DEEPFAKES AND THE DANGERS THEY POSE
We have talked about deepfakes on here before. Convincing deepfakes are generated by an AI that is trained using a neural model called a GAN. The more data you have, the more convincing your deepfake will be. As the amount of data we have access to grows exponentially, it will only get easier to create these fake photos/videos and potentially smear the reputation of someone in power. Very recently a deepfaked video of House Speaker Nancy Pelosi that aimed to portray her as being intoxicated surfaced on the internet. It was posted on Facebook and they refused to remove it even after it was proven the video was fake.
Pelosi later told California’s KQED “I think they have proven, by not taking down something they know is false, that they were willing enablers of the Russian interference in our election.” Governments and various activists have called for social networks to assist in the deletion of deepfakes. They have failed to respond in an appropriate manner. Now social network executives are being targeted to make them take this more seriously.
MARK ZUCKERBERG DEEPFAKED
In an interesting deepfake video as shown below the Facebook CEO is portrayed to say: “Imagine this for a second: One man, with total control of billions of people’s stolen data, all their secrets, their lives, their futures.”
Now this deepfake has been called non-malicious in nature and it is obvious that most people will know it is not real. The purpose of the video which was created by an Israeli startup company called Canny AI was simply to raise awareness and push social networks like Facebook to do something about this. Canny AI has also came out with deepfakes of President Trump and Kim Kardashian as part of this push for public awareness.
These deepfakes will only become more convincing with time.
CONGRESS HOLDS A HEARING ON DEEPFAKES
On June 13, 2019 the U.S. House of Representatives Select Committe on Intelligence held a hearing on deepfake videos. Researchers testified to Congress about the potential threat manipulated content poses to national security, particularly in light of the upcoming 2020 presidential election. Deepfakes have gained a lot of traction lately and it is about time Congress stopped ignoring issues related to technological advancement. More on this hearing can be found here.
At some point these deepfakes will become more and more lifelike. The line between what is real and what is fake will begin to blur. Eventually every single social media network and every single news source that displays photos/videos will need their own army of AIs that are programmed to detect these deepfakes with a high level of accuracy. They could then scan everything prior to displaying such things. It will also get to the point where individuals will have access to open sourced AIs of this caliber so that they can screen things for themselves. The programmers who create deepfakes will end up in a race with the programmers who create AIs to detect said deepfakes to see which kind of AI tech will progress the fastest.