Deepfakes and the Sexual Harassment of Women in Bangladesh

Across Bangladesh, a quiet and devastating crisis of superimposing any woman's face into pornographic contents is unfolding. Artificial intelligence has made it trivially easy to do this. This technology is called "deepfake". The targets are overwhelmingly women. And the consequences — social ruin, forced withdrawal from public life, and in the worst cases, death — are being absorbed almost entirely by the victims.

The Targets

The victims do not fit a single profile. They are students, activists, professionals, politicians, actresses, and anonymous private individuals.

In one of the most devastating cases documented in Bangladesh, a woman took her own life after an AI-edited video of her was shared with her family. The perpetrator understood precisely how Bangladeshi social structures work — family honor, community judgment, and the irreversibility of digital shame — and used a fabricated video to trigger all three at once. Her death was not an accident of technology. It was the intended result of its deliberate misuse. (The Daily Star)

Farzana Sithi is a student activist who rose to public attention during Bangladesh's 2024 quota reform movement. Between December 2024 and March 2025, a fake Facebook profile created in her name uploaded 133 confirmed AI-generated deepfake videos featuring her face superimposed onto explicit contents. The profile accumulated over 4.1 million views. It was monetized. The creator was earning Facebook Stars revenue from her humiliation. She eventually filed a police case in March 2025. The fake profile remained active.

Riya, a pseudonym used in academic research, is a university student in Rajshahi. Some peers circulated sexually explicit images with her face manipulated onto them through student networks. The content spread to friends and relatives. She was pressured to resign from every student organization she belonged to. Her mother told her to quit her studies and leave campus. Riya wanted legal help but feared that filing a complaint would draw media attention and amplify the harm. She did nothing. The content remained. (Perception of Deepfakes Among Bangladeshi Women, arXiv, Jan 2026)

Syeda Rizwana Hasan, an advisor to Bangladesh's Ministry of Environment, was targeted in early 2025 when a doctored image — her face placed on a body from an adult website — was circulated widely on Facebook through an account called "Chemical Ali," a repeat offender known for targeting prominent women. (VOICE Bangladesh, Aug 2025)

According to Strengthening Resilience Against Technology-Facilitated Gender-Based Violence (TFGBV) and Promoting Digital Development data, 89% of women social media users in Bangladesh have experienced online violence at least once.

The Perpetrators

The profile of the deepfake abuser in Bangladesh is varied and, in some cases, deeply unsettling.

The most documented type is the extortionist — usually a man, often someone the victim knows or has encountered online. He downloads photos from social media, uses a deepfake tool to generate explicit contents, and then contacts the victim with a threat: pay, comply, or the video goes to your family, your college, your employer.

It might also be an ex-partner, a rejected suitor, a neighbor, a classmate.

A third type is the content monetiser — someone who produces deepfakes of visible women for audience engagement and platform revenue, as in the Farzana Sithi case. The harassment is incidental to the financial motive.

And increasingly there is the politically motivated harasser — those who use deepfakes to silence women who speak publicly, enter politics, or challenge male authority. VOICE Bangladesh's 2025 report documents how coordinated digital attacks against women activists and female government advisors are deployed not just to embarrass, but to drive women out of public life entirely.

Globally, the numbers are staggering: 98% of deepfake videos online are sexually explicit, and 99% of those target women and girls. Since 2019, deepfake content has grown by 550%. (Global Partnership for Action on Gender-Based Online Harassment and Abuse, Dec 2025)

What Happens to the Victims

In Bangladesh's social context, an explicit video, real or fabricated, carries consequences that are immediate, severe, and often permanent. Families demand students leave academy. 

Marriages collapse or are cancelled. Employers terminate contracts. Neighborhoods ostracize. And even when a woman insists the content is fake, the burden of proof falls entirely on her, in a cultural environment primed to disbelieve her.

The psychological toll is severe. Victims report anxiety, depression, withdrawal from social life, academic failure, and forced career abandonment. In documented cases, women have taken their own lives.

What to Do

If someone is targeted by a deepfake, they should act immediately and methodically. Every step they take to document and report is evidence.

Preserving evidence first. Should take screenshot of the content, the profile name, the URL, and the date and time of every post. Digital content can be deleted within hours, and evidence not preserved in time is inadmissible in court.

Reporting to the platform. On Facebook, the victim report the content as non-consensual intimate imagery and report the fake profile separately. This creates an official record even if removal is slow.

Calling the Police Cyber Support Centre for Women (PCSW). The victim can also report through the Bangladesh Police website. 

Filing a formal case under Cyber Security Ordinance 2025 and Pornographic Act.

Contacting a civil society organisation. Bangladesh Mahila Parishad, Digitally Right, and ActionAid Bangladesh can provide legal guidance and support.

Ishtiaque Foysol is a Security Researcher and Trainer at Decodes Lab Ltd.