About Deepfake Technology

The myriad of deepfake technology, from its nascent presence since early 1990's to its ignominious specters today, is not only readily available but has also transgressed into copyright infringement. It is often weaponized by the masked criminals sustaining themselves over online interfaces to commit digital frauds. The term "deepfake" comes from the mere combination of "deep machine learning" and "fake". Deep faking is an advanced technology where the artificial intelligence is used to create deceptive realist images, audio as well as visual, which is often known as "synthetic media". It exists artificially yet the digital impersonation creates an impression of events that never actually took place, altering the reality.

Typically, Generative Adversarial Networks, in which two algorithms are pitted against, to create such synthetic media, are the most popular method for producing deepfake. This technology can be utilized in multifarious ways, educational sphere in order to produce interactive and creative content, in the film industry to substitute the lead actor to align a stunt double and even by social media platforms to enhance the user experience. However the technology holds vast potential to be utilized for legitimate purposes, its deployment for pernicious activities is on leap, including revenge porn, disparagement of political foes, fraudulent businesses, and harassing women.

In fact, since the availability of this technology has increased, 90–95% of deepfake movies are now non-consensual pornographic videos, and 90% of those videos target women, the majority of whom are minors Additionally, the use of deepfakes to disseminate false information about political parties and politicians has increased. A now-viral video from 2018 featured a deepfake President Obama saying things that weren't of his own, but were instead meant to emphasise how lifelike these artificial voices and graphics were. With its quick development, deepfake technology has the capacity to spread false information, harm lives, and destroy businesses if left unchecked.

The Tussle Between Copyright Infringement And Generative AI Deepfakes

Since, there exist easily accessible and widely available techniques that make them difficult to detect, deepfakes may quickly produce morphing photos and videos, which further drives their improvements. India lacks specific legislation to deal with deepfakes and AI-related issues, making it possible for someone's video/image or any copyrightable creation to be used without permission with startling accuracy. Actors' and authors' struggles to keep control over their digital identities are representative of a larger social issue.

Deepfakes cast doubt on the definition of originality and creativity. The question to still ponder upon is whether a deepfake still violate the copyright of the original work even if it modifies the original to the extent that it becomes a new and original creation? Since the law is ambiguous on these issues, legal development is necessary to handle this new problem. Its likely to see how courts handle these issues in light of present copyright law as long as deepfakes keep popping up.

Because deepfakes are often made by altering already published, copyrighted content, they are thus closely related to copyright law. Deepfakes thus raise issues of copyright infringement, notwithstanding the possibility of a strong fair use defence against infringement in certain circumstances. The Generative AI creates such an impersonation which may bring into existence such creations which may violate exclusive rights of the owners, especially the Right to publicly display their respective work.

The deepfake conundrum and the Indian stance

In the Indian context, Section 52 of the Indian Copyright Act, 1957 covers the idea of "fair dealing," albeit it doesn't provide a definition. This section effectively makes a distinction between legitimate and illegitimate users of protected works by providing a comprehensive list of works that are deemed to be infringing. However, the sections do not provide an exhaustive list of situations dealing with transformative use, leaving room for ambiguities and legal uncertainties for the deepfake works. From a wide perspective, it can be said that, the existing state of Indian copyright law can be used as an effective tool to combat deepfakes because of its rigour. Nonetheless, the Indian doctrine of fair dealing states that all deepfakes, including those made for amusement and with a valid purpose, violate copyright. This stance must be adjusted to take into account deepfakes produced for legitimate reasons.

Article 13 of the Trade Related Aspects of Intellectual Property Rights (TRIPS) states that "Members shall confine limitations or exceptions to exclusive rights to certain special cases which do not conflict with a normal exploitation of the work and do not unreasonably prejudice the legitimate interests of the right holder." The primary implication that can be drawn from this section in Indian context is, that certain alignment in legal mechanism and enforcement of copyright law, in a way that it not only stifles originality in creativity but also protects the right of creators from potential abuse.

Further, Copyright and intermediary liability are closely related concepts. Deepfakes and other similar content can typically reach a wider audience by way of social media and other intermediaries. Intermediaries may be required to remove illegal content under Section 79 of the Information Technology Act of 2000 upon receiving a court order or actual knowledge of it.[1] It is a safe harbour provision for the intermediaries ideally, but if adequate and prompt action to take down such deepfakes or prolonged availability of deepfakes even after sufficient knowledge shall not provide protection to such intermediaries anymore.

Conclusion
Due to their complexity and wide range of applications, deepfakes need that lawmakers and the courts take into account the motivation behind their development when determining whether or not to grant copyright protection to them. The Indian copyright legislation's comprehensive list of fair dealing provisions ignores the potential applications of deepfakes in the fields of entertainment, education, and medical technology. Because of this, even while it looks to be a step ahead in combating criminal usage of deepfake technology, it doesn't seem to be able to recognise the genuine applications of deepfakes.

The laws that control technology use must also advance with it. The work that lies ahead is difficult but necessary. We can perhaps create a system that safeguards individual rights without impeding technological advancement if we acknowledge the complexity of the issue and pursue all-encompassing legal, technological, and ethical solutions.

Written By:

1. Muskaan Suneja, a fourth-year law student at Christ University, Delhi
2. Amartya Shivam, a fourth-year law student at Christ University, Delhi