Hero image: MART PRODUCTION / Pexels
India Tightens Deepfake Rules, AI Content Label Mandate
The Indian government has amended IT Rules to require AI-generated content to be labelled and deepfakes to be removed within 24 hours of complaint, according to two independent outlets. The changes mark a shift from voluntary compliance to mandatory obligations for platforms and creators, raising questions about enforcement and user safety.
India’s Ministry of Electronics and Information Technology (MeitY) has moved to formalize obligations around synthetic media and deepfakes through amendments to the Information Technology (Intermediary Guidelines and Digital Media Ethics Code) Rules, 2021. Two independent outlets—MediaNews4U and India Today NE—report that the amendments introduce mandatory labelling of AI-generated content and impose a 24-hour deadline for platforms to remove deepfakes after receiving a complaint. This development follows rising public concern over manipulated media, including impersonation of public figures and disinformation campaigns during elections. While both outlets describe the same regulatory shift, they differ in emphasis: MediaNews4U highlights the legal obligations on intermediaries and creators, while India Today NE underscores the government’s stated intent to curb misinformation and protect digital citizens.
—
Introduction to India’s IT Rules and Deepfakes
The IT Rules, first notified in 2021, govern online intermediaries including social media platforms, messaging services, and digital publishers. They require due diligence, grievance redressal mechanisms, and periodic compliance reporting. Deepfakes—hyper-realistic synthetic media created using artificial intelligence—have emerged as a major vector for disinformation, impersonation, and reputational harm, prompting governments worldwide to consider stricter controls. India’s amendments reflect a global trend toward regulating synthetic content, with the European Union’s AI Act and U.S. state-level deepfake laws serving as reference points.
Both MediaNews4U and India Today NE describe the amendments as a response to the rapid proliferation of deepfakes, including those targeting celebrities, politicians, and private citizens. According to India Today NE, the government has framed the changes as necessary to protect “digital nagriks” (citizens) from harm caused by manipulated media. MediaNews4U, meanwhile, emphasizes the legal obligations imposed on intermediaries, including social media companies and AI platforms, to act within 24 hours of a verified complaint. The amendments reportedly apply to all “intermediaries” as defined under Section 2(w) of the IT Act, which includes cloud storage providers, social networks, and AI model deployers.
—
MediaNews4U and India Today NE Reporting: Where They Agree and Diverge
Core Agreement: Mandatory Labelling and 24-Hour Takedown
Both outlets report that the amended IT Rules now require:
- AI-generated content to be clearly labelled as synthetic or AI-created
- Intermediaries to remove deepfakes within 24 hours of receiving a verified complaint
- Platforms to publish monthly compliance reports on the number of complaints received and actions taken
MediaNews4U specifies that the labelling requirement applies to “all content generated or modified using artificial intelligence,” including text, images, audio, and video. India Today NE echoes this but adds that the government may issue further guidelines on what constitutes “adequate labelling,” suggesting potential ambiguity in enforcement.
Divergences in Emphasis and Detail
MediaNews4U places greater emphasis on the legal obligations imposed on intermediaries and creators, noting that failure to comply could result in loss of safe-harbour protections under Section 79 of the IT Act. It also highlights that the rules cover “all intermediaries,” which may include AI model developers and cloud storage providers, not just social media platforms.
India Today NE, by contrast, focuses more on the government’s stated intent to protect users and curb misinformation. It quotes unnamed officials as saying the amendments are aimed at “preventing misuse of technology” and “safeguarding democratic processes.” India Today NE also includes a brief mention of potential penalties for non-compliance, though it does not specify what those penalties might be. MediaNews4U does not address penalties beyond the loss of intermediary protections.
Additionally, India Today NE includes a reference to “public consultations” that informed the amendments, suggesting a participatory process, while MediaNews4U does not mention consultations at all.
