AI Deepfakes Are Outpacing the Law: Can India’s New Rules Close the Enforcement Gap? - Devdiscourse

India’s 2026 amendments shift practical responsibility toward intermediaries that enable, host, publish or distribute synthetic content. The rules do not create a standalone deepfake crime. Instead, they require platforms to help prevent, identify and limit harmful synthetic media while keeping existing legal remedies available.

Platforms must use appropriate technical measures against specified unlawful synthetic content. They must also prominently label synthetic material that is permitted. Where technically feasible, they must preserve provenance information, including identifiers linked to the resources used to generate or modify content. These duties target both circulation and traceability.

The most urgent change is procedural: the takedown deadline falls from 36 hours to three hours. This may reduce exposure to political deception, impersonation and fraud, but the researchers warn that detection systems have technical limits. Enforcement capacity, constitutional protections and the need to protect legitimate expression will shape whether the rules work in practice.

A deepfake is realistic synthetic audio or video that fabricates a person’s appearance, voice or actions. The technology can make identifiable people seem to say things they never said or perform acts that never occurred. Its legal importance depends on context, purpose and likely harm, not simply on the fact that artificial intelligence was used.

Ordinary digital editing can also alter media, but deepfakes are distinguished by convincing synthetic impersonation. Cloned voices and generated video may look or sound authentic enough to influence decisions. The article notes that relatively accessible software has made this capability available beyond specialist technical circles.

That apparent authenticity creates risks for privacy, identity, politics and money. A manipulated speech may distort public debate. An impersonated celebrity may suffer personality-rights harm. A fake investment promotion may deceive viewers financially. Because the consequences vary, deepfake regulation must work alongside criminal, consumer-protection, electoral and personality-rights mechanisms.

India’s 2026 amendment reduces the applicable takedown deadline for unlawful synthetic content from 36 hours to three hours. That compressed window reflects the speed at which fabricated videos, cloned voices and manipulated identities can spread. It also places pressure on platforms to detect harmful material quickly and make fast removal decisions.

The rules cover intermediaries involved in enabling, hosting, publishing or distributing synthetic content. They also require technical safeguards against specified unlawful content, prominent labels for permitted synthetic material and provenance preservation where technically feasible. Together, these mechanisms combine response speed with attempts to show how synthetic material was made.

The supplied article excerpt does not name the large platforms that face extra duties. It does show why enforcement is difficult: detection technology has limits, and platforms must balance rapid action with constitutional protections and legitimate expression. The researchers therefore describe the rules as an early intervention whose practical success remains uncertain.

Deepfakes create different kinds of real-world harm because they can make false material appear authentic. In elections, manipulated speeches and cloned voices can falsely attribute statements to public figures and disrupt democratic debate. In personal life, synthetic impersonation can damage reputation, privacy and personality rights.

The article cites a non-consensual deepfake involving actress Rashmika Mandanna and legal action protecting actor Anil Kapoor’s personality rights. It also discusses fraudulent investment promotions using prominent business figures’ identities. In one documented case, a Bengaluru woman reportedly lost ₹3.75 crore after encountering an AI-generated video featuring spiritual leader Sadhguru.

These examples show why one legal response cannot address every case. Electoral deception, sexual exploitation, identity appropriation and financial fraud involve different interests and remedies. India’s approach therefore combines platform rules with existing criminal, consumer-protection, electoral and personality-rights mechanisms, rather than treating all synthetic media as the same problem.

India’s amendments take a platform-focused route. They regulate intermediaries involved in enabling, hosting, publishing or distributing synthetic content, rather than establishing a standalone criminal statute for deepfakes. This approach treats circulation as a central enforcement point, because platforms can apply technical measures, labels and removal procedures across large volumes of content.

The rules add a three-hour takedown deadline for unlawful synthetic content. They also require safeguards against specified unlawful material and allow permitted synthetic content to remain when it is prominently labelled. Provenance information must be preserved where technically feasible. These duties operate alongside existing criminal, consumer-protection, electoral and personality-rights mechanisms.

The model is described as a reasonable early intervention, but its success is not guaranteed. Platforms may struggle with detection, speed and context. Enforcement institutions may also be stretched. At the same time, constitutional protections mean that efforts to suppress harmful deception must not unnecessarily restrict legitimate expression.

The study places India alongside China, the European Union and the United States in a comparative assessment of deepfake regulation. The supplied excerpt identifies India’s model as distinctive and hybrid. It combines content labelling, technical safeguards, accelerated removal and existing legal remedies rather than relying on one new criminal law.

For India, the concrete measures include a three-hour takedown deadline for unlawful synthetic content, safeguards against specified unlawful material and prominent labels for permitted synthetic content. Platforms must preserve provenance information where technically feasible. The excerpt does not provide equivalent details about China, the EU or the United States.

Therefore, a precise comparison cannot be made from the supplied text without adding facts not stated in it. The researchers’ broader warning applies to the comparative question: enforcement capacity, constitutional protections and detection limits may determine results. India’s approach is an early intervention, but its effectiveness remains uncertain.

Platform liability protections shape who must respond when harmful content appears online. India’s rules place greater responsibility on intermediaries that enable, host, publish or distribute synthetic content. Yet the supplied article does not detail the precise liability protections or safe-harbour rules that may apply. It presents the issue mainly as a balance between enforcement and rights.

Privacy rights matter because deepfakes can expose people to non-consensual sexual imagery, identity appropriation and reputational injury. Personality-rights litigation, such as the action protecting Anil Kapoor’s identity, shows that existing remedies can address some harms. Freedom of expression matters because not every synthetic work is unlawful; permitted content may remain if prominently labelled.

This balance makes enforcement difficult. Platforms need fast detection and removal, but automated systems have technical limits and may misunderstand context. The researchers warn that constitutional protections and legitimate expression could constrain overbroad action. India’s three-hour deadline may improve protection, but enforcement capacity will determine whether speed is effective and fair.