The ‘deepfakes’, digital creations used to harm women with fake pornographic videos

A shady place on the internet hides thousands of pornographic videos where the protagonists are famous, such as Taylor Swift or Mila Kunis. Despite the fact that their faces appear in the images, the body is not theirs. Like dolls, they were created with artificial intelligence, with a software that replaced the face of a porn actress with that of the artists.

Is about deepfakes, digital creations made with algorithms that learn through repetition and that after encoding dozens of images of faces from different angles can substitute the face and voice of one person for that of another.

They do not have to be harmful, they are also used in the film industry or in advertising, but the Sensity AI company estimates that between 90% and 95% of all deepfakes online are non-consensual porn and about 90% include women.

Silencing female leaders

In that virtual backyard, the vast majority of anonymous users use the faces of famous actresses or singers. Even so, the victims are also politicians or activists.

Agency EFE received in November a deepfake by activist Greta Thunberg showing the 18-year-old gesturing and dancing naked. The video was circulating in Brazil on WhatsApp and was accompanied by a message accusing her of “testing new tactics against climate change.”

With an internet search, EFE he found the video of the original actress whose face had been replaced by Thunberg’s.

Tips from the FBI to know if a video is a ‘deepfake’

Catalina Moreno, from Fundación Karisma, a Colombian organization that works to promote human rights in the digital world, points out that this type of deepfakes They seek to “silence, prevent women from being visible and in public debates”.

In the case of Thunberg, adds the expert, the deepfake wants to “question whether her activism in the environment is credible considering that there is a video circulating of her naked” and “leave her speech in the background, take away interest in the struggle she is advancing.”

Indian investigative journalist Rana Ayyub explains in an article how, after criticizing India for a rape in Kashmir of an 8-year-old girl, began to circulate a deepfake pornographic in which his face appeared and for which he suffered cyberbullying on the networks.

The consequences in the long run, he adds, were mental problems and self-censorship..

The UN ruled on the case, asking the Indian government to protect it from cyberbullying, but there were no repercussions for the perpetrators.

Fake videos can also put men’s lives at risk in countries where homosexuality is criminalized, in some with the death penalty.

Chile and Colombia, vigilant

Cybersecurity expert Juan David Cardona confirms EFE the presence of deepfakes in Latin America and reveals that the governments in Chile and Colombia, countries with elections last December and May 2022, respectively, monitor networks for attacks against policies.

Above all, since it is a technology that in recent years has taken a run and it is possible to create it with an application on the mobile.

Despite the deepfakes generated with free programs are not perfect, to discredit a woman it is enough that they are credible.

Weaponizing women’s sexuality is not new. A report of political violence on social networks in Ecuador collects the testimony of a politician who was advertised as a prostitute on a website and attached her phone number. Their campaign photo had been used in the ad.

Available to anyone

To generate deepfakes large amounts of images and videos are needed from different anglesThat is why until now it was difficult to use them against anonymous women. Even so, there are already applications that analyze photographs of women and with artificial intelligence they undress them in a few minutes.

Sensity AI found in 2020 a Telegram chatbot that had stripped at least 100,000 women. For its part, a study promoted by the European Parliament to address the deepfakes in European politics, it points out that apps are designed with gender biases.

“The ‘deepfake’ porn I was a victim of continue to give me nightmares”

The research points out that there is no regulation in most countries of the world and includes the examples of the US states of Virginia, Texas and California, the first in that country to enact laws on deepfakes.

Cardona insists on the importance of large technology companies taking action on the matter and stopping the flow of these videos on their platforms. Despite the fact that both the pornographic website PornHub and Twitter banned the deepfakes porno, EFE Verifica found that hundreds are hung on these platforms every day.

Some companies have developed programs to identify deepfakes, but they are not available to the user. Because, It is important to know some tips to determine if you are in front of a deepfake, such as noting if the gestures or the voice are delayed, if the color of the skin between the face and the body matches, or if the source from which the images come is reliable and public. (I)