Deepfake app to create nude pics of any woman was deleted
In recent days, a mobile app capable of “striping” a person just by taking a photograph went viral. After the debate among its fans and detractors, the app was removed from all platforms; according to cybersecurity services specialists, the creators mentioned that they could not allow their idea to be used “for harmful purposes”.
The app, called DeepNude, takes image editing
too far, as it contains a machine
learning algorithm to remove the person from the image’s clothes just in
seconds (a practice known as deepfake).
The app works especially efficiently when it comes to images of women, as
DeepNude neural network had records of millions of female nude images.
According to cybersecurity services
specialists, the free version of the app generated fake nudes and then added a
watermark to the image, so users had to pay about $50 USD to get the Premium
version, with which the watermark disappeared.
To defend against the controversy generated by
this app, the creators stated: “This project was born only as a form of
entertainment; every month some sales were made, always in a controlled way. To
be honest, the app doesn’t always work that well; you only get the expected results
with a specific type of image”.
DeepNude’s popularity suddenly increased last
weekend and apparently this was not a variable contemplated by the developers,
as the app’s Twitter account mentioned multiple reports of failures and falls
in the service. “We never thought it would go viral, we lost control of
the traffic.”
However, the decision to remove the app has
nothing to do with technical issues, experts in cybersecurity services mention.
“Increasing the number of users also increases the likelihood that our app
will be used for malicious purposes; it’s not the way we want to work,”
the developers said.
Specialists from the International Institute of
Cyber Security (IICS) claim that this does not necessarily end the problem, as
there must already be multiple copies of DeepNude circulating around hacking forums
or pirate application sites, so we could keep listening about this app in the
future.
(Visited 3 1 times)
Gloss