|Deepfake technology has really taken a turn for the worst in recent years / Photo by: Igor Stevanovic via 123RF|
Bill Hader sitting in an interview with David Letterman about his being part of the “Tropic Thunder” movie may not mean much if you’re not a fan of the comedian. A user on YouTube called “Ctrl Shift Face” recently gave the old video new life.
The interview was dated 2008, but it was only in 2019 that “Ctrl Shift Face” changed it, leading to its popularity. The video is a prime example of how deepfake technology truly presents itself to be both impressive and deeply heinous. In the video, we see Bill Hader talking enthusiastically about being part of the “Tropic Thunder” project, but every time he is asked about his thoughts on being in a film with Tom Cruise himself, his face itself becomes Tom Cruise.
What’s more, the face falls off when he stops answering Letterman’s Cruise-related questions, only to morph back seamlessly into Cruise’s face when Hader tells another anecdote related to him. Many in the comments are baffled and have expressed time and time again how the video had made them deeply uncomfortable. Deepfake technology has really taken a turn for the worst in recent years.
What are Deepfakes?
Barring Hader’s innocence in all this creepy technology, what is a deepfake and how does it work?
Deepfake technology, specifically video technology, in this case, is the “use of artificial intelligence to generate depictions of events that never actually happened.” It uses the vast river of information that the internet provides and uses techniques such as superimposing and imitating voices to do what YouTuber “Ctrl Shift Face” was able to do with Bill Hader with Tom Cruise’s face.
As reported by business news site Forbes, deepfakes are already troubling many lawmakers and technologists because these can be used in something more heinous than just Hader and his Cruise impression.
The Rise of the Deepfakes
Deepfakes are rising fast and almost all of it is happening beneath all of our noses. According to a report by the business magazine Fast Company, Giorgio Patrini’s Deep Trace Labs found, through an extensive examination of website, forums, and services, that there are currently 14,678 deepfake videos on the internet. If you think that’s scary, the research also found that 96% of those videos were pornographic.
Those most affected by these are women, especially Western female actresses. Over 90% of these deepfake videos featured Western female actresses.
The same thing is also being done to South Korean K-pop singers, who get the same terrifying treatment.
The reason why so many haven’t taken the issue seriously is that deepfakes which use Elon Musk, Donald Trump, or Nicholas Cage water down the issue so it feels comedic, but when it’s the name and reputation of someone innocent on the line, the lines of comedy and crime are blurred.
|Deepfake technology, specifically video technology, in this case, is the “use of artificial intelligence to generate depictions of events that never actually happened” / Photo by: Yulia Grogoryeva via 123RF|
Facebook, Snapchat, and TikTok’s Moves on Deepfake
Social media sites are where most of us stay on the internet, and in this case, they seem to be the places where most of these videos are shared. In an attempt to bar deepfakes from reaching the forefront of the internet through these sites, Facebook has begun its crackdown on deepfake videos, issuing a statement that they will begin removing deepfakes as well as “heavily manipulated media” from the platform.
According to American news and opinion website Vox, Facebook said videos that are edited in ways that would “likely mislead someone into thinking a subject of a video said words that they did not actually say” and AI and machine learning technology that “merges, replaces, or superimposes content onto a video, making it appear to be authentic” will all be taken down.
Generally, it is seen as a step in the right direction, but Amrita Khalid of the news site Quartz worries that it may not be enough, especially when other misleading videos still run rampant on the website.
Still, their efforts are much more commendable than that of Snapchat or TikTok, which should worry people: most of our youth spend time on those platforms. While Facebook is trying to curb these videos, Snapchat and TikTok’s move was to embrace it, even though their audience is not aware they are.
|Social media sites are where most of us stay on the internet, and in this case, they seem to be the places where most of these videos are shared / Photo by: scyther5 via 123RF|
According to Forbes, the popular use of filters on sites like Snapchat is already a form of embracing deepfake technology, and it’s being used unwittingly, every day, by Snapchat’s devoted 203 million-strong user base.
To take it a step further, they even acquired the AI Factory just last week, and then brought the new Snapchat “Cameo feature.”
TikTok’s move is a little more terrifying. The company has reportedly dipped their toes in the making of deepfake videos by investing in a software that “will let users swap face in recorded videos and will allow them to easily create their own deepfakes by scanning faces from multiple angles.”
Snapchat and TikTok are watering down the issue by advertising the technology for the use of fun. But until when is it going to be all fun and games?