Deepfake is a combination of two words, which are ‘Deep learning’ & ‘Fake’. By using Deepfake technology one can replace an existing person’s face in an image or video.
Deepfake is leveraging powerful techniques from machine learning and artificial intelligence, the main machine learning methods used to create Deepfakes are based on deep learning and involve training generative neural network architectures, such as autoencoders or generative adversarial networks (GANs).
Researchers at academic institutions have developed this technology in the early 1990s and later by amateurs in online communities. Technology steadily improved during the 20th century, and more quickly with digital video.
How does this Technique works
Deepfake uses an autoencoder which reduces an image to a lower-dimensional latent space. Deepfakes utilize this architecture by having a universal encoder that encodes a person into the latent space. The latent representation contains key features about their facial features and body posture.
This can then be decoded with a model trained specifically for the target. This means the target’s detailed information will be superimposed on the underlying facial and body features of the original video, represented in the latent space.
What is the main concern of this powerful technology?
Deepfakes can be a dangerous form of crime through artificial intelligence. Blackmailing to Financial Scams. The presence of deepfakes increases the difficulty of classifying videos as genuine or not.
Suppose you all are seeing a video of your President or Prime Minister, where he/she is saying some unethical stuff or requesting you to break the law, now before it can be proved that this is a Deepfake video some of you may go against him/her. So this can be a big threat to national security.
A deepfake of Donald Trump was easily created based on a skit Jimmy Fallon performed on NBC’s The Tonight Show. This was on 4 May 2016 where, Jimmy Fallon dressed up as Donald Trump and pretended to participate in a phone call with Barack Obama, conversing in a manner that presented him to be bragging about his primary win in Indiana.
American actor Jordan Peele, BuzzFeed, and Monkeypaw Productions created and produced a deepfake of Barack Obama (uploaded to YouTube on 17 Apr 2018) that depicted Barack Obama cursing and calling Donald Trump names.
Now for the above two examples, it was clear in the video itself that videos weren’t original, the intent was to show how dangerous this tech can be.
Let’s see some of the incidents that occur using Deepfake-
Many deepfakes on the internet feature pornography of people, often female celebrities whose likeness is typically used without their consent. Deepfake pornography prominently surfaced on the Internet in 2017, particularly on Reddit. A report published in October 2019 by Dutch cybersecurity startup Deeptrace estimated that 96% of all deepfakes online were pornographic. The first one that captured attention was the Daisy Ridley deepfake, which was featured in several articles.
Even we have seen this tech in Politicians’ videos. During the 2020 Delhi Legislative Assembly election campaign, the Delhi Bharatiya Janata Party used similar technology to distribute a version of an English-language campaign advertisement by its leader, Manoj Tiwari, translated into Haryanvi to target Haryana voters. In that video, Tiwari has seen criticizing Kejriwal & asking people to vote for BJP. Later revealed that it was a Deepfake.
Read more: Basics of International Space Station.
Now it doesn’t mean that Deepfakes don’t have any positive implication-
Powerful technology always has good agendas, for example –
Corporate training videos can be created using deepfake avatars and their voices.
Microsoft has developed an app called “Seeing AI” which uses artificial intelligence and deepfake technology to narrate the world around you to aid individuals who are blind and/or have low vision.
What are the steps taken till now-
Twitter is taking active measures to handle synthetic and manipulated media on its platform. In order to prevent misinformation from spreading, Twitter is placing a notice on tweets that contain manipulated media and/or deepfakes that signal to viewers that the media is manipulated.
Facebook has taken efforts towards encouraging the creation of deepfakes in order to develop state-of-the-art deepfake detection software. Facebook has also detailed that the platform will be taking down media generated with artificial intelligence used to alter an individual’s speech.
However as we know this is a powerful tech, not all deepfakes can be detected accurately and immediately, sometimes it can take a considerable time to be found and taken down.
‘Our advice will be don’t believe everything you see online.’