GitHub’s Deepfake Porn Crackdown still doesn’t work

“When we look at the abuse of intimate images, most of the tools and weaponized use have come from the open source space,” says Ajder. But they often start with well-intentioned developers, he says. “Someone creates something they think is interesting or cool and someone with bad intentions recognizes its malicious potential and weapons it.”

Some, like the repository disabled in August, have purpose-built communities around them for explicit uses. The model has positioned itself as a tool for deepfake porn, says Ajder, becoming a “funnel” for abuse, which mainly targets women.

More videos uploaded to the porn-streaming site from an account crediting AI models downloaded from GitHub presented the faces of famous deepfake targets, celebrities Emma Watson, Taylor Swift, and Anya Taylor-Joy, as well as other less famous but very real women, superimposed in sexual situations.

The creators freely described the tools they used, including two scrubbed by GitHub, but whose code survives in other existing repositories.

Deepfake-seeking perpetrators congregate in many places online, including secret community forums on Discord and on Reddit, compounding deepfake prevention attempts. A Redditor offers his services using the repository software archived on September 29. “Could anyone make my cousin,” another asked.

The torrents of the main repository banned by GitHub in August are also available in other corners of the web, which show how difficult it is to clean up. open-source deepfake software across the board. Other deepfake pornographic tools, such as the DeepNude app, have been similarly brought down before new releases come out.

“There are so many models, so many different forks in the models, so many different versions, it can be hard to track them all,” says Elizabeth Seger, director of digital policy at UK think tank Demos. “Once a template becomes open source publicly available for download, there is no way to make a public rollback of it,” he adds.

A deepfake porn creator with 13 explicit manipulated videos of female celebrities has credited a prominent GitHub repository marketed as a “NSFW” version of another project encouraging responsible use and explicitly asking users not to use it for the nudity “Learn all Face Swap AI available from GitHUB, without using online services,” says his profile on the tube site, cheekily.

GitHub had already disabled this NSFW version when WIRED identified the deepfake videos. But other repositories marked as “unlocked” versions of the model were available on the platform on January 10, including one with 2500 “stars”.

“It is technically true that once (a model is) out it cannot be reversed. But we can always make it more difficult for people to access,” says Seger.

If it is not checked, he adds, the potential harm of “porn” deepfake is not only psychological. Its knock-on effects include the intimidation and manipulation of women, minorities and politicians, as has been seen with political deepfakes affecting female politicians around the world.

But it’s not too late to get the problem under control, and platforms like GitHub have options, Seger says, including intervening at the point of upload. “If you put a model on GitHub and GitHub said no, and all the hosting platforms said no, for a normal person it becomes more difficult to get that model.”

The reinforcement in deepfake porn made with open source models is also based on policy makers, technology companies, developers and, of course, the creators of abusive content themselves.

At least 30 US states also have legislation addressing fake pornography, including bans, according to the non-profit Public Citizen. legislation trackeralthough definitions and policies are disparate, and some laws only cover minors. Deepfake creators in the UK are also soon to feel the force of the law after the government announced criminalize the creation of sexually explicit deepfakesand also the sharing of them, on January 7.