Deepfakes Hit Journalists in 27 Countries, With Women the Main Targets

PARIS — Reporters Without Borders said it documented 100 journalists targeted by deepfakes in 27 countries between December 2023 and December 2025, with women accounting for 74% of the victims.

The press freedom group, known by its French initials RSF, said 13% of the women in its sample were targeted with pornographic deepfakes, which it said were often used as part of broader harassment campaigns.

The figures, published by RSF in February, add detail to a threat that has moved from election warnings and celebrity scams into newsroom security. The harm described in the report is not limited to reputational damage. RSF said some journalists reduced their public work, changed production methods or faced police complaints and audience abuse after fake videos circulated online.

Cristina Caicedo Smit, a Voice of America reporter, discovered in February 2025 that her voice and image had been replicated in videos posted on X, according to RSF. The group said the fabricated clips were designed to portray VOA as an activist outlet opposed to President Donald Trump.

RSF said the case was part of a wider pattern in which synthetic audio and video were used to manipulate public debate by borrowing the authority of known presenters and reporters. It also cited a fake clip of Portuguese journalist Pedro Benevides and a campaign targeting South African broadcaster Leanne Manas.

In Manas's case, RSF said deepfakes were used in bogus advertisements for pharmaceutical products and cryptocurrency schemes. The organization said members of the public later confronted her at work or sent messages demanding compensation after they believed and acted on the fake endorsements.

The technology has been discussed differently across regions. In Europe and North America, it is often framed as a platform governance and election integrity problem. In parts of Latin America, Africa and South Asia, the same tools are showing up as direct attacks on named journalists, especially women, who then face both reputational damage and personal harassment.

The United Nations has echoed that concern. Its page for the International Day to End Impunity for Crimes Against Journalists says women journalists are increasingly exposed to AI-driven threats including gendered disinformation, surveillance, deepfakes and harassment.

CPJ, the Committee to Protect Journalists, said in material published for International Women's Day 2026 that gender continues to shape how journalists experience threats, with women facing coordinated online harassment and digital abuse alongside the broader risks attached to reporting.

What makes deepfakes harder to contain is that debunking does not always work. RSF quoted Benevides saying some viewers continued to believe the message in a fake video even after he said publicly that it was fabricated.

The legal response remains uneven. RSF cited the case of Slovak journalist Monika Todová, who filed a defamation complaint after a fake audio clip alleged election fraud planning, only for the investigation to stall before police closed the case without identifying the perpetrator.

That mismatch between speed and accountability is becoming central to the story. The synthetic clip can be made in hours and distributed in minutes. Verification, takedowns and court action usually move much more slowly.

For journalists, the burden is not only technical. RSF said some victims described the experience as seeing themselves in "a different reality." Some cut back appearances or redesigned how they worked online in an attempt to reduce the source material available for future fakes.

The issue also reveals a gap in how the same attack is understood. In many newsrooms in Paris, London or Washington, a deepfake is treated as a fact-checking and moderation problem. For a local presenter in Johannesburg, Buenos Aires or Kinshasa, it can become a workplace security problem, a family problem and a source of direct financial claims from strangers.

RSF's figures do not show how many cases go unreported, and the organization said its tally was not exhaustive. The next phase of the issue will depend on whether platforms, police and lawmakers can speed up response systems before synthetic impersonation becomes routine in political crises and daily news work alike.