Regulating Speech or Controlling It? A Critical Analysis of the 2026 IT Second Amendment Rules
Introduction
The Ministry of Electronics and Information Technology (MEITy) released the Draft Information Technology (Intermediary Guidelines and Digital Media Ethics Code) Second Amendment Rules, 2026 on March 30, 2026, inviting public comments with a response window closing on April 14. This is a limited 15-day period for public input on proposed rules that will have major constitutional impacts. The brevity and timing of this opportunity demonstrate debatable commitment to stakeholder engagement and meaningful consultation by the drafting agency.
While MEITY describes the proposed amendments as "clarificatory and procedural nature," an analysis shows they will have substantive effects. Collectively, the amended language changes significantly how online speech will be regulated in India by providing the executive with more concentrated regulatory authority, limiting the required transparency of content enforcement, mandating greater retention of data without proportionality-based safeguards, and placing excessive compliance burden on intermediaries. Each of these changes has consequences beyond just changes in process and together, these changes collectively raise substantial concerns regarding compliance with Articles 14, 19, and 21 of the Constitution of India.
The Constitutional Baseline: Shreya Singhal and the Limits of Intermediary Liability
India’s Supreme Court decision in Shreya Singhal v Union of India (2015) 5 SCC 1 provides the foundation for intermediary liability, wherein the Court read down Section 79(3)(b) of the IT Act, 2000, holding that intermediaries are required to act upon receiving actual knowledge only through a court order or a valid notification by the appropriate government authority. The Supreme Court’s decision intended to provide a constitutional protection to intermediaries from being subjected to informal, unverified executive pressure to take down content by requiring that any such order be subject to some level of legal objective credibility or threshold.
Rule 3(4) of the proposed amendments places that balance under significant strain. By requiring intermediaries to comply with advisories, directions, standard operating procedures, codes of practice, and guidelines issued by the Ministry — and tying non-compliance to the loss of safe harbour — the draft effectively lowers the constitutional threshold that Shreya Singhal was designed to maintain. Compliance obligations now potentially arise from instruments that carry no judicial sanction and no mandatory public disclosure.
Rule 3(4): Delegated Legislation or Executive Overreach
The rule-making power conferred on the Central Government under Section 87 of the IT Act is limited to carrying out the provisions of the Act. It does not authorise the creation of new substantive obligations. This principle has been consistently affirmed in Indian Express Newspapers v. Union of India (1985) 1 SCC 641 and Confederation of Ex-Servicemen Associations v. Union of India (2006) 8 SCC 399, where the Court held that delegated legislation must remain within the four corners of the parent statute.
Rule 3(4) tests those limits. It converts executive advisories into binding compliance instruments without a clear statutory foundation in either Section 79 or Section 87. Although the proposed rule requires that such instruments specify their legal basis, there is no requirement that they be published or made publicly accessible. This creates a framework in which legality risks becoming circular — instruments claimed to be lawful solely by reference to a provision that does not clearly authorise them, shielded from scrutiny by their own opacity. Justice Chandurkar’s judgment in Kunal Kamra v. Union of India identified precisely this defect in the Fact Check Unit amendment. Rule 3(4) replicates the structural problem in a broader form.
Compliance Pressure and the Logic of Over-Censorship
The practical consequence of Rule 3(4) lies not only in its legality but in how it reshapes incentive structures for platforms. An intermediary facing the permanent threat of safe harbour loss will not wait to assess the legal merit of each advisory. The rational calculation is to comply early, broadly, and without friction. Lawful content — particularly satire, political commentary, and journalism — becomes vulnerable not because it is unlawful, but because it presents regulatory risk.
This dynamic was visible on 18 March 2026, when stand-up comedian Pulkit Mani (@hunnywhoisfunny) found his satirical Instagram reel being restricted across India. The video had accumulated over 16.5 million views. Users encountered a notice citing Section 79(3)(b) of the IT Act. No reasons were publicly provided. No prior hearing was offered. The same night, several political parody and satire accounts were withheld on X.
Data Retention, Privacy, and the Proportionality Test
The amendments to Rules 3(1)(g) and 3(1)(h) extend data retention obligations by making them additional to requirements under any other law. The existing 180-day floor for retained user data — covering removed content, registration information, and associated records — becomes a minimum rather than a ceiling. No maximum is specified, and no proportionality requirement accompanies the extension.
