Connect with us

Business

DeepNude App Banned on GitHub After Spreading to Multiple Platforms

Published

on

  • Reuploaded replicas of the app DeepNude have been popping up on social media platforms including Twitter, YouTube, and Reddit.
  • The app, which removed clothing from pictures of women to make them look naked, had previously been removed by its creator after an article published by Vice’s technology publication Motherboard created backlash. 
  • Discord and GitHub have since banned replica versions of the app after it was spread on their sites.
  • Over the last week, dozens of women in Singapore have had pictures from their social media accounts doctored and put on porn websites. Those pictures are believed to have been made with a version of the DeepNude App.

DeepNude App Explained

The open source software platform GitHub has banned all code from the controversial deepfake app known as DeepNude, a desktop application that removes clothing from pictures of women and generates a new photo of them appearing naked.

The app was originally released last month, but it did not receive notoriety until Vice’s tech publication Motherboard broke the story several days after it launched. The day after Motherboard’s exposé, the DeepNude creators announced they were pulling the app.

“The probability that people will misuse it is too high,” the creators said in a statement on Twitter. “Surely some copies of DeepNude will be shared on the web, but we don’t want to be the ones who sell it.”

“The world is not yet ready for DeepNude,” the statement concluded.

GritHub Bans DeepNude Replicas

Apparently, the world thought otherwise, because copies of the DeepNude app were shared and still are being shared all over the internet. 

The program was an app that was meant to be downloaded for use offline, and as a result, it could be easily replicated by anyone who had it on their hard drive.

That is exactly what happened. People who replicated the software reuploaded it on various social media platforms, like GitHub, which banned the app for violating its community guidelines.

“We do not proactively monitor user-generated content, but we do actively investigate abuse reports,” a GitHub spokesperson told Motherboard. “In this case, we disabled the project because we found it to be in violation of our acceptable use policy. We do not condone using GitHub for posting sexually obscene content and prohibit such conduct in our Terms of Service and Community Guidelines.”

According to The Verge, the DeepNude team itself actually uploaded the core algorithm of the app to GitHub.

“The reverse engineering of the app was already on GitHub. It no longer makes sense to hide the source code,” The Verge said the team wrote on a now-deleted page. “DeepNude uses an interesting method to solve a typical AI problem, so it could be useful for researchers and developers working in other fields such as fashion, cinema, and visual effects.”

However, Rogue Rocket was still able to find at least one GitHub repository that claimed to have DeepNude software for Android.

“Deep nudes for android. this is the age of FREEDOM, NOT CENSORSHIP! hackers rule the future!” the page’s description said. 

DeepNude Spreads

GitHub was not the only platform that the replicated app was shared on. 

Even with just a cursory search on Twitter, Rogue Rocket was able to locate two Twitter accounts that provided links to replicated versions of the app. One of the accounts links to a website called Deep Nude Pro, which bills itself as “the official update to the original DeepNude,” and sells the app for $39.99.

The other account links to a DeepNude Patreon where people can either download the app or send the account holder pictures they want to generate and then buy.

When Rogue Rocket searched YouTube, there appeared to be multiple videos explaining how to download new versions of the app, many of which had links to download the app in the description.

Others have also shared links on Reddit, and The Verge reported that links to downloads were being shared on Telegram channels and message boards like 4chan.

To make matters even worse, a lot of the replicated software includes versions that claim they removed the watermarks included in the original app, which were used to denote that the generated pictures were fake.

While it has been reported that a lot of the links to the reuploaded software are malware, download links to the new versions are still incredibly easy to find.

GitHub is also not the only platform to ban the app. According to Motherboard, last week Discord banned a server that was selling what was described as an updated version of the app, where customers could pay $20 in Bitcoin or Amazon gift cards to get “lifetime access.”

Source: VICE

The server and its users were removed for violating Discord’s community guidelines.

“The sharing of non-consensual pornography is explicitly prohibited in our terms of service and community guidelines,” a spokesperson for Discord told Motherboard in a statement.

