Brazil: AI deepfakes pose "excessive risk" to women and girls

A report released on Tuesday, April 28, 2026, by the Brazilian research institute InternetLab warns that the proliferation of AI-generated deepfakes has become an "excessive risk" that disproportionately targets women and girls.

The document, highlighted by Global Voices Advox, argues that current artificial intelligence technologies are being weaponized to create non-consensual sexual content, reinforcing historical patterns of gender-based violence in the digital sphere, News.Az reports, citing ADVOX.

Citing research from Security Hero, the report notes that 98% of all deepfake videos online are sexually explicit, and 99% of those depicted are women. In Brazil, this technology has moved beyond speculative harm to real-world abuse; in 2025, the country amended its criminal code to increase penalties for psychological violence against women involving AI-manipulated imagery. The report also highlights a growing crisis in schools across Latin America, where female students are increasingly victimized by classmates using easily accessible AI apps to generate and circulate explicit images.

InternetLab's analysis calls for a shift in how AI risks are classified, urging regulators to view these tools not just as technical glitches but as products designed with business models that prioritize profit over the safety of vulnerable groups. The organization emphasizes the need for "safety-by-design" principles and greater corporate responsibility from digital platforms, arguing that without independent audits and transparency mechanisms, AI will continue to amplify digital defamation and sexual extortion at an unprecedented scale.

News.Az 

By Leyla Şirinova