—
The Claim: Mandating AI Content Labelling and Faster Deepfake Removal
AI Content Labelling as a Regulatory Standard
Both outlets report that the amended rules require AI-generated content to be labelled as such. MediaNews4U describes this as a blanket requirement across all AI-generated media, while India Today NE frames it as a response to “deepfake proliferation.” The labelling requirement appears to apply to both user-generated content and content generated by AI systems deployed by platforms.
Notably, neither outlet provides a definition of “AI-generated content,” leaving open questions about whether the requirement applies to content lightly edited using AI tools (e.g., AI-enhanced filters) or only to fully synthetic media. This ambiguity may create enforcement challenges.
24-Hour Takedown Deadline for Deepfakes
Both outlets agree that intermediaries must remove deepfakes within 24 hours of receiving a verified complaint. MediaNews4U specifies that the complaint must be “verified,” though it does not define what verification entails. India Today NE states that the government will issue further guidelines on verification, indicating that the process may involve a government or third-party verification mechanism.
MediaNews4U emphasizes that the 24-hour deadline applies to all intermediaries, including cloud storage providers and AI model deployers, while India Today NE focuses on social media platforms and messaging services. This divergence suggests potential confusion about which entities are covered and how the deadline would be enforced across different types of intermediaries.
—
Cross-Referencing Outlets: A Combined Evidence Approach
Convergence on Key Requirements
Both outlets converge on three core elements:
- Mandatory labelling of AI-generated content
- 24-hour removal of deepfakes after verified complaint
- Monthly compliance reporting by intermediaries
This convergence strengthens the credibility of the reported amendments, as two independent outlets with different editorial focuses have arrived at similar descriptions of the rules. The consistency suggests that the amendments are likely real and substantial, rather than minor clarifications or rumours.
Divergences in Scope and Enforcement Details
Where the outlets diverge, the differences are largely in emphasis and detail rather than factual contradiction. MediaNews4U’s focus on legal obligations and safe-harbour implications suggests a more technical reading of the rules, while India Today NE’s emphasis on user protection and democratic safeguards reflects a policy-oriented perspective.
The lack of detail on verification mechanisms, definitions of “AI-generated content,” and potential penalties creates gaps in the public record. These gaps are not contradictions between the outlets, but rather areas where the rules themselves appear to be under-specified.
—
Original Analysis: What the Pattern Across Sources Suggests
Taken together, these reports suggest that India is moving toward a regulatory model that treats synthetic media as a form of harmful content requiring swift removal, rather than as a form of protected speech or artistic expression. The 24-hour takedown window is unusually rapid by global standards—comparable to the EU’s Digital Services Act’s 24-hour deadline for illegal content removal, but applied here to deepfakes specifically. This speed reflects a prioritization of harm prevention over due process, which may reduce false positives but could also chill legitimate uses of AI, such as satire or creative expression.
The labelling requirement, while seemingly straightforward, introduces significant ambiguity. AI tools are increasingly embedded in everyday software, from photo editors to email clients. Requiring every output to be labelled as “AI-generated” could create a labelling burden that is both impractical and misleading—users may not know whether an AI tool was used in the creation of a piece of content, and platforms may struggle to detect AI-generated content in real time. The absence of a clear definition in either report raises concerns about arbitrary enforcement.
Moreover, the focus on intermediaries—rather than creators—suggests that the government views platforms as the primary locus of control. This aligns with India’s long-standing approach to online regulation, where intermediaries bear much of the compliance burden. However, it may overlook the fact that many deepfakes originate from private accounts or encrypted platforms, where intermediaries have limited visibility into content.
Finally, the lack of detail on penalties or verification mechanisms suggests that the rules are still in a formative stage. The government may be relying on the threat of losing safe-harbour protections to incentivize compliance, but without clear penalties or a transparent verification process, enforcement could be inconsistent. This could lead to a situation where some platforms comply rigorously while others ignore the rules, creating a patchwork of protections for users.