We will investigate and take immediate action against any reported terms of service violation by a server or user. Non-consensual pornography warrants an instant shut down on the servers and ban of the users whenever we identify it, which is the action we took in this case.”

DeepNude App Used in Singapore

The rapid diffusion of the app on numerous social media platforms has now become an international problem.

On Wednesday, The Straits Times reported that over the past week “dozens of women in Singapore” have had pictures of them taken from their social media accounts and doctored to look like they are naked, then uploaded to pornographic sites.

Those photos are believed to have been doctored using a version of the DeepNude app, which have been shared via download links on a popular sex forum in Singapore.

Lawyers who spoke to The Straits Times told them that doctoring photos to make people look naked is considered a criminal offense in Singapore.

Even though the artificial intelligence aspect is new, one lawyer said that the broad definitions under the law could allow people to be prosecuted for doing so.

Another lawyer backed that up, saying that under Singapore’s Films Act, people who make DeepNude pictures can be jailed for up to two years and fined up to $40,000. They can also be charged with insult of modesty and face a separate fine and jail term of up to a year. 

Legal Efforts in the U.S.

The legal precedent in Singapore raises questions about laws that regulate deepfakes in the United States. While these efforts appear stalled on the federal level, several states have taken actions to address the issue.

On July 1, a new amendment to Virginia’s law against revenge porn, that includes deepfakes as nonconsensual pornography, went into effect. Under that amendment, anyone caught spreading deepfakes could face 12 months in prison and up to $2,500 in fines.

The idea of amending existing revenge porn laws to include deepfakes could be promising if it is effective. According to The New York Times, as of early this year, 41 states have banned revenge porn.

At the same time, lawmakers in New York state have also proposed a bill that would ban the creation of “digital replicas” of individuals without their consent. 

However, the Motion Picture Association of America has opposed the bill, arguing that it would “restrict the ability of our members to tell stories about and inspired by real people and events,” which would violate the First Amendment.

The opposition to the law in New York indicates that even as states take the lead with deepfake regulation, there are still many legal hurdles to overcome.

See what others are saying: (VICE) (The Verge) (The Strait Times)

Advertisements

Business

Amazon and Instacart Workers Launch Strike, Demanding Safer Conditions During Pandemic

Published

on

  • Amazon workers in Staten Island are staging a walkout, demanding that the warehouse be thoroughly cleaned. This comes after workers say not enough measures sanitary measures were taken when a coronavirus case was confirmed at that facility, but Amazon says it has increased deep cleanings.
  • Instacart workers nationwide are also striking, saying they will not fulfill orders until they receive sanitation supplies, hazard pay, and better access to paid sick leave.
  • Delivery service workers have been facing uphill battles when it comes to sick leave, with many companies only offering 2 weeks paid if an employee tests positive for the virus, despite tests being far and few between.

N.Y. Amazon Workers Strike

Amazon and Instacart workers are striking, demanding their respective companies give them tools to work in safer and cleaner conditions as they become essential figures during coronavirus lockdowns.

Frustrations at an Amazon facility in Staten Island, New York grew after one of the workers there tested positive for coronavirus. Employees have been concerned that not enough safety measures were taken after this, and are demanding during a Monday walkout that the building be thoroughly cleaned while they are not present. 

“The plan is to cease all operations until the building is closed and sanitized,” employee Christian Smalls, who is actually in a 14-day precautionary quarantine recommended by Amazon, told CNN. “We’re not asking for much. We’re asking the building to be closed and sanitized, and for us to be paid [during that process].”

Early counts suggest that around 100 workers attended the walkout. Videos show participants carrying signs, with many standing apart from one another to practice social distancing. Some signs contained phrases like “Our Health Is Also Essential.”

Smalls also told CNN that Amazon is not being transparent with the public about how many workers at the Staten Island warehouse have tested positive. He believes that the facility, which he called “breeding grounds for this pandemic,” could have as many as seven cases. 

An Amazon spokesperson told CNBC that Smalls claims were “misleading” and that the facility was being deep cleaned on an increased basis. Amazon as a whole is also giving those who are diagnosed or those who come into contact with someone diagnosed with the virus an extra two weeks paid sick leave so they can quarantine. Workers are also seeing a pay boost of $2 an hour through April.

