3 hour AI content removal rule: What it means for Instagram, YouTube and Facebook

The Indian government has introduced a strict three-hour deadline for social media platforms to remove AI-generated or synthetic content flagged by authorities or courts. The move, part of amendments to India’s IT rules, aims to curb the spread of deepfakes, manipulated videos, and misleading synthetic media while ensuring accountability of online platforms.

What the three-hour deadline means

• Platforms including Instagram, YouTube, Facebook, and X must remove AI-generated or synthetic content within three hours of receiving a takedown notice from the government or a court.
• This replaces the earlier 36-hour window, significantly reducing the time available for action.
• If a platform fails to comply, it is deemed to have violated due diligence obligations under India’s IT Act and can face penalties.

How the process works

Flagging the content: The government, courts, or regulatory authorities identify content that is illegal, harmful, or deceptive.

• Issuing a notice: A formal takedown notice is sent to the platform hosting the content.
• Verification and removal: Platforms must verify whether the flagged content is AI-generated or synthetic using automated tools and remove it within three hours.
• Metadata and labels: All AI content must have persistent labels and embedded identifiers. These cannot be removed or altered, ensuring traceability of the content’s origin.
• User warnings: Platforms must also remind users every three months, in English or any Eighth Schedule language, about penalties for misuse of synthetic content.

Scope of content covered

Videos, audio, and audio-visual content created or altered using AI that appears real or authentic. Includes deepfakes, synthetic voices, and altered visuals that could mislead or misrepresent individuals or events.

Exemptions include routine editing, accessibility improvements, educational materials, presentations, research drafts, and illustrative content that does not distort meaning.

Legal implications of non-compliance

SGI (synthetically generated information) involving child sexual abuse, impersonation, explosives, false documents, or obscene content is treated as a criminal offence under relevant laws, including POCSO Act and the Explosive Substances Act.

Platforms not acting within three hours may lose safe harbour protections under Section 79 of the IT Act if violations are deliberate or due diligence is not observed.

Why the government has implemented this rule

• To tackle the rapid spread of deepfakes and synthetic media capable of causing reputational damage, misinformation, or election manipulation.
• To ensure users can distinguish between real and synthetic content, empowering them to make informed decisions online.
• To increase platform accountability, forcing social media companies to proactively monitor and manage AI-generated content.

What industry and experts say

Earlier proposals requiring watermarks on 10% of screen space were dropped after industry feedback, citing difficulty across formats.

Platforms are now required to label AI content prominently without fixed-size watermarks, while still embedding identifiers.

IT Minister Ashwini Vaishnaw stated: “Users must know whether content is synthetic or real. Mandatory labelling and rapid takedown ensure transparency and curb misuse of AI.”

India’s three-hour takedown deadline marks a significant tightening of rules for AI-generated content, combining swift action, clear labelling, and strict accountability for social media platforms. It reflects the government’s commitment to prevent misuse of deepfakes while safeguarding public trust in digital information.