The Right to Know, The Duty to Label: India’s AI Governance Framework and the Constitutionalisation of Synthetic Media
The digital public sphere is no longer a space of authentic human expression alone. It has become a hybrid environment, in which content generated by algorithms increasingly coexists with—and in some cases, becomes indistinguishable from—content created by human beings. This transformation, powered by advances in generative artificial intelligence, presents a profound challenge to the foundations of democratic discourse. When citizens cannot reliably distinguish between authentic communication and synthetic fabrication, the very concept of informed consent, deliberative debate, and electoral accountability is threatened at its core.
India’s response to this challenge, articulated in the accompanying essay by S. Krishnan, Secretary of the Ministry of Electronics and Information Technology, is not merely a regulatory intervention; it is a constitutional claim. It asserts that the right to know the origin of the information we consume is not a luxury of the literate or a privilege of the technologically sophisticated. It is a fundamental entitlement of democratic citizenship, as essential to the exercise of informed choice as the right to vote or the freedom of speech.
The amended Information Technology (Intermediary Guidelines and Digital Media Ethics Code) Rules, 2021, read alongside the India AI Governance Guidelines, 2025, operationalise this claim through a carefully calibrated architecture of definitional precision, ex ante safeguards, mandatory transparency, and tiered accountability. This is not a reactionary crackdown on technological innovation or a symbolic gesture toward anxious voters. It is a systematic, principled, and globally relevant framework for governing the most disruptive communications technology since the invention of the printing press.
The framework’s central innovation is the introduction of a precise, operational definition of “synthetic material generated from information.” This definition captures content that is “artificially or algorithmically created or altered in a manner that appears authentic or indistinguishable from reality,” while explicitly excluding routine, good-faith activities such as editing, alteration for accessibility, educational and research materials, and legitimate creative expression. This is not semantic nitpicking; it is the foundation of legal certainty. Without a clear definition of what constitutes regulated synthetic content, neither intermediaries nor citizens can know their rights and obligations. With it, the framework establishes a common language for adjudicating disputes, assigning responsibility, and protecting fundamental rights.
The framework’s second pillar is its shift from reactive moderation to ex ante governance. Previous approaches to harmful online content operated primarily after the fact: content was published, complaints were filed, platforms were notified, and takedowns were ordered. This reactive model, however well-intentioned, was structurally incapable of addressing the scale and speed of synthetic media dissemination. By the time harmful content was identified and removed, it had often already been viewed, shared, and internalised by millions. The damage, once done, could not be undone.
The amended rules break decisively with this reactive paradigm. They require intermediaries that enable or facilitate the creation or dissemination of synthetic content to deploy reasonable and appropriate technical measures to prevent the circulation of unlawful content at the point of creation or dissemination. This is not a mandate for pre-publication censorship or a requirement to monitor all user content. It is a proportional, risk-based obligation to integrate safeguards into the design of platforms and tools, rather than relying exclusively on after-the-fact enforcement.
For lawful synthetic content, the framework mandates clear and prominent labelling, supported by permissions and transparency mechanisms. This is not a restriction on expression; it is an enhancement of the informational environment. Citizens are empowered to distinguish authentic content from synthetic fabrication in real time, exercising their own judgement about the credibility and trustworthiness of what they see, hear, and read. The labelling requirement operationalises the right to know—a constitutional value that the framework explicitly recognises and protects.
The Architecture of Accountability: Tiered Obligations and Systemic Responsibility
The amended Intermediary Guidelines recognise that not all platforms are created equal. A small, niche community forum that occasionally hosts user-generated content does not pose the same systemic risks as a major social media platform with hundreds of millions of users and algorithmic amplification of viral content. The framework’s tiered approach to intermediary obligations reflects this reality.
Significant social media intermediaries—platforms whose scale and influence give them outsized impact on the digital public sphere—are subject to additional, commensurate obligations. They are required to obtain user declarations regarding synthetic content, to deploy proportionate technical measures to assess the accuracy of those declarations, and to ensure that synthetic material is not published without appropriate identification. Failure to comply with these requirements may constitute a false cause of action with attendant statutory consequences.
