
Also, many of the videos were clearly distorted or deviated from the prompt, making it easy to tell they were fake. Extra limbs, fingers, and heads were also a common feature in many videos, as prompts were reused and remixed via the app.
Meta AI vibes are similar to Instagram Reels or TikTok videos, except they are made entirely or partially by AI. Users can also add background sounds and text captions, such as trending music or lip-synced dialogue.
Meta introduced Vibes in September 2025, describing it as “designed to make it easier to find creative inspiration and experiment with Meta AI’s media tools.”
Apart from creating AI videos, the Meta AI app allows users to manage their AI glasses, and the Meta AI assistant answers questions and helps generate multimedia content.
A recent investigation by Swedish media outlets Svenska Dagbladet (SvD) and Goteborgs-Posten (GP) alleges that highly sensitive personal footage recorded via customers’ Meta AI glasses was made available to Meta’s subcontracted workers in Kenya. The investigation said this included footage of sexual encounters, pornography, banking information, bathroom visits and more.
British regulators have contacted the social media giant for further information.
issued – March 12, 2026 12:49 PM IST
