India Mandates AI Content Labels, 3-Hour Deepfake Takedown Rule Comes Into Force from Feb 20, 2026

AI-generated photos, videos, and audio must now carry mandatory labels; social media platforms face stricter liability under new IT Amendment Rules, 2026.

February 20, 2026: In a major regulatory move to curb deepfakes, misinformation, and digital impersonation, the Government of India has made it mandatory for all AI-generated audio, video, and images to carry clear identification labels. Under the new IT Amendment Rules, 2026, social media platforms are now legally bound to remove objectionable or illegal AI content within three hours of receiving a complaint or official order.

The notification for the new rules was issued on February 10, 2026, and the regulations officially came into force today, February 20, 2026, marking one of India’s strictest digital governance reforms in the AI era.

PM’s Push for Content Authenticity

A day before the rules took effect, Prime Minister Narendra Modi addressed the Global AI Summit on February 19, calling for an “authenticity label” on digital content.

Drawing a parallel with food packaging, the Prime Minister said that just as products carry nutrition labels, digital content must also carry authenticity labels so citizens can easily distinguish between real and AI-fabricated material.

Key Provisions of the New AI Rules

1. Mandatory AI Labels (“Digital Stamp”)

All AI-generated photos, videos, and audio files must carry a visible “AI-generated” label.
Example: If a video shows a politician delivering a speech created using AI, the video must clearly display an AI-generated tag on-screen.

2. Technical Marker System (“Digital DNA”)

Each AI-generated file will carry hidden metadata — described as “digital DNA” — containing:

• Date of creation
• AI tool/software used
• Platform of first upload

This technical marker will enable law enforcement agencies to trace the source of AI-generated crimes, including impersonation, fraud, and misinformation.

3. Tamper-Proof Labels

Removing or editing AI labels, watermarks, or metadata is now illegal.
If any platform detects tampering with AI labels or metadata:

• The content will be auto-deleted
• The uploader and platform may face legal action

4. Criminalisation of Harmful AI Use

Use of AI for the following will be treated as serious criminal offences:

• Child sexual abuse material
• Obscenity
• Fraud and financial crimes
• Weapons-related content
• Identity theft and impersonation
• Deepfake political manipulation

5. 3-Hour Takedown Rule (Earlier: 36 Hours)

The content removal deadline has been reduced from 36 hours to just 3 hours after:

• A user complaint
• A court order
• A government directive

An officer of at least DIG rank in the police department is authorised to initiate takedown proceedings.

6. Platform Liability for False Disclosure

Social media platforms must now:

• Take a user declaration on whether content is AI-generated
• Deploy AI verification tools to check false claims

If AI content is published without disclosure, the platform will be held legally responsible.

Government’s Stand

The Ministry of Electronics and Information Technology stated that the new rules aim to build an:

“open, safe, trusted, and accountable internet ecosystem.”

The government clarified that the reforms are designed to tackle:

• Misinformation
• Election manipulation
• Digital impersonation
• AI-based fraud
• Identity theft
• Synthetic media crimes

Legal Provisions and Penalties

Violations will attract action under existing laws:

• Bharatiya Nyaya Sanhita (BNS) Section 353 – Up to 3 years imprisonment for spreading fear or hatred using false information
• BNS Section 336 – Up to 2 years imprisonment for AI-based impersonation
• IT Act, 2000 – Section 79 – Platforms lose safe harbour protection if they fail to comply

Safe Harbour Explained Simply

• Legal Shield: Platforms are not automatically liable for user content
• Conditional Protection: Protection applies only if rules are followed
• Violation Consequence: Failure to remove AI content within 3 hours results in loss of legal immunity and legal action

What is a Deepfake?

A deepfake is AI-generated content where a person’s face or voice is digitally replaced, making fabricated videos or audio appear completely real, often used for fraud, misinformation, and impersonation.