India Mandates AI Content Labels and Three-Hour Takedown Rule for Deepfakes
Starting February 20, 2026, India has enforced strict new regulations governing artificial intelligence-generated content across digital platforms. The Central Government’s directive, first announced on February 10, requires all AI-created photos, videos, and audio files to carry a clear “AI Generated” label. Social media companies now face a drastically reduced deadline of just three hours to remove flagged objectionable or illegal content, down from the previous 36-hour window.
These measures aim to curb the spread of deepfakes, misinformation, and digital fraud while building a safer, more trustworthy internet ecosystem. The rules apply to all platforms operating in India, holding both users and companies accountable for compliance.
Prime Minister’s Call for Digital Authenticity Labels
Just one day before the rules took effect, Prime Minister Narendra Modi addressed an AI summit on February 19, emphasizing the need for authenticity labels on digital content. Drawing a parallel to nutritional labels on food products, he argued that digital content should also clearly indicate whether it is genuine or AI-generated. Such transparency, he noted, would help citizens distinguish between real and fabricated material online.
Key Provisions of the New AI Content Rules
Mandatory “AI Generated” Stamp on Visual and Audio Content
Every piece of AI-generated content must now display a visible label such as “AI Generated” in a corner of the video, image, or audio file. For instance, a video featuring an AI-synthesized speech by a political leader cannot be posted without this disclosure. The label must remain intact and legible throughout the content’s lifecycle.
Embedded Digital DNA in Metadata
Beyond visible labels, the rules require AI-generated files to carry embedded metadata—often described as a digital DNA. This hidden information records when the content was created, which AI tool generated it, and the first platform where it was uploaded. Law enforcement agencies can trace this technical marker to identify the original source in cases involving AI-assisted crimes, fraud, or identity theft.
Strict Penalties for Removing Labels
Previously, users could edit or remove watermarks from AI content without consequence. Under the new framework, tampering with labels or metadata is illegal. Platforms must deploy technology that automatically deletes or blocks content if anyone attempts to alter or strip the AI disclosure markers.
Zero Tolerance for Harmful AI Content
The government has adopted a particularly firm stance on AI-generated content involving children. Using artificial intelligence to create child sexual abuse material, facilitate fraud, spread information about weapons, or impersonate individuals will be treated as serious criminal offenses. Authorities have warned of stringent legal action in such cases.
Revised Takedown Timeline for Social Media Platforms
Under the amended IT rules, social media companies now have only three hours from receiving a complaint to remove unlawful or objectionable content. This represents a significant reduction from the earlier 36-hour window, substantially increasing platform accountability. Non-compliance could result in penalties or loss of legal protections under India’s intermediary liability framework.
User Declarations and Platform Responsibility
When uploading content, users must now explicitly declare whether the material was created using AI. Platforms are required to implement verification tools to check the accuracy of these user claims. If AI-generated content gets published without proper disclosure, the platform itself shares responsibility for the violation.
The Ministry of Information Technology has stated that these regulations are designed to foster an “open, safe, trusted, and accountable internet.” Officials believe the rules will help control challenges posed by generative AI, including the spread of false information, identity theft, and electoral manipulation.
Understanding Deepfakes
Deepfake technology uses AI tools to superimpose a real person’s face or voice onto another video or audio recording, making the result appear convincingly authentic. This capability has raised serious concerns about consent, privacy, and public trust, prompting the government to tighten regulations around its use.
What This Means for Content Creators and Platforms
- Creators must label all AI-generated content before publishing
- Platforms need to invest in detection and verification technologies
- Both parties face legal consequences for non-compliance
- Three-hour takedown window applies to all objectionable content
- Metadata manipulation is now a punishable offense
The new framework represents one of the most comprehensive approaches to AI content regulation globally. By combining mandatory labeling, metadata tracking, and rapid takedown requirements, India aims to strike a balance between technological innovation and public safety in the digital space.
