Someone You Know, Saying Something They Never Said: Inside India's Deepfake Problem

This Blog is Written by Hanshika Singh, 3rd Year, BA LLB, Christ Academy Institute of Law. 

Blog 23 | Edition VII

Introduction

Now imagine that you have obtained a video clip of somebody you know. His or her voice is unchanged and their face is clearly visible, and apparently, he or she says something scandalous. You believe the footage since you do not notice anything that would indicate its falsity. Eventually, you learn that the person has not done or said anything of the kind. It is the harsh reality of deepfakes that is becoming increasingly common. A photograph, video, or audio clip that has been fabricated with the help of AI to make an individual appear to say or do something they never did is known as a deepfake. Technology has many positive aspects such as creative, educational, and entertaining applications, but when abused, it can endanger the individual's identity, reputation, privacy, dignity, and even economic prosperity. Furthermore, it enables the creation of non-consensual pornographic content, fraud, disinformation, and blackmail. What is the most troublesome aspect of the situation is the fast pace at which technology works. It takes minutes to create and distribute a false video, but it may take much time to identify the perpetrator and obtain any kind of justice. Therefore, the gap between law and technology emerges. In case of harm, apart from punishment, there needs to be prevention, awareness, accountability, and treatment for India.

India’s Legal Framework

There is no law in India that has a single definition for the "deepfake misuse". Rather, various laws could be in play depending on the type of technology being used. There are laws in the Information Technology Act, 2000, that address some crimes associated with deepfakes. These sections—sections 66C, 66D and 66E—cover privacy breaches, identity theft and personation by means of computer resources. For deepfakes that include pornographic or sexually explicit electronic content, Sections 67 and 67A may also apply. If a deepfake is used to commit an existing offence such as cheating, impersonation, intimidation or defamation, the Bharatiya Nyaya Sanhita, 2023 (BNS) could be relevant. However, there is no general offence in the BNS called ‘deepfake misuse'. Personal data relating to identifiable persons may be relevant in this context and the Digital Personal Data Protection Act, 2023 (DPDP Act) may also be applicable. However, it cannot address all issues with the deepfake. Where a deepfake infringes on multiple rights at the same time, such as right to privacy, defamation, fraud, sexual exploitation, and personality, multiple legal remedies may be required. The 2026 change to the Information Technology Rules pertaining to artificially generated information is another significant development. Not all AI-generated content is banned in the Rules. Synthetic content, when permitted, should be identified and, if possible, include permanent content metadata or other means of identifying the content source. The responsibilities for removing may apply to illegal synthetic material.

Indian Courts and the Protection of Identity

A significant case for understanding the legal protection of identity is Anil Kapoor v. Simply Life India & Ors., 2023 SCC OnLine Del 6914. The Delhi High Court granted temporary relief from unlawful exploitation and recognized the protection of personality rights which cover name, appearance, voice, likeness and persona. The case is important because it demonstrates the extent to which existing personality-rights principles can be applied to novel technologies, even if the case was’nt one involving a deepfake. In the era of AI, the importance becomes more apparent. With technology these days, it's easy to recreate a face or voice and trick people into thinking they were there when they were not. Meanwhile, any expression of real news, criticism, satire, parody, and other forms of legitimate expression protected by the constitution ought to be balanced with identity protection.

The Victims We Often Forget

Without a doubt, when the discussion of deepfakes comes up, it's celebrities, politicians and women that are often in the spotlight. But there are others who are equally deserving of attention: children, transgender, persons with disabilities. Children are especially susceptible to material created using AI which is sexual in nature and may have severe repercussions in terms of child-protection and criminal law. Digital content can also be copied and distributed again and again, causing harm beyond the initial removal of the content. Transgender individuals can be subjected to deepfake images that defy their gender identity, reveal personal data, generate sexualised images or validate discriminatory stereotypes. Manipulated representations of a person's appearance, behaviour, or medical condition can be experienced by persons with disability and reinforce social prejudice that they may already have. So, deepfake regulation should be technology agnostic, but impact on vulnerabilities. The Law must safeguard all while acknowledging digital abuse may have varying impacts in each community.

The Digital Divide: When People Do Not Know What a Deepfake Is?

But there is another issue which often gets overlooked digital literacy. Possession of a smartphone does not necessarily equate to knowledge about the workings of AI. While many people may know how to send and receive a video, they may not be aware that the face and/or voice in the video can be a result of artificial intelligence. Reports from the government have constantly pointed out disparity in internet usage between rural and urban populace. Thus, public awareness becomes necessary. India requires deepfake literacy programs in schools, colleges and localities, especially in regional languages. Three basic things must be taught to people what a deepfake is, how to verify suspicious material, and where to report it. A video containing a known face or voice must not automatically be believed.

What Can India Learn From Other Countries?

Some other countries provide important lessons for India, but there is no necessity to adopt any particular scheme entirely. Firstly, the EU’s AI Act provides an example of the transparency and labelling regime for the AI-generated content. The transparency obligations under Article 50 of the AI Act come into effect on 2 August 2026. Hence, this act should be considered as a transparency regime rather than the full-scale solution to deepfake problems. Secondly, China’s Deep Synthesis Provisions focus on such issues as identification, labelling, traceability and responsibility of providers of deep-synthesis services. Thirdly, the Criminal Code Amendment (Deepfake Sexual Material) Act 2024 of Australia improves criminal protection against some kinds of deepfake non-consensual sexual materials. All these countries provide three valuable lessons for India: transparency from the EU, traceability from China and criminal protection from Australia.

What Should India Do?

There needs to be a framework in India which includes a preventive system, labeling, quick removal, accountability, redress for the victims and awareness. The platforms will have to offer an easily understandable system through which complaints can be made in case of illegal content. In cases of victimization, screenshots, URL, user names and dates when the content was uploaded need to be saved prior to lodging the complaint. The content can be reported to the platform, the National Cyber Crime Reporting Portal, the local police station or the cybercrime authority. The technology of AI can also be used for the same purpose, as it helps detect whether the images, audio and video files have been altered or not. However, AI detection is not fool-proof as well; sometimes it incorrectly detects the original content and in some cases does not detect the advanced forms of manipulation.

Conclusion

Deepfake demonstrates that the rate at which technological innovation can pose challenges to the concept of identity is quick. India has already begun to deal with the issue by way of the Information Technology Act, BNS, DPDP and the changing IT Rules. But the law must keep evolving with technology. The point is not to ban the use of AI or punish every form of synthetic content. The scope should rather be to prevent any misuse, detect synthetic content, take down any illegal material quickly, ensure liability of responsible individuals and platforms, provide recourse to the victim and educate the people. The lack of knowledge should not become an excuse for the lack of protection. Technology is going to progress further. It is thus imperative for the law to act before the next wave of deepfakes makes detection and regulation even more difficult. When technology can create someone’s identity, can the law protect the individual behind it?

(Write to the author at Hanshikasingh220@gmail.com.)