By Hafiz Rahat Usama 
The Saree Portrait Trend’s Disturbing Revelation
What began as an innocent social media trend across South Asia – women uploading photos to create AI-generated saree portraits – took a sinister turn when one Pakistani woman discovered her generated image included a mole that wasn’t visible in her original photo. This unsettling incident exposed how artificial intelligence can infer and reveal personal details without consent, raising serious concerns about privacy violations in the digital age.
How AI Systems Learn Too Much
Generative AI platforms like Google’s Gemini analyze vast datasets of online images, learning to recognize patterns and predict details beyond what’s visible in source photos. While this capability demonstrates AI’s sophistication, it also creates alarming privacy risks, particularly in countries like Pakistan where data protection laws remain inadequate. The unpassed Personal Data Protection Bill, modeled after Europe’s GDPR, leaves citizens vulnerable in a legal vacuum regarding AI applications.
The Epidemic of Digital Violence Against Women
Across South Asia, AI technology is being weaponized against women at an alarming rate. Deepfake pornography has evolved from a niche threat to a widespread danger, with victims discovering manipulated explicit images circulating online. A Pakistani content creator recently faced public shaming when her Instagram photos were altered into fake nude images, while in India, women journalists found themselves digitally “auctioned” with their faces superimposed on pornographic content.
Legal Gaps and Societal Challenges
Digital rights expert Sadaf Khan notes that while Pakistan’s PECA law addresses some digital crimes, it fails to adequately cover AI-related violations. The situation is compounded by societal attitudes that often blame victims rather than perpetrators. “Deepfakes blur reality, exposing women to safety threats and stigma,” Khan explains, emphasizing that legal protections alone cannot prevent the psychological and social consequences of such violations.
Building Defenses Against Digital Threats
Protecting women in the digital space requires a multi-pronged approach:
- Education must teach digital literacy that includes understanding AI risks
- Legal reforms should urgently pass data protection laws with AI-specific provisions
- Regional cooperation could establish protocols to combat cross-border digital crimes
- Personal precautions like avoiding high-resolution uploads can reduce vulnerability