However, this Staten Island facility is just one of many Amazon locations seeing a number of issues amid the coronavirus outbreak. At least 13 Amazon warehouses have reported confirmed cases of the coronavirus. A warehouse in Queens was also temporarily closed after a case was confirmed there. According to CNBC, workers at numerous facilities have been forced to ration essential things like hand sanitizer and disinfectant wipes, if there even are any available.

Instacart Strike and Delivery Workers

Workers at Instacart are staging a nationwide strike of their own starting Monday. Contractors for the grocery delivery service say they want increased hazard pay of $5 per order, a better tipping system, more paid sick leave, and to be provided sanitation supplies like disinfecting wipes and hand sanitizer. Some of their gig workers say they will not fulfill orders until their demands are met. 

In response, Instacart has said they will distribute hand sanitizer and change its tipping settings. This is still not enough for their workers, who go into crowded grocery stores every day so people in lockdowns can stay inside. 

“Actions speak louder than words. Instacart worker Sarah Polito told NPR. “You can tell us that we’re these household heroes and that you appreciate us. But you’re not actually, they’re not showing it. They’re not taking these steps to give us the precautions. They’re not giving us hazard pay.”

Instacart workers are among many delivery service workers who do not feel their employer is properly responding to the coronavirus. While companies like DoorDash, Postmates, Uber and more have given two weeks paid sick leave to workers diagnosed with the coronavirus, employees are still left in a tricky place because here are just not enough tests. Employees who think they might have COVID-19 but cannot access a test are out of luck.

One DoorDash worker told the L.A. Times that after he felt shortness of breath and had a cough, a doctor wrote him a note saying he should quarantine for two weeks.

“Patient may return to work on April 3, 2020 pending management of pain and symptoms,” the note read. “Patient is instructed to self quarantine to avoid acquiring viral illness or exposure to others.”

Upon receiving this note, DoorDash denied his sick pay request because the doctor did not outright mention the coronavirus. He was then suspended for two weeks without pay for safety reasons.

Support for Strikes

Because so many workers feel they are not getting the benefits they deserve during this outbreak, there was a lot of support for workers at Amazon and Instacart striking. Online, many encouraged people to not use those services to show solidarity with the workers. 

Rep. Alexandria Ocasio-Cortez also tweeted about it. “One of the best ways to thank essential workers is to support the fight to improve their lives,” she wrote. 

See what others are saying: (Forbes) (Reuters) (Vice)

Advertisements
Continue Reading

Business

Zoom’s Sudden Popularity Draws Attention to App’s Privacy Risks

Published

on

  • As more and more people use Zoom for virtual gatherings, several have raised concerns about privacy issues in the app.
  • One issue is that meeting hosts have the ability to save meetings to a cloud and monitor some behavior of attendees.
  • Many using the app have also experienced “zoombombers,” which are trolls making their way into calls, showing graphic and explicit content. 
  • Zoom has responded to one major criticism: its ability to share data with Facebook. Vice’s Motherboard found that the app could do so on Thursday and by Friday, Zoom got rid of that code.

Host Capabilities

As video chatting app Zoom increases in popularity while students and employees work from home, critics are afraid the app may have glaring privacy issues that users are unaware of. 

Zoom has become widely-used since millions of people across the country were forced inside because of the coronavirus. From meetings, to lectures to virtual boozy Sunday brunches, it has become the app of choice for video chatting in quarantine. Even Prime Minister Boris Johnson has used it to conduct government meetings in the U.K.

Calls on the app can be set up by a “host” who initiates scheduling the call, but many allege that these hosts are given too much power on Zoom. The app offers tools that, depending on the subscription tier-one belongs to, allow hosts to access what some may consider private information. 

One feature called “attention tracking” lets the host of a meeting see if an attendee does not have Zoom in focus for more than 30 seconds. This means that if an attendee is active in a window other than Zoom– to look at other documents, message a colleague, or watch the world collapse live on Twitter for 30 seconds– the host is made aware of this. They don’t see what the attendee is specifically doing, just that the Zoom window has become inactive. 

