AI deepfakes of Rashmika Mandanna, Alia Bhatt, Amitabh Bachchan and others have made personality rights suddenly urgent. But is the existing law enough to protect us from AI? Honestly — only partly. Here’s where it stands, and where it must go.
The threat
As technology races ahead, so does its misuse. Convincing deepfake videos — AI-manipulated to look genuine — have targeted many celebrities: Rashmika Mandanna, Alia Bhatt, Katrina Kaif in morphed lewd videos; and during the Lok Sabha elections, fake videos of Ranveer Singh and Aamir Khan on political matters. Deepfake tech, available to anyone with a device and internet, is misused to show stars endorsing illegal gaming apps, advertising products they never signed up for, or in fabricated lewd videos. Even OpenAI faced an allegation from Scarlett Johansson over a chatbot voice resembling hers despite her refusal.
In response, many celebrities have started trademarking their names to protect their goodwill.
The limits of existing IP law
Creating and circulating deepfakes violates the Right to Privacy, and contravenes Sections 66 and 67 of the IT Act, 2000 plus MeitY guidelines. But do IP laws offer recourse? Only limited scope:
- Copyright — helps only if the morphed video edits an original the personality had copyright in;
- Trademark — helps if the celebrity trademarked their name, or a registered mark appears in the video, or under passing off.
The existing legislation isn’t fully equipped for AI’s complexities.
Courts and personality rights
Courts have stepped in via personality rights — protecting unauthorised use of names, images, voice, likeness, dialogues or traits. Various High Courts have recognised the rights of Jackie Shroff, Amitabh Bachchan, Anil Kapoor and others — ordering takedowns of infringing media and domains, advisories to intermediaries, and compensatory damages.
But is it enough?
The effectiveness is questionable. These remedies give temporary relief against the primary source, but can’t undo the confusion or dilution already caused. Once a video goes viral, removing it from the global web is nearly impossible — there’s no reversing an upload. So new or updated provisions across IP, technology, media and telecom law are an indispensable necessity.
The author’s view: trademark law’s role
Trademark law — properly enforced — can play a crucial role. Its very objective is to distinguish source and prevent misuse causing confusion or dilution — exactly what personality rights aim at for a person’s goodwill. Just as passing off protects unregistered owners, celebrities should be protected even without registration. The law must become more stringent, with stricter punishments for unauthorised use of identity and likeness, and strict regulation of posting such media on third-party platforms. Regulating AI has become the need of the hour — a necessary evil whose effective regulation is the only way forward.
The takeaways
- AI deepfakes outpace existing IP law — copyright and trademark help only partly.
- Courts protect via personality rights — takedowns, advisories, damages (Jackie Shroff, Anil Kapoor, etc.).
- Remedies are temporary — viral content can’t be fully undone.
- The law must evolve — stronger provisions, stricter penalties, and AI/platform regulation.
Frequently asked questions
Do celebrities have legal protection against AI deepfakes in India? Partly — through personality rights (recognised by courts), the IT Act (Sections 66/67), privacy law, and limited IP remedies. Courts have ordered takedowns and damages, but the framework isn’t fully equipped for AI.
Can copyright or trademark stop a deepfake? Only in limited cases — copyright if the deepfake edits a work the person owns, trademark if their name is registered or a registered mark appears, or via passing off.
What have courts done about deepfakes? Recognised personality/publicity rights for celebrities like Jackie Shroff, Anil Kapoor and Amitabh Bachchan, ordering takedowns of infringing media and domains, advisories to intermediaries, and damages.
Why aren’t current remedies enough? They offer only temporary relief; once a deepfake goes viral, it can’t be fully removed from the web, so stronger, updated laws and AI regulation are needed.
Legislation referred to
- The Information Technology Act, 2000
- The Trade Marks Act, 1999