India has mandated social media platforms to enhance their efforts in regulating deepfakes and AI-created impersonations, while also significantly shortening the window for complying with content removal directives. This action has the potential to alter how global technology companies manage content in one of the world’s largest and most rapidly expanding internet service markets.
The amendments to India’s 2021 IT Rules, released (PDF) on Tuesday, formally establish a regulatory framework for deepfakes. These revisions necessitate the labeling and traceability of synthetic audio and visual material, alongside dramatically reduced compliance deadlines for platforms. This includes a three-hour limit for official takedown requests and a two-hour window for certain pressing user-initiated complaints.
The new regulations hold considerable weight due to India’s prominence as a digital market. Housing over a billion internet users, largely a young demographic, the South Asian nation represents a crucial market for major platforms such as Meta and YouTube, suggesting that compliance strategies adopted in India will likely shape global product development and content moderation practices.
Pursuant to the updated regulations, social media platforms that enable the upload or sharing of audio-visual content are now required to mandate disclosures regarding whether content is synthetically generated. They must also deploy tools to authenticate such claims and ensure that deepfakes are distinctly labeled and contain embedded, traceable provenance information.
The rules explicitly prohibit specific categories of synthetic content, including deceptive impersonations, non-consensual intimate imagery, and material associated with severe criminal activity. Failure to comply, particularly in instances reported by authorities or users, could jeopardize companies’ safe-harbour protections under Indian law, exposing them to increased legal liability.
Meeting these obligations is heavily reliant on automated systems, with platforms expected to implement technical tools for verifying user disclosures, identifying and labeling deepfakes, and proactively preventing the creation or dissemination of prohibited synthetic content.
Rohit Kumar, a founding partner at The Quantum Hub, a New Delhi-based policy consulting firm, observed that “The amended IT Rules signify a more considered method for regulating AI-generated deepfakes.” He further noted, “The considerably reduced grievance timelines — such as the two- to three-hour takedown windows — will significantly escalate compliance burdens and warrant careful examination, particularly given that non-compliance is linked to the loss of safe harbour protections.”
Aprajita Rana, a partner at the prominent Indian corporate law firm AZB & Partners, stated that the regulations now specifically target AI-generated audio-visual content, rather than all online information, and make exceptions for routine, aesthetic, or efficiency-driven applications of AI. Nevertheless, she cautioned that the stipulation requiring intermediaries to remove content within three hours of becoming aware of it diverges from established free-speech principles.
“The law, however, continues to require intermediaries to remove content upon being aware or receiving actual knowledge, that too within three hours,” Rana reiterated, mentioning that the labeling stipulations would be applied across various formats to curtail the dissemination of child sexual abuse material and misleading content.
The Internet Freedom Foundation, a digital advocacy group based in New Delhi, expressed concerns that the regulations might accelerate censorship due to significantly shortened takedown periods. They argued this leaves minimal opportunity for human review and could compel platforms towards excessive automated content removal. In a statement posted on X, the group also raised objections regarding the expansion of prohibited content categories and clauses allowing platforms to disclose user identities to private complainants without requiring judicial scrutiny.
“These impossibly short timelines eliminate any meaningful human review,” the group stated, warning that these changes could jeopardize free-speech safeguards and due process.
According to two industry sources who spoke with TechCrunch, the amendments emerged from a restricted consultation process, with only a small number of suggestions being incorporated into the final rules. While the Indian government seemingly accepted proposals to narrow the scope of information covered—focusing specifically on AI-generated audio-visual content rather than all online material—other recommendations were not adopted. The sources indicated that the magnitude of alterations between the draft and final rules justified an additional round of consultation to offer companies clearer guidance on compliance expectations.
Government authority to order content takedowns has been a persistent source of dispute in India. Social media platforms and civil-society organizations have consistently criticized the extensive and obscure nature of content removal directives. Even Elon Musk’s X initiated legal action against New Delhi regarding orders to block or remove posts, asserting that these actions constituted overreach and lacked adequate safeguards.
Requests for comments from Meta, Google, Snap, X, and the Indian IT ministry went unanswered.
These most recent revisions occur merely months after the Indian government, in October 2025, decreased the number of officials permitted to issue orders for internet content removals. This action was taken in response to a legal challenge by X concerning the extent and transparency of takedown authorities.
The amended regulations are scheduled to become effective on February 20, leaving platforms with limited time to adapt their compliance systems. This implementation coincides with India’s hosting of the AI Impact Summit in New Delhi, from February 16 to 20, an event anticipated to attract leading global technology executives and policymakers to the nation.