AI-Generated Deepfakes & Obscenity Laws

1. The Challenge of Synthetic Media

Before we dive into the legal text, it's essential to grasp the nature and scope of the problem. AI-generated media, or "deepfakes," are not a single category of content; they can be used for parody, art, political messaging, and malicious harm. Understanding this spectrum is key to appreciating the legal and ethical complexities.

To see this in action, please watch the following video from BBC News. It provides vivid, real-world examples of how deepfake technology is being used and misused, from political campaigns in India to non-consensual explicit images of public figures.

The video demonstrates that the term "harmful content" is broad. It can refer to:

• Defamation and Reputational Harm: Falsely depicting someone saying or doing something damaging.
• Privacy Invasion and Harassment: Creating non-consensual intimate or explicit imagery.
• Public Mischief and Disinformation: Spreading false information to manipulate elections or incite unrest.

With this context in mind, let's analyze the primary legal tool India currently uses to combat such content: the Information Technology Act, 2000.

2. Applying Pre-AI Laws to Post-AI Content

The IT Act was enacted long before generative AI became mainstream. Consequently, regulators and courts must interpret its existing provisions to address new forms of harm.

Let's start by examining the core paper that outlines the legal landscape. This will serve as our guide to the relevant statutes.

As the paper outlines, there isn't one "deepfake law" but a collection of provisions that can be applied depending on the nature of the harmful content. Let's break them down.

Provisions for Obscene and Sexually Explicit Content

The most direct application of the IT Act to harmful deepfakes concerns obscene and explicit material. For this, we turn to a specialized analysis which breaks down the relevant sections.

Based on this reading, the key provisions are:

• Section 67: Punishment for publishing or transmitting obscene material in electronic form. This is the general provision against obscenity, carrying a penalty of up to three years imprisonment and a fine.
• Section 67A: Punishment for publishing or transmitting of material containing sexually explicit act, etc., in electronic form. This is a more stringent provision targeting sexually explicit content, with penalties rising to five years imprisonment and a higher fine.
• Section 67B: (Not detailed in the article but important to know) This section specifically deals with material depicting children in sexually explicit acts, carrying the most severe penalties.

These sections are the primary legal weapons against the creation and distribution of non-consensual deepfake pornography.

Provisions for Other Harms

As we saw in the BBC video, not all harmful deepfakes are obscene. For other harms like impersonation and privacy violation, other sections of the IT Act are used:

• Section 66C (Identity Theft): Punishes the fraudulent use of another person's electronic signature, password, or other unique identification feature. This could apply if a deepfake voice is used to bypass biometric security.
• Section 66D (Cheating by Impersonation): Punishes cheating by impersonating someone using a computer resource. This is highly relevant for deepfakes used in financial fraud or to spread disinformation.
• Section 66E (Violation of Privacy): Punishes anyone who "intentionally or knowingly captures, publishes or transmits the image of a private area of any person without his or her consent."

The Analytical Gap: The Challenge of "Synthetic" Reality

Your background in legal research will help you appreciate the nuance here. While these provisions seem applicable, they contain language that a defense counsel could challenge in the context of purely synthetic media.

Let's return to our core academic paper to explore a critical analytical gap.

The paper raises a crucial point: Section 66E penalizes the act of "capturing" or "transmitting" an image of a person's private area. A defense could argue that a deepfake is not a "captured" image of a real event but an entirely new, synthetically generated artifact. It doesn't depict something that happened; it creates a depiction of something that never happened. This textual ambiguity highlights the need for laws that are technologically neutral or specifically designed for AI-generated content.

3. The Evolving Regulatory Response

Recognizing these gaps, the Indian government has been actively working to strengthen the regulatory framework beyond the baseline provisions of the IT Act. This has been a multi-step process.

First, the Ministry of Electronics and Information Technology (MeitY) issued advisories, putting pressure on social media intermediaries to proactively tackle deepfakes.

Following these advisories, the next logical step is to amend the rules to make these obligations explicit. The article from Khurana & Khurana, which you read earlier, discusses amendments to the IT Rules that create a more direct regulatory framework for AI-generated content. These amendments focus on:

1. Formal Definition: Creating a statutory definition of “synthetically generated information” to remove ambiguity.
2. Mandatory Labeling: Requiring platforms and tool providers to clearly and permanently label AI-generated content, empowering users to distinguish between authentic and synthetic media.
3. Platform Accountability: Reinforcing the obligations of intermediaries to remove harmful deepfakes upon notification to retain their "safe harbour" protection under Section 79 of the IT Act.

This evolution—from general criminal provisions to intermediary advisories to specific labeling rules—demonstrates a maturing regulatory approach that tries to keep pace with technology.

Conclusion

In this lesson, we have analyzed how India's legal framework, primarily the IT Act of 2000, applies to obscene and harmful AI-generated content like deepfakes. We have seen that while the Act provides a foundational toolkit, its effectiveness is challenged by the nature of synthetic media.

Key Takeaways:

• Existing Provisions: Harmful deepfakes are primarily tackled using Sections 67 and 67A (obscene/explicit content) and Sections 66C, 66D, and 66E (impersonation/privacy violation) of the IT Act.
• Analytical Gaps: The language of these pre-AI statutes, such as the word "captures" in Section 66E, creates potential legal loopholes when applied to purely synthetic content.
• Regulatory Evolution: The government is closing these gaps through a combination of advisories and amendments to the IT Rules, 2021. The new focus is on creating clear definitions for synthetic media, mandating labeling for transparency, and reinforcing platform accountability.
• Layered Approach: The legal strategy is not a single law but a multi-layered approach combining criminal liability for creators with due diligence obligations for intermediaries.

Preview of the Next Lesson:

This lesson concludes our module on the IT Act. We have covered intermediary liability, cybersecurity, and now, content regulation. We will now pivot to a new and fascinating domain in Module 7: AI in the Justice System. Our first lesson will explore the foundational concept of 'access to justice' and analyze how AI technologies can be a double-edged sword, both enhancing and impeding it. This will prepare us for later discussions on predictive justice and the automation of legal practice.