AI-assisted harassment is growing, with deepfakes and manipulated content used to undermine women’s professional credibility; Italian Prime Minister Giorgia Meloni faced a campaign involving sexually suggestive AI-generated images and warned that many women lack the resources to respond.

The scale is significant: a UN Women survey across 119 countries found that nearly one in four women in public-facing roles had experienced AI-assisted online violence.

In response, Article 50 of the EU AI Act took effect in summer 2026, requiring providers to mark synthetic audio, images, video and text in machine-readable, detectable formats.

Several approaches are emerging: Apple, Nikon and Canon offer on-device image verification; OpenAI, Google and Anthropic add machine-readable watermarks to outputs; and the C2PA open standard records a verifiable chain of custody and file changes.

These authentication features from companies including Google, OpenAI and Apple are designed to make digital content’s origin and authenticity easier to verify, potentially helping protect women’s identities and reputations.

Platforms such as YouTube already have programs to detect public-figure impersonations and undisclosed AI-generated content, and the article argues that EU transparency requirements could accelerate such tools, with particular benefits for women facing AI-enabled harassment.