Regulation Of Deepfakes And Synthetic Media-Legal Gaps And Proposals

Introduction

Technology today allows us to create almost anything on a screen. With artificial intelligence, videos, photos, and even voices can be fabricated so convincingly that it becomes nearly impossible to tell what is real. These are called deepfakes—a form of synthetic media.

While such tools can be useful in films, education, or accessibility projects, they are more often linked with harm. Deepfakes have been misused in India and abroad to create non-consensual intimate videos, political propaganda, financial scams, and reputational attacks.

India does not yet have a specific law on deepfakes. Victims rely on scattered provisions of the Information Technology Act, 2000 (IT Act), the Bharatiya Nyaya Sanhita, 2023 (BNS), and the IT Rules, 2021. But these laws were not drafted with artificial intelligence in mind. This article explains how Indian law currently deals with deepfakes, what the gaps are, what other countries are doing, and what reforms could help.

What Are Deepfakes and Synthetic Media?

Deepfakes are made using machine learning models such as generative adversarial networks (GANs). These systems are trained to swap faces, mimic voices, or generate entirely new but realistic media. Unlike ordinary editing, deepfakes are much harder to spot.

Deepfakes are not always harmful. They can be used in movies, digital art, gaming, advertising, or accessibility tools (for example, to help the speech-impaired). But in practice, most attention comes from their abuse. A 2019 report found that 96% of deepfakes online were pornographic, often targeting women without consent.¹ The danger is not only personal harm but also the erosion of public trust in information.

Current Legal Framework in India

Information Technology Act, 2000

The IT Act has some provisions that can apply to deepfakes:

• Section 66C punishes identity theft.²
• Section 66D punishes cheating by impersonation using computers.³
• Sections 67 and 67A criminalise publishing obscene or sexually explicit material online.⁴
• Section 69A allows the government to block unlawful content.

These provisions give some tools but were not designed for AI-generated manipulation.

Bharatiya Nyaya Sanhita, 2023

The new BNS also has sections that could be stretched to cover deepfakes:

• Section 336 – Forgery of documents or electronic records.
• Section 351 – Use of false electronic records.
• Section 354 – Sexual harassment (relevant for non-consensual pornographic deepfakes).
• Section 356 – Criminal defamation.⁵

These provisions may apply if a deepfake harms reputation, impersonates someone, or violates dignity. But again, none of them mention deepfakes specifically.

Intermediary Liability

The IT (Intermediary Guidelines and Digital Media Ethics Code) Rules, 2021 require platforms to remove illegal content once notified.⁶ However, there is no duty to proactively detect deepfakes or to label AI-generated media. This makes enforcement weak.

Global Approaches

Different countries are moving ahead faster than India:

1. European Union

   The AI Act (2024 draft) requires clear disclosure and labelling of deepfakes.⁷

2. United States

   States like California and Texas have banned election-related deepfakes close to polling.⁸

3. China

   The Deep Synthesis Provisions (2023) require watermarks on all AI-generated media and user identity verification.⁹

4. United Kingdom

   The Online Safety Bill (2023) addresses harmful manipulated content.

The common themes are definitions, labelling, platform duties, and stronger victim protection.

Legal Gaps in India

Even with the IT Act, BNS, and IT Rules, the Indian framework has major shortcomings:

1. No definition – “Deepfake” or “synthetic media” has no legal meaning.
2. Weak remedies for victims – Takedowns are slow, and compensation is rare.
3. Limited platform responsibility – No proactive detection or watermarking requirements.
4. Jurisdiction issues – Deepfakes spread across borders, while Indian law is territorially limited.
5. Risk to free speech – Over-broad laws could restrict satire, parody, or legitimate creative uses.

Judicial and Policy Developments

Indian courts have not yet ruled directly on deepfakes, but existing judgments guide the debate:

• Justice K.S. Puttaswamy v. Union of India (2017) recognised the right to privacy under Article 21.¹⁰ Circulating deepfake pornography directly violates this right.
• Shreya Singhal v. Union of India (2015) struck down vague restrictions on online speech, stressing that any law must be precise.¹¹ Future deepfake laws must respect this standard.
• Swami Ramdev v. Facebook Inc. (2019, Delhi HC) ordered global takedowns of defamatory content, showing courts are willing to extend online liability across borders.¹²

Policy discussions are also ongoing. NITI Aayog’s “Responsible AI” report (2021) called for transparency and accountability in AI, but it did not propose deepfake-specific rules.

Proposals for Reform

To close these gaps, India needs a comprehensive deepfake law. Some possible steps include:

1. Define Deepfakes in LawCreate a clear definition of “deepfake” and “synthetic media.” Distinguish harmful uses (e.g., fraud, porn, misinformation) from harmless or beneficial uses (e.g., films, research).
2. New Offences under BNSAdd specific offences for creating, sharing, or profiting from malicious deepfakes, with higher penalties for non-consensual sexual content and election interference.
3. Platform DutiesMandatory watermarking or labelling of AI-generated media. Faster takedown timelines (24–48 hours). Penalties for repeated non-compliance.
4. Victim RemediesA statutory right to erasure for victims of deepfakes. Compensation through fast-track civil claims. Special cyber cells to handle deepfake complaints.

Support research in AI-based deepfake detection. Use blockchain or digital signatures to verify authentic media. Run public awareness campaigns and media literacy programs.

Conclusion

Deepfakes are more than just a technological trend—they are a direct challenge to privacy, dignity, democracy, and trust in information. India’s current legal tools under the IT Act, BNS, and IT Rules are useful but fragmented.

Learning from global models, India must move towards a specific, balanced legal framework: one that criminalises malicious deepfakes, mandates platform responsibility, and protects victims, while still allowing space for innovation and creativity.

The longer the regulation is delayed, the more risks society faces—from harassment of individuals to manipulation of entire democratic processes. The law must step in before trust in digital information is permanently eroded.