This raises direct concerns under Article 21 as interpreted in Justice K.S. Puttaswamy v. Union of India (2017) 10 SCC 1, which held that any state intrusion into privacy must satisfy the triple test of legality, necessity, and proportionality. Undefined retention periods, with no statutory ceiling and no requirement of purpose limitation, risk failing all three. The longer user data is held, including metadata, device information, and records of removed content, the greater the exposure to surveillance, unauthorised access, and use beyond the original justification.
Circumventing Judicial Scrutiny Through Procedural Redesign
The Bombay High Court, in its August 2021 order, stayed provisions of the IT Rules’ oversight mechanism as prima facie violative of Article 19(1)(a). The Madras High Court in T.M. Krishna v. Union of India affirmed that stay, cautioning that government-controlled media oversight risked undermining press independence. Both matters remain pending before the Delhi High Court.
The amendments to Rules 8(1) and 14 restructure the same oversight machinery through a modified procedural design. By extending the Inter-Departmental Committee’s jurisdiction to cover “matters” referred by the Ministry with no requirement of a complainant, no defined subject matter, and no guaranteed prior hearing, the proposed rules effectively reconstitute what courts found constitutionally suspect. Individual users posting news and current affairs content are now brought within reach of blocking mechanisms originally designed for institutional publishers.
Conclusion
As seen above, the Draft IT Rules 2026 are unable to meet the constitutional and judicial requirements to regulate free speech. What the proposed amendments construct is a durable system in which platforms self-censor under liability pressure, data is retained without proportionate justification, and content oversight expands through procedural adjustment rather than parliamentary legislation. Regulation of the digital public sphere is both legitimate and necessary. But it must be anchored in law, not in the quiet authority of executive advisories. The law must ultimately remain anchored in constitutional values, guided by the enduring principles of justice, equity, and good conscience.
The comment period closes on 14 April 2026.
Submissions may be sent to itrules.consultation@meity.gov.in.
References
- https://www.meity.gov.in/static/uploads/2026/03/30591fc6e322dcbcc9dae84a0f02e9e7.pdf
- https://www.meity.gov.in/static/uploads/2026/03/a71a21d35c107f2e528363d3eb17646a.pdf
- https://www.meity.gov.in/static/uploads/2026/02/550681ab908f8afb135b0ad42816a1c9.pdf
- https://neopolitico.com/india/government-blocks-viral-satirical-reel-impersonating-pm-modi-raising-fresh-questions-on-free-speech-and-digital-regulation/
- https://internetfreedom.in/sound-the-alarm-iffs-first-read-on-meitys-draft-it-rules-second-amendment-2026/
Related Blogs
Introduction
Earlier this month, lawmakers in Colorado, a U.S. state, were summoned to a special legislative session to rewrite their newly passed Artificial Intelligence (AI) law before it even takes effect. Although the discussion taking place in Denver may seem distant, evolving regulations like this one directly address issues that India will soon encounter as we forge our own course for AI governance.
The Colorado Artificial Intelligence Act
Colorado became the first U.S. state to pass a comprehensive AI accountability law, set to come into force in 2026. It aims to protect people from bias, discrimination, and harm caused by predictive algorithms since AI tools have been known to reproduce societal biases by sidelining women from hiring processes, penalising loan applicants from poor neighbourhoods, or through welfare systems that wrongly deny citizens their benefits. But the law met resistance from tech companies who threatened to pull out form the state, claiming it is too broad in scope in its current form and would stifle innovation. This brings critical questions about AI regulation to the forefront:
- Who should be responsible when AI causes harm? Developers, deployers, or both?
- How should citizens seek justice?
- How can tech companies be incentivised to develop safe technologies?
Colorado’s governor has called a special session to update the law before it kicks in.
What This Means for India
India is on its path towards framing a dedicated AI-specific law or directions, and discussions are underway through the IndiaAI Mission, the proposed Digital India Act, committee set by the Delhi High Court on deepfake and other measures. But the dilemmas Colorado is wrestling with are also relevant here.
- AI uptake is growing in public service delivery in India. Facial recognition systems are expanding in policing, despite accuracy and privacy concerns. Fintech apps using AI-driven credit scoring raise questions of fairness and transparency.