—
Expert and Institutional Response to India’s Deepfake Regulations
Industry and Civil Society Reactions
While neither outlet quotes specific experts or civil society organizations, both reports suggest that the amendments are likely to spark debate. MediaNews4U’s emphasis on legal obligations implies that industry groups may push back on the scope of coverage, particularly regarding AI model developers and cloud providers. India Today NE’s focus on user protection suggests that civil society groups may welcome the rules but raise concerns about over-censorship or lack of due process.
Given India’s status as a major digital market, the amendments are likely to draw attention from global platforms such as Meta, Google, and X (formerly Twitter), all of which operate in India and are subject to the IT Rules. These companies may seek clarity on definitions, verification processes, and penalties before adjusting their content moderation policies.
Government’s Stated Objectives
India Today NE reports that government officials framed the amendments as necessary to “prevent misuse of technology” and “safeguard democratic processes.” This framing aligns with India’s broader approach to online regulation, which has increasingly focused on curbing misinformation and protecting national security. MediaNews4U does not quote officials but implies a similar intent through its description of the rules.
The mention of “public consultations” in India Today NE’s report suggests that the government may have engaged with stakeholders, though neither outlet provides details on the nature or outcome of these consultations. This could indicate a more consultative process than previous IT Rule amendments, which were often criticized for lack of transparency.
—
Red Flags and Debunking Checklist for Deepfake Detection
Detecting deepfakes requires a combination of technical tools, media literacy, and critical thinking. Below is a checklist of red flags and legitimate signals to help users and platforms identify potentially manipulated media.
- Inconsistent Lighting or Shadows: Deepfakes often fail to replicate realistic lighting, especially in facial regions. Look for unnatural shadows or highlights that don’t match the scene.
- Unnatural Facial Movements: Blinking, lip sync, and micro-expressions may appear exaggerated, delayed, or absent in deepfakes. Watch for jerky or robotic movements.
- Audio-Visual Mismatch: Deepfake audio may not sync with lip movements, or the voice may sound unnaturally modulated. Use tools like Adobe’s VoCo or ElevenLabs to compare voice patterns.
- Background Anomalies: Deepfakes often struggle to render background elements accurately. Look for blurring, warping, or inconsistencies in objects or scenery.
- Metadata and Source Verification: Check file metadata for clues about origin, editing software, or timestamps. Reverse-image search tools can help trace the original source of an image or video.
- Contextual Inconsistencies: Deepfakes are often shared out of context or with misleading captions. Verify the source and timeline of the content before sharing.
- AI Detection Tools: Platforms and third-party tools (e.g., Microsoft Video Authenticator, Deepware Scanner) can flag synthetic media with varying degrees of accuracy. Use them as a first pass, but verify results independently.
- User Reports and Community Flags: Crowdsourced verification can help identify deepfakes, but be cautious of coordinated disinformation campaigns that may exploit reporting systems.
—
What to Do About Deepfakes in India: A Call to Action
For Users: How to Respond When You Encounter a Deepfake
If you suspect a deepfake, do not share it without verification. Report it to the platform using the designated grievance mechanism, and file a complaint with the Ministry of Electronics and Information Technology (MeitY) through its official grievance portal. Provide as much context as possible, including the original source (if available) and any inconsistencies you’ve observed. If the deepfake involves a public figure or could cause reputational harm, consider reaching out to the individual or organization directly to alert them.
Media literacy is your strongest defence. Learn to spot the red flags listed above, and encourage others to do the same. If you’re unsure whether content is real, assume it might not be until verified by a trusted source.
For Platforms: Implementing the New Rules
Platforms must update their content moderation policies to comply with the 24-hour takedown deadline and labelling requirements. This will require:
- Investing in AI detection tools to identify deepfakes and synthetic content at scale.
- Establishing clear verification processes for complaints, including collaboration with fact-checkers and government agencies where necessary.
- Training moderators to handle deepfake complaints efficiently and consistently.
