Deep Fake Tech

4 min read

Introduction

photo_2026-01-25_20-51-19
  • Deep fake technology has emerged as a new tool of the disinformation threat.
  • Deepfake technology is a hyper-realistic digital falsification technique for manipulating videos, images, and audio using powerful computers, deep learning, and artificial intelligence.
  • Deepfake can alter existing content by replacing one person with another or create new content depicting false events..
  • It can be used to damage reputations, fabricate evidence, defraud the public, and undermine trust in democratic institutions.

Section 66E of the IT Act, 2000

Section 66E of IT Act 2000, addresses deepfake crimes involving unauthorized use of someone's images, infringing on privacy, with penalties of up to three years in jail or a fine up to ₹2 lakh.

Amendments to IT Rules 2021

In October 2023, India released draft amendments to the Information Technology (Guidelines and Digital Media Ethics Code) Rules, 2021.

Current regulatory mechanism

The current regulatory mechanism by the Ministry of Electronics and Information Technology approach. The primary responsibility is placed on the intermediaries (social media platforms) for the following processes:

  1. Identification - platforms must use technical tools to detect synthetic media.
  2. Labeling - any contents generated or modified by EL must be clearly labeled to inform the user.
  3. Takedown timelines - intermediaries are required to remove non-consensual sexual content or disinformation within specified hours of being flagged.
  4. Grievance regression - establishing a mechanism where users can report defects and receive timely resolutions.

Pro-activeness of the judiciary

Individuals' right to their voice likeness and persona

The Delhi High Court in Anil Kapoor v. Simply Life India judgement protected an individual's right to their voice, likeness, and persona against the AI misuse.

Commercial value of a person's identity

Faces involving celebrities in Aishwarya Rai Bachchan highlight how deepfakes infringe upon the commercial value of a person's identity.

Major challenges and concerns

The burden on the intermediaries

Social media platforms may become arbiters of truth. To avoid the legal liability, they might over censor content harming satire and free speech.

Technical limitations

By the time a deep fake is detected, the viral damage is already being done.

Lack of provenance

Simple labels can be cropped out or removed. Without embedded metadata, the origin of a file remains untraceable.

Enforcement gap

India likes specialized media forensic labs at the scale needed to handle millions of pieces of content daily.

European Union

The European Union's AI Act focuses on transparency and mandates machine-readable markers for AI content.

United States of America

The United States of America's Take It Down Act specifically targets non-consensual intimate imagery with strict criminal liabilities.

United Kingdom

The United Kingdom's Online Safety Act criminalizes the creation of sexually explicit deep fakes regardless of whether they are shared.

Case of South Korea Deep Fake

Surging non-consensual deep fake pornography

The country is facing a surge in non-consensual deep fake pornography, particularly targeting women, with teenage boys as the primary perpetrators. This issue reflects broader societal problems, including misogyny and weak digital regulations.

Laws passed, but enforcement challenges still persist.

The country passed laws now criminalize viewing or possessing explicit deep fake content, but enforcement challenges persist.

Key contributing factor

  • Lack of education on digital ethics and
  • insufficient regulation of minors on social media
  • Misogynistic attitudes (strongly prejudiced against women.)

Trust eroded between genders

Trust between genders, especially in schools, has been eroded with the Deep Fake technology.

Way forward

Four-pillar strategy

To move from reactive to proactive governance in the case of Deep Fake Tech, India should adopt a four-pillar strategy.

1. Institutional Oversight: A Single Regulator

  • Currently, the regulation is fragmented between MeitY (IT rules) and the MIB (broadcasting) rules.
  • What India needs is an autonomous digital or AI regulator so that a single body would harmonize standards and conduct mandatory audit of AI algorithms.

2. Technological traceability

  • The credit mark system is about moving beyond visible watermarks.
  • India should adopt digital provenance.
  • The concept is all about a digital birth certificate for every AI file that travels with the content across the mediums.
  • Aligning with C2PA (Coalition for Content Provenance and Authenticity) ensures that Indian content is verifiable globally.

3. Risk Tiered Accountability

  • Not all deepfakes are equally dangerous.
  • Low risk, like entertainment, filters, and satires, require basic disclosure.
  • High risk, like elections, state security, and financial markets, which require pre-release certification and rapid kill switch protocols.

4. Regulatory Sandboxes

  • India should allow start-ups to develop detection and watermarking tools in a supervised environment.
  • This safe innovation loop ensures that regulation does not kill the domestic AI industry and help startups.
Built with LogoFlowershow