- Evidence of regular circulation of edited photos and videos of female celebrities on a Facebook page named ‘Cholo Bodle Jai’.
- At least 116 pieces of disinformation identified in the last seven months.
- 79 pieces of disinformation identified in August alone.
- The page belongs to a person from Gazipur named Imran Miah, who is disabled.
- Female celebrities are being defamed/suffer loss of honor due to the controversial posts.
The daily activities of the Facebook page start by greeting everyone with ‘Good morning’. Following this, religious greetings, or ‘Good afternoon,’ ‘Good evening,’ or ‘Good night’ posts are regularly seen. However, interspersed with these ordinary posts, numerous fake contents claiming to be photos and videos of celebrity women are being constantly circulated. Rumor Scanner has found one such Facebook page named ‘Cholo Bodle Jai’ (Let’s Change) where, consistent with the page’s name, there is an ongoing mission to regularly alter women’s faces. Due to these objectionable contents, the women involved are frequently becoming victims of adverse comments. The person behind the page is a disabled individual from Gazipur.
Rumor Scanner found the page named ‘Cholo Bodle Jai‘ on April 16 of the current year. On that day, in a post found through regular monitoring, it was claimed that some pictures were of Dhallywood actress Pori Moni. The post has received over five thousand reactions and nearly a thousand comments to date. Netizens have made both positive and negative comments. However, a review shows that the volume of negative comments is higher, with the majority being body-shaming in nature.

Rumor Scanner verified and found that these were not Pori Moni’s photos at all. The claim was being circulated by using face-swap technology to place Pori Moni’s face onto the photos of a different woman from Shimla, Himachal Pradesh, India. Face swap is a technology where one person’s face is digitally exchanged with another’s. This makes it look as if the first person is actually looking or acting like the second person. This is typically done using Artificial Intelligence (AI) and deepfake technology. In recent years, the misuse of these AI tools has surged due to their easy availability. Pages like ‘Cholo Bodle Jai’ have emerged as an example of this. Rumor Scanner kept this page under regular surveillance since April. Fact-checks were also performed on various posts of the page in the subsequent months. As of September 23, Rumor Scanner has published fact-checks identifying 37 posts of the page as disinformation.

This discussed page was launched on September 22, 2022. The page has been run under the same name, ‘Cholo Bodle Jai’ (Let’s Change), since its inception. The page is being managed by a single administrator from Bangladesh. Gazipur is mentioned as the location of the page. Since its launch, the page, which has 28,000 followers, has been regularly posting pictures of various women. The circulation of AI-based fake and edited content on the page was noticed at least from last year. Following this trend, its volume began to increase this year.
In the investigation regarding this page, it was learned that a person named Imran Miah is the entrepreneur behind the page. His home is in Senpara village of Nagari Union in Kaliakair Upazila of Gazipur District. He himself revealed this information in a video on September 13. See the video here. In the video, he stated that 17 years ago, he fell from a tree, which caused his body, from the neck down, to become paralyzed. Since then, he has been unable to move in any way. He cannot even eat with his own hands. In the video, he claimed this page to be his own.
Another video of Imran was circulated on the page on September 12. See the video here. A different person speaks in the video, claiming to be Imran’s cousin, with Imran present. However, the said person was not seen in front of the camera. He claimed in the video that Imran has to use his mobile phone with the help of others.
From a post on Imran’s Facebook account, it is known that he was injured after falling from a tree on December 22, 2008. He had to accept lifelong disability (paralysis).
According to Imran and his cousin, Imran cannot do any work with his hands. In that case, there are questions about who makes the posts on the page ‘Cholo Bodle Jai’. Rumor Scanner spoke with Imran to find out more about this. When initially asked if he makes the videos for the page himself, he was not keen to talk, but later he stated, “I have a little bit of strength in my hand, with that, I can somehow slowly operate the mobile.”
Imran told Rumor Scanner that he used to edit pictures and videos after watching YouTube and Facebook tutorials and upload them to the page. He wanted to earn some money. That’s why he edited the videos and pictures. He said, “I posted edited pictures and videos on the page for a long time. But I haven’t been able to earn a single taka until today. I didn’t even get monetization.”
He has now stopped creating this type of content. He said, “I’m thinking of making some videos of myself, but I don’t have the people for that, so I can’t make videos.”
Imran says, “Everyone makes videos and pictures using AI, so tell them something, they edit and make a lot of bad pictures and videos.”
To review the activities of the ‘Cholo Bodle Jai’ page, Rumor Scanner analyzed all the posts published by the page last August. A total of 238 posts were made on the page in that single month. Every day in August, a minimum of five to a maximum of eleven posts were made on the page. Out of these, 135 posts were general (including ‘Good Morning’, ‘Good Night’, and various similar posts). The remaining posts contained various ordinary and objectionable photos and videos of women.Rumor Scanner’s analysis showed that 97 posts were made about 36 different women on the page in August. Out of these, Rumor Scanner found evidence that 79 posts circulated edited photos and videos of different women, falsely claiming them to be the women in question. Among these posts, 44 were photo-centric content and 35 were video-centric content. Furthermore, 18 posts whose original sources could not be traced were marked as unproven. However, through ‘Deepfake-o-meter‘, a joint project of the University of Buffalo and the National Science Foundation (NSF) in the United States for detecting AI and deepfake content, Rumor Scanner learned that all 18 of these contents had a probability of being edited by AI, ranging from a minimum of 46 percent to a maximum of 72 percent. Rumor Scanner confirmed that the remaining four contents (1, 2, 3, 4) were original photos/videos.
The analysis shows that the most-involved individuals in the content identified as fake are female stars from the Bangladeshi entertainment industry. Out of 36 women, 29 are from Bangladesh. Among them, 28 are entertainment celebrities. The two small-screen actresses who have been made victims of the most disinformation—six pieces each—are Tanjim Saiyara Totini and Tasnia Farin.
Besides Bangladesh, the page was also seen circulating eight pieces of disinformation involving entertainment celebrities from Kolkata, India. In August, the most disinformation (2 posts) was spread on the page involving Tollywood actress Nusrat Jahan.
Analyzing the page’s August content, it was found that the content of one woman is often used as the main source for circulating content involving another woman. This was observed in the cases of Tasnia Farin and Sadia Ayman. Two pieces of content circulated involving Farin were created by editing a video of an Indian woman named Rashi Singh. Similarly, two pieces of content circulated involving Sadia Ayman were created by editing a video of another Indian woman named Shayra Zilawat. Interestingly, the photo of an Indian woman named Divya Pachal was circulated once claiming to be Ashna Habib Bhabna and another time claiming to be Mahiya Mahi. All in all, in the 79 fake posts involving 36 women that month, evidence was found of misuse by editing the content of 76 different women.
Methodology
This research was conducted under the supervision of the Rumor Scanner team. For this purpose, the necessary content from the page named ‘Cholo Bodle Jai’ was documented, analyzed, and reviewed. Subsequently, the original content corresponding to the circulated content was found and compared. Assistance was also taken from an AI detection tool. Finally, the findings were presented through infographics and text.