- Accountability is unclear. If an Indian AI-powered health app gives faulty advice, who should be liable- the global developer, the Indian startup deploying it, or the regulator who failed to set safeguards?
- India has more than 1,500 AI startups (NASSCOM), which, like Colorado’s firms, fear that onerous compliance could choke growth. But weak guardrails could undermine public trust in AI altogether.
Lessons for India
India’s Ministry of Electronics and IT ( MEITy) favours a light-touch approach to AI regulation, and exploring and advancing ways for a future-proof guideline. Further, lessons from other global frameworks can guide its way.
- Colorado’s case shows us the necessity of incorporating feedback loops in the policy-making process. India should utilise regulatory sandboxes and open, transparent consultation processes before locking in rigid rules.
- It will also need to explore proportionate obligations, lighter for low-risk applications and stricter for high-risk use cases such as policing, healthcare, or welfare delivery.
- Europe’s AI Act is heavy on compliance, the U.S. federal government leans toward deregulation, and Colorado is somewhere in between. India has the chance to create a middle path, grounded in our democratic and developmental context.
Conclusion
As AI becomes increasingly embedded in hiring, banking, education, and welfare, opportunities for ordinary Indians are being redefined. To shape how this pans out, states like Tamil Nadu and Telangana have taken early steps to frame AI policies. Lessons will emerge from their initiative in addressing AI governance. Policy and regulation will always be contested, but contestations are a part of the process.
The Colorado debate shows us how participative law-making, with room for debate, revision, and iteration, is not a weakness but a necessity. For India’s emerging AI governance landscape, the challenge will be to embrace this process while ensuring that citizen rights and inclusion are balanced well with industry concerns. CyberPeace advocates for responsible AI regulation that balances innovation and accountability.
References
- https://www.cbsnews.com/colorado/news/colorado-lawmakers-look-repeal-replace-controversial-artificial-intelligence-law/
- https://www.naag.org/attorney-general-journal/a-deep-dive-into-colorados-artificial-intelligence-act/
- https://carnegieendowment.org/research/2024/11/indias-advance-on-ai-regulation?lang=en
- https://the-captable.com/2024/12/india-ai-regulation-light-touch/
- https://indiaai.gov.in/article/tamilnadu-s-ai-policy-six-step-tamdef-guidance-framework-and-deepmax-scorecard
Executive Summary:
A video is circulating on social media claiming to be footage of the aftermath of Iran's missile strikes on Israel. The video shows destruction, damaged infrastructure, and panic among civilian casualties. After our own digital verification, visual inspection, and frame-by-frame inspection, we have determined that the video is fake. The video is just AI-generated clips and not related to any incident.
Claim:
The viral video claims that a recent military strike by Iran resulted in the destruction of parts of Israel, following an initial missile attack launched by Iran. The footage appears current and depicts significant destruction of buildings and widespread chaos in the streets.
FACT CHECK:
We conducted our research on the viral video to determine if it was AI-generated. During the research we broke the video into individual still frames, and upon closely examining the frames, several of the visuals he showed us had odd-shaped visual features, abnormal body proportions, and flickering movements that don't occur in real footage. We took several still frames and checked them in image search sites to see if they had appeared before. The search results revealed that several clips in the video had appeared previously, in separate and unrelated circumstances, which indicates that they are neither recent nor original.
While examining the Instagram profile, we noticed that the account frequently shares visually dramatic AI content that appears digitally created. Many earlier posts from the same page include scenes that are unrealistic, such as wrecked aircraft in desolate areas or buildings collapsing in unnatural ways. In the current video, for instance, the fighter jets shown have multiple wings, which is not technically or aerodynamically possible in real life. The profile’s bio, which reads "Resistance of Artificial Intelligence," suggests that the page intentionally focuses on sharing AI-generated or fictional content.
We also ran the viral post through Tenorshare.AI for Deep-Fake detection, and the result came 94% AI. All findings resulting from our research established that the video is synthetic and unrelated to any event occurring in Israel, and therefore debunked a false narrative propagated on social media.
