The world is moving at an extremely fast pace. There are various new technologies coming up and every technology has positives and negatives to it. Understanding the changing technologies can be overwhelming for a normal person who is not so acquainted with them. Another such emerging technology is Deepfake technology. It is a useful technology but it has an immense potential of harming the normal man. Thus, it is necessary to understand this technology and look into its legal implications.
What is Deepfake technology and where did it emerge?
In 2017, a Reddit user by the ‘Deepfake’ uploaded pornographic content of celebrities like Taylor Swift. This thread spread and became popular. The technology used to edit these fake videos was then given the name of Deepfake technology.
Deepfake technology takes its name from “deep learning” and “fake” (a portmanteau), a form of AI. In Deepfake AI, deep learning algorithms that teach themselves how to solve problems with large data sets, are used to swap faces in videos, images, and other digital content to make the fake appear real. It can also be used to create sock puppets, non-existent people, who are active both online and in traditional media. It is mostly used for entertainment purposes. One such example is when Deepfake was used in Star Wars films to depict characters as they were as children or to replace characters who had passed away. But there are dark sides to the use of technology too. The problem arises when it is used for nefarious purposes like ruining a person’s reputation, pornography, or bullying.
How are deepfakes created?
Deepfakes are created using a general adversarial network called GAN. This algorithm uses two algorithms within it. The first one called the Generator creates fake images which are then detected by the second algorithm called the Distributor. The data generated by the Distributor is then again fed into the Generator. Thus, this technology keeps on improving itself on its own making it harder to detect Deepfakes over time.
Dangers of Deepfake Technology
Deepfake technology has developed up to a point where it can be extremely difficult to tell fake content from real ones. The most commonly affected people are celebrities, politicians, and women. Women are targeted heavily and their faces are plastered over sexual content. Deepfake porn can ruin a woman’s life causing immense emotional damage and can ruin her life completely. Such incidents can be traumatizing for the victim and render them completely helpless. Even if such content is removed from the internet. The victims still live in a fear of it resurfacing and being re-traumatized.
On the other hand, it can also lead to economic losses to persons like business owners. It can also have an effect on society in general when famous personalities like influencers and politicians are targeted. When politicians are misrepresented, it can affect their election campaigns and thus the elections as a whole as they influence public opinion.
Another negative use of Deepfakes is for blackmailing people. Threats to post nudes and other sexual content ends up forcing the person being coerced to give in to the person threatening them. It also ends up undermining the credibility of video evidence in legal cases for proving crimes as now fake videos can be created easily.
Shallowfakes and Deepfakes: The Difference
Shallowfakes, also called Dumbfakes, are just like Deepfakes but are made using much more basic editing tools rather than the ones which are highly professional. The video gets glitchy when slowed or sped up which is an easy way to identify Shallowfakes. They can be called a less complex version of deep fakes. An example is the clip from Nancy Pelosi’s address at the Center for American Progress which was edited to slow down the video and change the audio’s pitch to give the impression that she was intoxicated and was shared widely on social media. According to critics, this was a Shallowfake rather than a Deepfake. Even though heavy technology is not used for making Shallowfakes, they can cause equal damage as Deepfakes.
Legal Implications and solutions
Many tools exist to detect deepfakes but they might not work effectively all the time and are not available for use widely. Social media platforms like Twitter and Facebook are taking active steps to curb this menace and are putting forth solutions and implementing them up to an extent on their respective platforms. For example, malicious content is labelled as false on Twitter and a warning is given to users before liking, sharing, or retweeting them. In some cases, the content is directly removed. However, much more needs to be done by social media platforms to ensure that Deepfakes are spotted and removed on time.
Another solution is to generate more awareness among people about how to identify Deepfakes. There are some very simple signs for doing so. Some of them are blurriness, glitching, awkward body movement, discolouration, and the absence of other normal human activities such as no blinking, etc.
Governments around the world have started introducing laws to regulate Deepfake Technology. In the US, a variety of legislation has been suggested at the State and Federal levels, while in the EU, the AI act encourages the labelling of synthetic media for consumer safety. The laws in the People’s Republic of China that prohibit Deepfakes and other forms of “fake news,” as well as recently proposed legislation in the Philippines, are two instances of “fake news” regulations in the Asia-Pacific area. One thing to be wary of with these laws is when they define audio-visual forgery extremely broadly, taking into account significant types of free speech like satire, or when they allow governments a lot of discretion and power to determine what is “fake.”
The Indian Government should also take note of the laws which are being introduced and of the growing use of Deepfakes. Additionally, it is hard to control the uploading of Deepfakes online as there are no proper laws in place to regulate these issues. There is a need to introduce such laws. Therefore, it is high time that we work on regulating these technologies and introduce proper laws in India before more damage is caused.
Author(s) Name: Jui Purwat (Symbiosis Law School, Pune)