This tiered architecture is not arbitrary; it is constitutionally grounded. The framework recognises that platforms with greater systemic impact must shoulder higher governance obligations. This is not discrimination against successful businesses; it is the regulatory corollary of the principle that with great power comes great responsibility. A platform that shapes the information environment for half a billion citizens cannot claim the same privileges of non-interference as a neighbourhood bulletin board.
The framework also clarifies the legal status of good-faith moderation actions. Intermediaries that deploy automated tools or other reasonable technical measures to identify and label synthetic content do not thereby forfeit their statutory protections. This clarification is essential to creating a compliance-enabling environment. Without it, platforms would face a perverse incentive: the more they invested in responsible content governance, the greater their exposure to liability for erroneous or overbroad enforcement. The framework resolves this dilemma by affirming that reasonable, good-faith efforts to comply with regulatory obligations do not undermine statutory protections.
The Constitutional Foundation: Rights, Dignity, and Democratic Stewardship
The AI Governance Guidelines, 2025, articulate the normative framework that animates India’s approach to synthetic media regulation. They emphasise transparency, accountability, human-centric design, and risk awareness as the guiding principles for responsible AI adoption. These are not merely technical recommendations; they are constitutional values translated into regulatory practice.
The guidelines explicitly operate within the boundaries of existing law. They do not create new statutory obligations or substitute for the enforceable legal rules contained in the IT Act and the Intermediary Guidelines. Rather, they provide direction and guidance to developers, deployers, and institutions, reinforcing the expectation that AI systems—particularly those capable of generating synthetic content—must be designed and deployed with due regard to societal impact and public trust.
The essay’s invocation of Parliament’s legislative authority to regulate the use of AI in the public interest is not a assertion of regulatory excess; it is a reaffirmation of democratic stewardship. The framework explicitly disclaims any claim to regulatory finality. It is not a permanent, immutable settlement of the governance challenges posed by synthetic media. It is an adaptive, evolving response to a rapidly changing technological landscape, grounded in constitutional principles and open to revision as experience accumulates and capabilities advance.
This posture of regulatory readiness—neither reactive nor pre-emptive, neither laissez-faire nor authoritarian—is the framework’s most distinctive and valuable feature. It acknowledges that we are in the early stages of understanding the societal implications of generative AI. It recognises that today’s solutions may be inadequate for tomorrow’s problems. It insists, nevertheless, that the absence of perfect knowledge cannot excuse the absence of principled action. We must govern what we can, with the tools we have, while remaining open to learning and adaptation.
The Global Relevance: India’s Model for Democratic AI Governance
Krishnan’s concluding claim—that India’s response to synthetic media offers “not only a domestic governance solution, but a globally relevant model for democratic, rights-respecting regulation in an AI-mediated world”—is not nationalistic boosterism. It is a substantive argument about the comparative merits of different regulatory approaches.
The European Union’s AI Act, for all its comprehensiveness, is a prescriptive, rules-based framework that imposes detailed, legally binding obligations on a wide range of AI systems. It is strong on legal certainty but weak on adaptability; its detailed provisions may prove difficult to amend as technology evolves. China’s approach to AI governance is authoritarian and control-oriented, prioritising state security and social stability over individual rights and democratic accountability. It is effective at suppressing dissent but incapable of fostering the kind of open, pluralistic public sphere that democratic societies require.
India’s framework occupies a distinctive third space. It is legally binding where necessary—on labelling, takedowns, and intermediary obligations—but flexibly guided where appropriate. It is grounded in constitutional rights and democratic values, not in state security or bureaucratic convenience. It is designed to be adaptive and proportionate, not rigid and exhaustive. It recognises the legitimate interests of technology developers and platforms while insisting on the primacy of citizen welfare and public trust.
This is not to claim that India’s framework is perfect or that it will succeed where other approaches have failed. The governance of synthetic media is a wicked problem, for which no fully satisfactory solution exists or may ever exist. The technology is advancing too rapidly; the social consequences are too complex; the trade-offs between competing values are too painful. Any regulatory framework will be incomplete, contested, and subject to unforeseen consequences.
But India’s framework has the singular virtue of asking the right questions. How do we preserve the right to know in an environment where seeing is no longer believing? How do we hold platforms accountable for systemic harms without imposing disproportionate burdens on innovation? How do we govern technologies that we do not yet fully understand? These are not questions that admit of definitive answers. But they are questions that must be asked, and India’s framework provides a rigorous, principled, and transparent process for asking them.
