Mint Explainer: How celebrities are confronting deep fakes, copyright concerns

Actor Rashmika Mandanna’s recent experience with AI-generated deepfakes that went viral on social media has only added to the concern around such content.

The rapid proliferation of AI-generated texts, voiceovers, deepfakes, and memes through accessible AI programs like ChatGPT, Midjounreny, Stable Diffusion, and AI voice startups like ElevenLabs, presents an increasing challenge for public figures, including actors, authors, and other personalities. As AI technology advances, the difficulties faced by these celebrities in protecting their likeness and content have intensified.

Mint sheds light on the challenges faced by these celebrities, assessing the effectiveness of existing Indian copyright and Intellectual Property Rights (IPR) laws in upholding and safeguarding their rights in the face of these technological advancements.

Risks Posed by Generative AI to Celebrities

The recent case involving Mandanna is a perfect illustration of how vulnerable celebs are to AI-generated content. In this instance, a morphed video shows a woman, resembling Mandanna, entering a venue wearing revealing clothes. Without a discerning eye, it is challenging to distinguish that it’s not the actor but someone else. This level of AI-generated deepfakes and morphed pictures has caused panic among celebrities. Before fact-checking, such content swiftly spreads, reaching millions of viewers across social media platforms.

Mandanna is not alone. AI-generated voiceovers of Prime Minister Narendra Modi have flooded platforms like Instagram. A digitally generated voice mimics Modi singing famous songs in multiple Indian languages. This content is particularly popular among creators using these memes for political campaigns.

This trend concerns celebrities globally with many Hollywood workers going on strike, demanding better pay and protection against AI exploitation. Reports indicate a London-based AI Startup, Realeyes, offering $150 per hour to participate in an AI research project. This project involves creating virtual avatars of Hollywood actors by capturing their emotions and expressions. This initiative solidifies concerns that actors could be easily replaced in big productions using generative AI tools, leading to commercial exploitation without consent leading to more financial losses, dilution of their personal brand value and social media trolling.

Legal Intervention

In September, Anil Kapoor obtained a stay from the Delhi high court on the unauthorised usage of his signature ‘Jhakaas’ movement by AI platforms in memes. Last year, actor Amitabh Bachchan also obtained a similar restraining order from a court for unauthorised usage of his voice and his pictures for a machine-generated message based on the iconic show Kaun Banega Crorepati. However, as of the writing of this article, there is no report about Mandanna taking any legal action against her privacy infringement. 

The Ministry of Electronics and Information Technology said these deepfakes can be dealt with within the Information Technology Act, 2000, which includes Section 66D and the IT Intermediary Rules. Rule 3(1)(b)(vii) mandates social media intermediaries to inform users not to host content that impersonates another person within their regulations. Rule 3(2)(b) requires intermediaries to promptly remove or disable access to impersonation content, including electronically manipulated images, within 24 hours of receiving a related complaint.

Union minister of state for electronics and IT Rajeev Chandrasekhar on  Tuesday urged people who have been impacted by deepfakes to file police complaints and seek remedial measures under the IT Act, which provides for jail time and financial penalties against miscreants. He added that platforms have been asked to take actions against the spread of deepfakes.

Generative AI and Current Copyright and IPR Regime in India

According to legal experts, a key challenge with the current law lies in the assumption within India’s Copyright Act that the author must be human. “Even where the work is computer-generated, the person who caused such work to be generated virtually is considered the author, not the program itself. It’s clear that the definition excludes AI systems as authors of any work,” experts point out. 

“This stance has been taken by Indian courts in various judgments. However, the situation is not that simple. With the onset of ChatGPT and other such AIs, the lines between originality and borrowed work are becoming blurred,” said Ekta Rai, advocate, Delhi high court.

Aviral Kapoor, partner at Alagh & Kapoor Law Offices, suggested that in dealing with the issues posed by AI and digital content creation, it could be wise to reconsider the definitions of originality and authorship. This would involve defining rights concerning content produced by AI and guaranteeing that legal systems adapt alongside technological progress.

Swati Sharma, partner at Cyril Amarchand Mangaldas said, “That is currently a developing law. There are no parameters defined as to how AI will function. All governments globally are working on a framework to define parameters for the functionality of AI, and therefore, for us to say that the Indian law is ready or globally, are any of the laws of any country ready to deal with the challenges of AI? I would say no because AI, in itself, is so transformative at the moment. It will take a long time for us to come to a conclusion about the framework or parameters within which they will function, and then accordingly, laws will have to be amended.”