Bangladeshi Activist Targeted with Deepfake Videos, Violating Meta’s Policies
Two videos (1, 2) featuring Bangladeshi activist and model Farzana Sithi, known for her involvement in the anti-discrimination student movement in 2024, have been circulating on Facebook in recent months. These videos were shared from a profile operating under her name and identity, but Dismislab’s verification has found that the videos are deepfake content, not real. An Analysis of the profile also shows that the account itself is fake, created using Sithi’s name and has been active since December 13, 2024. This profile, categorized as “Professional,” has been regularly posting such videos, where deepfake technology has been used to alter the individual’s face. These edited videos are being widely shared and are receiving millions of views, resulting in cyber harassment of a female activist. These videos and the fake profile also violate multiple Facebook policies.
On March 4 and March 6, 2025, two reels were shared on Facebook from the profile named “Farzana Sithi”. In the first reel, Sithi is seen dancing in a black outfit. It has received over 100,000 views and more than 1,500 reactions. In the second reel, which is the most-viewed video on the profile so far, she appears in an orange sari. By the time of writing this article, the video had received over 2 million views, been shared more than 100 times, and garnered reactions from over 14,000 users.
A reverse image search using key frames from the two reels shows that the original sources of the video clips are two Indian content creators, not Farzana Sithi. The first reel was originally posted on February 10, 2025, by Indian actress and content creator Tanya Desai on her Instagram account. The second reel was found on the Instagram account of another Indian content creator, Pooja Mallick. This makes it evident that the woman seen in the reels is not Sithi, but Tanya and Pooja, whose faces were altered using AI technology and replaced with Sithi’s face before posting them on the fake Facebook profile.
Analysis shows that this profile has been regularly posting such reels (1, 2, 3, 4, 5) and videos (1, 2, 3) of Farzana Sithi for more than four months under her name and identity. She told Dismislab that the current profile picture isn’t hers either. The profile features an image of Sithi as the profile photo, claiming that she is a professional model and that the profile belongs to her. However, this claim is false. When Dismislab reached out, Farzana Sithi shared her verified profile, which she said is her only profile.
Between December 13, 2024, and March 17, 2025, a total of 166 reels and videos were shared from the fake profile in question. Among them, 147 were reels and videos, with 144 on Farzana Sithi. Analysis shows that 133 of these are AI-generated deepfake content. The remaining 11 include interviews Sithi gave to journalists at different times and other general videos. Since December 13 last year, the profile has been sharing deepfake content featuring Sithi’s face almost daily. An analysis of the five most-viewed videos (1, 2, 3, 4, 5) reveals that the profile has garnered over 4.1 million views so far. Facebook’s transparency section indicates that the account is being operated from Bangladesh.
Each deepfake video featuring Sithi’s face has been monetized. Some video clips have the creator’s own caption, “Give me star” or “Star give me.” Some reels and videos have been posted directly asking for stars, further clarifying the motive for monetization.
According to Facebook, the Stars feature on Reels allows content creators to monetize their video and audio content. Viewers can purchase Stars and send them to creators through reels as a form of support. Facebook pays creators $0.01 per Star. This means, if a content receives 10,000 Stars, the creator will earn $100 from Facebook.
Deepfake technology uses a form of artificial intelligence known as deep learning to create images of fabricated events or people. The technology first appeared online in 2017 when a Reddit user posted doctored pornographic videos featuring Hollywood actresses. While it has been misused for pornography, its applications also extend to spoof, satire, and other forms of mischief. In a report by The Guardian, Danielle Citron, a professor of law at Boston University, quotes: “Deepfake technology is being weaponised against women.”
Farzana Sithi is a student at a government college in Dhaka and was active in the 2024 student-led quota reform movement. She has frequently been in the limelight on social media. In addition to her studies, she works as a model. Sithi told Dismislab that the deepfake videos and fake profiles have caused significant problems in her real life. She said that these videos are objectifying her. “These videos are an attempt to tarnish my character. I have been subjected to various forms of bullying and shaming. It has affected not only my social media presence but also my personal life. Whenever any of my family members or I go outside, people start asking weird questions related to these videos. Even my character is always being questioned. I am frustrated.”
Sithi said she reported to Facebook to remove the deepfake videos and the fake profiles on multiple occasions, and is awaiting its action. “I also tried seeking help from third-party IT experts, but they demanded $100 to $150 to remove all the fake accounts and videos, which I cannot afford,” Sithi explained.
The fake account using Farzana Sithi’s name and identity violates Facebook’s policies. Sithi told Dismislab that she has only one personal profile and that none of the others belong to her. According to Facebook’s impersonation policy, “You may not use someone else’s name or image in a way that misleads others into believing that the account represents a particular person or organization.”
However, an analysis of the comments on the reels posted from the fake account shows that many Facebook users were misled into believing that the videos were Sithi’s own dance performances. Some users were seen making offensive and provocative remarks. One user commented under a reel, “Rapists, don’t you see these? You only target 6-year-olds and 8-year-olds.” Another user wrote – “…if this is your condition, you should be raped by 50 men every day.”
According to Meta’s Manipulated Media Community Standards, users can not post “Videos that have been edited or synthesized, beyond adjustments for clarity or quality, in ways that are not apparent to an average person, and would likely mislead an average person to believe: The video is the product of artificial intelligence or machine learning, including deep learning techniques (e.g., a technical deepfake), that merges, combines, replaces, and/or superimposes content onto a video, creating a video that appears authentic.”
The fake profile operating under Farzana Sithi’s name remains active on Facebook and is being operated as a “Professional” profile. Since the profile is monetized, a “Give Star” option appears at the beginning of each content indicating that they are generating revenue through Facebook’s Stars feature.
On March 11, Farzana Sithi filed a case with Sher-e-Bangla Nagar Police Station in Dhaka against Bangladeshi content creator Khalid Mahmud Hridoy Khan over alleged threats of sexual violence in a Facebook live session. Later Khalid Mahmud Hridoy was arrested by Savar police for harassing girls while pretending to be intellectually disabled.
Sithi had also reached out to the media in November amid severe cyber harassment. At the time, she told The Daily Star that misleading content was being deliberately spread against her, with edited and distorted versions of her video clips. She said she was facing constant cyberbullying, with most of the attacks targeting her character, clothing, and activism.