Still, the idea of this happening while you could be completely unaware has made a lot of people uneasy. Justin Brookman, director of privacy and technology policy at Consumer Reports  said this kind of feature should not exist. 

“If you’re teleworking on a home computer, your boss shouldn’t be able to monitor what’s on your screen,” he said in an article on Consumer Reports. “Zoom should get rid of attention tracking mode, or at the very least make participants aware when it’s on.” 

And this isn’t the only thing hosts can do that some see as potentially dangerous. There are several options that allow Zoom meetings to be recorded. One that some find particularly concerning is cloud recording, which is exclusively for paid subscribers and can only be done by hosts. It allows the video, audio, and a transcription of the meeting to be stored in the Zoom cloud. From there it can be accessed and downloaded by authorized employees at a company so that people who were not part of the meeting can read or watch it back. 

“Zoombombing”

Zoom’s issues extend past the powers a host has. There have also been reports about trolls being able to hack into Zoom meetings, something that has been called “zoombombing.” According to a report from TechCrunch, zoombombers are hopping into meetings and showing graphic content like pornography or violent imagery.

In one case, a public Zoom Work From Home Happy Hour was attacked with sexually explicit video and images. Despite the hosts’ many attempts to boot the zoombomber out of the meeting, they were able to re-enter under a new name. To stop this from happening, the hosts had to end the call.

That’s not the only time something like this has happened. NBC talked to a couple that read children’s books to kids stuck at home via Zoom. Ruha Benjamin, an associate professor of African American studies at Princeton University, was leading the call and told NBC that while she was reading to the kids, an image of a “chubby white man in a thong” popped up.

At first, she did not know if everyone could see it, but then a male voice began to repeatedly say the n-word for all 40 kids on the call to hear. She then had to shut the call down and told the outlet, “we knew it was a malicious, targeted thing. My husband and I are both African American.”

Virtual classrooms, religious services, and various other places have also been targets of this kind of harassment. Zoombombers have the ability to do this for a couple of reasons. First, if a Zoom call is public or if the link has been made public, anyone who wants to join can. Second, Zoom’s default settings allow anyone in a call to get screen time. A host does not need to grant an attendee access. Some of this can be changed in Zoom’s advanced settings if a user knows to look for it, but otherwise, this is the way the app will do things on its own.

Entrepreneur Alex Miller shared a Twitter thread giving tips on how to best protect your Zoom calls from hackings like this. 

You can disable the “join before host” feature so that no one can enter a chat and do something inappropriate without the host knowing. Zoom users can also add a co-host so that multiple people can remain on guard. Screen sharing can also be changed to host only.

On top of this, users can also disable file transfers and prevent removed people from joining the call again.

Info Sharing With Facebook

Zoom has also responded to another issue that was found within the app. A Thursday report from Vice’s Motherboard found that Zoom could send data to a company that is perhaps most well known for data privacy controversies: Facebook. This could happen even if you don’t even have a Facebook account.

One day after this report came out, Zoom removed the code that allowed this. According to Motherboard, Zoom would connect to Facebook’s Graph API, which is the main way developers get data in or out of Facebook. Zoom would then notify Facebook when a user opens the app and give details on the device they are doing so from, including the model, location, phone carrier, and a “unique advertiser identifier created by the user’s device which companies can use to target a user with advertisements.” Nothing in their privacy policy explicitly addressed this. 

When Zoom told Motherboard they were getting rid of this code, they explained that the issue had to do with their SDK, or software development kit, which is a bunch of code that can be used to implement app features, but can also send data to third parties.

“Zoom takes its users’ privacy extremely seriously,” they said in a statement to Motherboard. “We originally implemented the ‘Login with Facebook’ feature using the Facebook SDK in order to provide our users with another convenient way to access our platform. However, we were recently made aware that the Facebook SDK was collecting unnecessary device data.”

Zoom also confirmed that the information being collected was not personal user information, but device information, which lined up with Motherboard’s findings. 

