The rapid rise of generative artificial intelligence (AI) has created a major challenge for India's digital liability laws. Instead of simply hosting user data, generative AI systems actively recreate human identity. This shift disrupts the core assumptions of Section 79 of the The Information Technology Act, 2000, which protects online platforms from being held liable for user-generated content.^[1] While the Information Technology (Intermediary Guidelines and Digital Media Ethics Code) Amendment Rules of 2026^[2] create fast-tracked takedown times for "Synthetically Generated Information" (SGI), they still rely on an outdated model where platforms only react after someone uploads infringing content.

Moving beyond temporary court orders, three structural reforms are captured here: passing a law to establish a "Right to Identity Continuity," setting up automatic, smart-contract-backed royalties to cover AI training data, and mandating digital tracking markers within the AI software itself to protect individual digital sovereignty.

I. The End of Platform Neutrality: How AI Creates Identity

The intersection of generative AI and personality rights has exposed a deep crack in Indian digital governance and intellectual property law. As courts tackle unauthorized AI-generated voices, faces, and traits, we must ask a fundamental question: can a system that actively recreates a real person's identity claim immunity as a neutral platform? For twenty-five years, internet regulation has relied on the "Safe Harbor" rule under Section 79 of the Information Technology Act, 2000. This rule assumes platforms are neutral pipelines digital highways not responsible for what people post on them.

However, generative AI breaks this logic entirely. When platforms use deep neural networks to mimic a person's voice, behavior, or appearance, they are no longer neutral. They actively ingest, process, and rebuild human traits to create synthetic personas. This means they act like traditional publishers who choose what to publish, rather than neutral distributors. As a result, we must urgently overhaul our digital liability rules.

II. AI Authorship: Why the 2026 Rules Fall Short

Our regulatory gap becomes clear when looking at the Information Technology (Intermediary Guidelines and Digital Media Ethics Code) Amendment Rules, 2026. Under these rules, the government introduced a framework for "Synthetically Generated Information" (SGI). The rules define SGI as audio or video content created or changed using computers to look or sound real.

The 2026 rules require platforms to take down deepfakes or impersonation content within two to three hours of receiving a complaint.^[3] To balance this heavy workload, the rules include a "Good Samaritan" provision. If a platform uses automated tools to label or block synthetic content, it will not lose its safe harbor immunity under Section 79.

Yet, this setup does not work for generative AI. The rules assume that an independent user uploads an illegal file and the platform merely distributes it. But generative AI works the other way around. When a user enters a prompt, the AI generates a voice or face from data it was trained on. The platform's own system creates the content. The tool is not just sharing an infringement; it is exploiting a person's identity for commercial gain. By shifting from host to creator, generative AI platforms act as publishers, using safe harbor rules as a shield.

III. How AI Dilutes Personal Identity and Dignity

In India, courts protect "personality rights" the right of a person to control how their identity is used. Historically, this has two roots:

1. It is part of the right to privacy and personal liberty under Article 21 of the Constitution, as established in the landmark case K.S. Puttaswamy v. Union of India (2017)[4] (2017 INSC 616) and R. Rajagopal v. State of Tamil Nadu (1994) (1994 INSC 521).

2. It relies on the civil law of "passing off," which stops businesses from falsely claiming a celebrity endorses their product. Previously, you had to prove actual commercial confusion like a fake billboard or product line.

Generative AI, however, introduces "Identity Diffusion." When online tools allow anyone to create endless clones of a person, having unique conversations, reading text, or generating video in their voice, the harm goes far beyond financial loss. It erodes a person's control over their own self.

If platforms can freely turn a person's voice and expression into training data under the excuse of "transformative use," human identity becomes a commodity for platforms to exploit. This violates the core promise of Article 21, which protects human dignity. Indian courts have begun to recognize this threat in major cases, such as the Delhi High Court's protective order in Anil Kapoor v. Simply Life India & Ors. (2023)^[5] and the Bombay High Court's ruling in Arijit Singh v. Codible Ventures LLP & Ors. (2024).^[6] In both cases, the courts stopped platforms from using AI to copy the actors' and singers' voices, expressions, and traits without consent.

IV. Threats to the Public Sphere and Digital Trust

Beyond individual harm, copying human identities without accountability damages public trust. Flooding social networks with synthetic audio and video that look entirely real makes it hard for citizens to trust real journalism.

Traditionally, identity theft is seen as a private dispute. But when algorithms automate identity replication on a massive scale, it becomes a systemic threat. It can be used to fake public consensus, manipulate markets with deepfakes, or destabilize societies.

A prime example is how courts have had to step in to protect public figures from wider structural exploitation, as seen in the Bombay High Court's order in Karan Johar v. India Pride Advisory Private Ltd. (2024),^[7] where a filmmaker's name and identity were unauthorizedly weaponized. Because safe harbor rules protect platforms, developers face zero penalties for these wider societal harms. This allows wealthy platforms to profit from deception while genuine human participation gets pushed aside.

V. A Practical Blueprint for Structural Reform

To fix this gap, Indian law must move away from reacting to individual lawsuits with temporary court orders. We need three structural reforms:

First, we must pass a law establishing a "Right to Identity Continuity." This law would declare an individual's biometric features, voice, and unique gestures to be their personal, inalienable property. Any system that wants to model or clone these features would need a time-limited "Identity License." The burden of proof would shift to the platform to show they have a secure digital consent token before generating content that looks or sounds like a real person.

Second, we should create a "Compensated Access Framework." If an AI model trains on a person's public recordings, books, or art, it must track this usage on a secure digital ledger. Every time a user interacts with that synthetic voice or persona, a micro-royalty should be paid automatically to the original creator. This turns unfair scraping into a fair, licensed market.

Third, the law should mandate "Technical Provenance Protocols." Any system that generates human likenesses or voices must embed an unalterable digital watermark into the file's metadata. If an AI tool generates a voice that is highly similar to a real person's voice without a consent token, the system should block the file automatically before it can be shared online. This stops identity theft at the point of creation, rather than waiting for slow and expensive court battles. It ensures platforms can no longer hide behind outdated safe harbor protections while profiting from synthetic identities.

1. The Information Technology Act, 2000, § 79, No. 21, Acts of Parliament, 2000 (India). ↑

2. Ministry of Electronics and Information Technology (MeitY), Government of India, Gazette Notification G.S.R. 120(E), Information Technology (Intermediary Guidelines and Digital Media Ethics Code) Amendment Rules, 2026 (notified Feb. 10, 2026; effective Feb. 20, 2026). Access the official notification PDF via LiveLaw here. ↑

3. See LiveLaw's detailed analysis here . ↑

4. K.S. Puttaswamy v. Union of India (2017) 10 SCC 1 (2017 INSC 616). ↑

5. Anil Kapoor v. Simply Life India, 2023 SCC OnLine Del 6905. ↑

6. Arijit Singh v. Codible Ventures LLP, 2024 SCC OnLine Bom 2445, decided on 26-07-2024. ↑

7. Karan Johar v. Indian Pride Advisory Pvt. Ltd. & Ors., Commercial IP Suit (L) No. 17863 of 2024 with Interim Application (L) No. 17865 of 2024. ↑

   Views are personal.