Conclusion:
Our research found that the video is fake and contains AI-generated images and is not related to any real missile strike or destruction occurring in Israel. The source is specific to fuel the panic and misinformation in a context of already-heightened geopolitical tension. We call on viewers not to share this unverified information and to rely on trusted sources. When there are sensitive international developments, the dissemination of fake imagery can promote fear, confusion, and misinformation on a global scale.
- Claim: Real Footage of Iran’s Missile Strikes on Israel
- Claimed On: Social Media
- Fact Check: False and Misleading
Introduction
The United Nations General Assembly (UNGA) has unanimously adopted the first global resolution on Artificial Intelligence (AI), encouraging countries to take into consideration human rights, keeping personal data safe, and further monitoring the threats associated with AI. This non-binding resolution proposed by the United States and co-sponsored by China and over 120 other nations advocates the strengthening of privacy policies. This step is crucial for governments across the world to shape how AI grows because of the dangers it carries that could undermine the protection, promotion, and right to human dignity and fundamental freedoms. The resolution emphasizes the importance of respecting human rights and fundamental freedoms throughout the life cycle of AI systems, highlighting the benefits of digital transformation and safe AI systems.
Key highlights
● This is indeed a landmark move by the UNGA, which adopted the first global resolution on AI. This resolution encourages member countries to safeguard human rights, protect personal data, and monitor AI for risks.
● Global leaders have shown their consensus for safe, secure, trustworthy AI systems that advance sustainable development and respect fundamental freedom.
● Resolution is the latest in a series of initiatives by governments around the world to shape AI. Therefore, AI will have to be created and deployed through the lens of humanity and dignity, Safety and Security, human rights and fundamental freedoms throughout the life cycle of AI systems.
● UN resolution encourages global cooperation, warns against improper AI use, and emphasizes the issues of human rights.
● The resolution aims to protect from potential harm and ensure that everyone can enjoy its benefits. The United States has worked with over 120 countries at the United Nations, including Russia, China, and Cuba, to negotiate the text of the resolution adopted.
Brief Analysis
AI has become increasingly prevalent in recent years, with chatbots such as the Chat GPT taking the world by storm. AI has been steadily attempting to replicate human-like thinking and solve problems. Furthermore, machine learning, a key aspect of AI, involves learning from experience and identifying patterns to solve problems autonomously. The contemporary emergence of AI has, however, raised questions about its ethical implications, potential negative impact on society, and whether it is too late to control it.
While AI is capable of solving problems quickly and performing various tasks with ease, it also has its own set of problems. As AI continues to grow, global leaders have called for regulations to prevent significant harm due to the unregulated AI landscape to the world and encourage the use of trustworthy AI. The European Union (EU) has come up with an AI act called the “European AI Act”. Recently, a Senate bill called “The AI Consent Bill” was introduced in the US. Similarly, India is also proactively working towards setting the stage for a more regulated Al landscape by fostering dialogues and taking significant measures. Recently, the Ministry of Electronics and Information Technology (MeitY) issued an advisory on AI, which requires explicit permission to deploy under-testing or unreliable AI models related to India's Internet. The following advisory also indicates measures advocating to combat deepfakes or misinformation.
AI has thus become a powerful tool that has raised concerns about its ethical implications and the potential negative influence on society. Governments worldwide are taking action to regulate AI and ensure that it remains safe and effective. Now, the groundbreaking move of the UNGA, which adopted the global resolution on AI, with the support of all 193 U.N. member nations, shows the true potential of efforts by countries to regulate AI and promote safe and responsible use globally.
New AI tools have emerged in the public sphere, which may threaten humanity in an unexpected direction. AI is able to learn by itself through machine learning to improve itself, and developers often are surprised by the emergent abilities and qualities of these tools. The ability to manipulate and generate language, whether with words, images, or sounds, is the most important aspect of the current phase of the ongoing AI Revolution. In the future, AI can have several implications. Hence, it is high time to regulate AI and promote the safe, secure and responsible use of it.
Conclusion
The UNGA has approved its global resolution on AI, marking significant progress towards creating global standards for the responsible development and employment of AI. The resolution underscores the critical need to protect human rights, safeguard personal data, and closely monitor AI technologies for potential hazards. It calls for more robust privacy regulations and recognises the dangers associated with improper AI systems. This profound resolution reflects a unified stance among UN member countries on overseeing AI to prevent possible negative effects and promote safe, secure and trustworthy AI.
References