- Publishing monthly compliance reports as mandated, detailing the number of complaints received, actions taken, and any challenges faced.
Platforms should also consider proactive measures, such as watermarking AI-generated content at the point of creation or deployment, to reduce the burden on users to identify synthetic media.
For Creators and AI Developers: Ethical Considerations
Creators and AI developers have a responsibility to ensure their tools are not used to deceive. Consider implementing safeguards such as:
- Default labelling of AI-generated outputs in user interfaces.
- Restrictions on the use of AI tools to impersonate real individuals without explicit consent.
- Collaboration with platforms and regulators to develop industry-wide standards for synthetic content.
Developers should also document the capabilities and limitations of their tools to help users and platforms assess potential risks.
For Policymakers: Addressing Gaps in the Rules
The current amendments leave several critical questions unanswered. Policymakers should clarify:
- The definition of “AI-generated content” and what constitutes “adequate labelling.”
- The verification process for complaints, including the role of government agencies, fact-checkers, and platforms.
- The penalties for non-compliance, including whether intermediaries can be held liable for failing to meet the 24-hour deadline.
- The scope of coverage for intermediaries, particularly AI model developers and cloud storage providers.
Without these clarifications, the rules risk being unevenly enforced or challenged in court. A transparent consultation process with industry, civil society, and technical experts could help address these gaps.
—
Red Flags vs. Legitimate Signals: A Comparative Table
| Category | Red Flags (Indicators of Deepfake) | Legitimate Signals (Indicators of Authentic Content) |
|---|---|---|
| Facial Movements | Unnatural blinking, lip sync errors, or robotic expressions | Smooth, natural blinking and facial expressions consistent with the context |
| Lighting and Shadows | Inconsistent lighting or shadows that don’t match the scene | Uniform lighting and shadows that align with the environment |
| Audio-Visual Sync | Mismatch between lip movements and spoken words | Lip movements and audio are perfectly synchronized |
| Background Details | Blurring, warping, or inconsistencies in background objects | Clear, consistent background details that match the scene |
| Metadata | Missing or altered metadata, or metadata indicating heavy editing | Intact metadata with timestamps and source information |
| Source Verification | Content shared from unknown or suspicious accounts without context | Content shared from verified accounts with clear context and attribution |
—
FAQ
What are the IT Rules, and how do they relate to deepfakes?
The IT Rules, notified in 2021, govern online intermediaries including social media platforms, messaging services, and digital publishers. The 2026 amendments introduce mandatory labelling of AI-generated content and a 24-hour deadline for platforms to remove deepfakes after receiving a verified complaint. These changes aim to address the proliferation of synthetic media and its potential for harm.
Who is covered by the new rules?
According to both MediaNews4U and India Today NE, the rules apply to “all intermediaries” as defined under Section 2(w) of the IT Act. This likely includes social media platforms, messaging services, cloud storage providers, and AI model developers. MediaNews4U emphasizes that the rules cover a broad range of intermediaries, while India Today NE focuses more on social media platforms.
What counts as “AI-generated content” under the rules?
Neither outlet provides a clear definition. MediaNews4U describes the requirement as applying to “all content generated or modified using artificial intelligence,” while India Today NE does not define the term. This ambiguity could lead to inconsistent enforcement, particularly for lightly edited content or AI-enhanced media.
How will the 24-hour takedown deadline work in practice?
Both outlets report that intermediaries must remove deepfakes within 24 hours of receiving a verified complaint. MediaNews4U specifies that the complaint must be “verified,” though it does not define the verification process. India Today NE states that the government will issue further guidelines on verification, suggesting that the process may involve government or third-party verification mechanisms.
What are the penalties for non-compliance?
MediaNews4U notes that failure to comply could result in the loss of safe-harbour protections under Section 79 of the IT Act, which shields intermediaries from liability for user-generated content. India Today NE mentions potential penalties but does not specify what they might be. Neither outlet provides detailed information on enforcement mechanisms or penalties.
—