See what others are saying: (The Guardian) (Forbes) (BBC)

Advertisements
Continue Reading

Business

TikTok Suppressed Content From “Ugly,” Poor, and Disabled Users, Report Says

Published

on

  • A report from The Intercept claimed that in an effort to attract new users, TikTok had policies in place for its moderators to suppress content from users deemed “ugly,” poor, or disabled.
  • The documents also showed that TikTok outlined bans to be placed on users who criticized “political or religious leaders” or “endangered national honor.”
  • Sources said the policies were created last year and were in use as recently as the end of 2019.
  • A TikTok spokesperson said the majority of the guidelines were never in use or are no longer in use, but the ones targeting users’ appearances were aimed at preventing bullying.
  • However, the documents reviewed by The Intercept do not explicitly mention anti-bullying efforts.

Leaked Policies

Newly released documents reveal that TikTok creators directed their moderators to censor posts from users believed to be poor, disabled, or “ugly,” among other guidelines.  

The leaked policies were first reported by The Intercept on Monday, exposing an inconsistency within the highly popular video-sharing app whose tagline is “Real People. Real Videos.” However, based on this recently-exposed information, it seems like TikTok only wants to funnel certain types of “real people” on the “For You” feed, its page dedicated to promoting select content to its millions of users. 

The Intercept noted that the documents appear to have originally been printed in Chinese — the language of the app’s home country — but had been translated into sometimes-choppy English for global distribution. Of the multiple pages of policies the news outlet posted, one outlines characteristics that the app considers undesirable such as “abnormal body shape, chubby, have obvious beer belly, obese, or too thin.” 

The rules also encourage restrictions of “ugly facial looks” including wrinkles, noticeable scars, and physical disabilities. Criteria for the backgrounds of videos were also included in the policies, discouraging “shabby and dilapidated” environments including slums, dirty and messy settings, and old decorations. 

As far as the reasoning for these guidelines, TikTok wrote: “If the character’s appearance or the shooting environment is not good, the video will be much less attractive, not [worthy] to be recommended to new users.” 

A spokesperson for the app told The Verge that the guidelines reported by The Intercept are regional and “were not for the U.S. market.”

The other policies that The Intercept released detail more types of content that should be banned across the platform, including defamation or criticism of “civil servants, political or religious leaders,” as well as family members of these leaders. Moderators were instructed to punish any users who “endang[er] national honor” or distort “local or other countries’ history,” using May 1998 riots in Indonesia, Cambodian genocide, and Tiananmen Square incidents as examples.

The Intercept reported that sources told them the policies were created last year and were in use until at least late 2019. 

TikTok’s Response

A spokesperson for the app told The Intercept that “most of” these exposed rules “are either no longer in use, or in some cases appear to never have been in place.”

The spokesperson also told the outlet that the policies geared toward suppressing disabled, seemingly impoverished, or unattractive users “represented an early blunt attempt at preventing bullying, but are no longer in place, and were already out of use when The Intercept obtained them.”

These intentions have been pushed by the platform in the past — in December, TikTok admitted that at one point they prevented the spread of videos from disabled, LGBTQ, or overweight users, claiming it was an attempt to curb bullying. 

A TikTok spokesperson told The Intercept that these newly-released policies “appear to be the same or similar” as the ones revealed in December, but the guidelines published this week are notably different — they don’t mention anti-bullying motives and instead focus on how to appeal to more users. 

Criticism of TikTok’s Moderation and App’s Response

TikTok has faced scrutiny in the past for appearing to censor certain content, including pro-democracy protests in Hong Kong and criticism of the Chinese government.  

It’s also worth noting that the app has been under fire for its data-sharing policies and the U.S. government has even suggested this is a national security threat. 

TikTok said this week that it will stop using China-based moderators to review overseas content, noting that these employees hadn’t been monitoring content in U.S. regions. 

And in further attempts to counter the criticism of their moderation tactics, TikTok announced last week that it plans to open a “transparency center” in Los Angeles in May. This center will allow outside observers to better understand how the platform moderates its content.

See what others are saying: (The Intercept) (The Verge) (Business Insider)

Advertisements
Continue Reading