Conclusion: Trust, Technology, and the Democratic Project
The challenge posed by synthetic media is, at its core, a challenge of trust. Trust that the information we consume accurately represents the intentions and beliefs of its creators. Trust that the platforms we use will act as responsible stewards of the public sphere, not as indifferent conduits for manipulation and fraud. Trust that the institutions charged with governing technology will respond effectively and proportionately to citizen harm, not with paralysis or overreach.
Trust cannot be mandated by law or engineered into code. It must be earned through consistent, transparent, and accountable practice. The regulatory framework that Krishnan describes creates the conditions under which such practice can emerge. It establishes clear rules and expectations. It empowers citizens with information and remedies. It holds powerful intermediaries accountable for systemic impacts. It provides guidance and direction for responsible innovation.
But the framework itself is not the solution; it is the scaffolding within which solutions can be built. The actual work of building trust must be performed by the thousands of engineers, product managers, policy professionals, and compliance officers who design, deploy, and govern the platforms and tools that constitute our digital public sphere. It must be performed by the millions of citizens who exercise their right to know, who demand transparency and accountability from the platforms they use, who refuse to accept the erosion of their informational environment as an inevitable cost of technological progress.
India’s framework provides the scaffolding. The rest is up to us.
Q&A Section
Q1: What is the operational definition of “synthetic material generated from information” introduced in the amended Intermediary Rules, and why is this definitional clarity legally significant?
A1: The amended Rules define synthetic material as content that is “artificially or algorithmically created or altered in a manner that appears authentic or indistinguishable from reality.” This definition is carefully calibrated to exclude routine, good-faith activities such as editing, alteration for accessibility, educational and research materials, and legitimate creative expression. The definitional clarity is legally significant for several reasons. First, it provides a clear, operational standard for determining what content falls within the regulatory framework, enabling both intermediaries and citizens to understand their rights and obligations. Second, by expressly bringing synthetic content within the scope of “information” under the IT Act, the Rules ensure that harms associated with such content are addressed through formal legal processes rather than left to informal, inconsistent platform moderation practices. Third, the exclusion of legitimate activities from the definition prevents overbroad application that could chill innovation, creativity, and accessibility. This is not semantic nitpicking but the foundation of legal certainty. Without a clear definition, neither platforms nor citizens can know what is required of them; with it, the framework establishes a common language for adjudicating disputes, assigning responsibility, and protecting fundamental rights.
Q2: What distinguishes the amended framework’s approach of “ex ante governance” from traditional reactive content moderation, and what obligations does it impose on intermediaries?
A2: Traditional reactive content moderation operates after the fact: content is published, complaints are filed, platforms are notified, and takedowns are ordered. This model is structurally incapable of addressing the scale and speed of synthetic media dissemination; by the time harmful content is removed, it has often already been viewed and internalised by millions. Ex ante governance shifts the regulatory focus upstream, requiring intermediaries to deploy preventive measures at the point of creation or dissemination. The amended Rules require intermediaries that enable or facilitate synthetic content to deploy reasonable and appropriate technical measures to prevent the circulation of unlawful content before it reaches users. This is not a mandate for pre-publication censorship or universal monitoring; it is a proportional, risk-based obligation to integrate safeguards into platform design. For lawful synthetic content, the framework mandates clear and prominent labelling supported by permissions and transparency mechanisms. This ex ante approach recognises that when damage can be instantaneous and irreversible, prevention is not merely preferable to cure—it is the only viable strategy for protecting the informational environment.
Q3: How does the framework’s tiered approach to intermediary obligations reflect constitutional principles of proportionality and democratic accountability?
A3: The framework imposes additional, commensurate obligations on “significant social media intermediaries”—platforms whose scale and influence give them outsized impact on the digital public sphere. These platforms are required to obtain user declarations regarding synthetic content, deploy proportionate technical measures to assess declaration accuracy, and ensure synthetic material is not published without appropriate identification. Failure to comply may constitute a false cause of action with statutory consequences. This tiered architecture is not arbitrary discrimination but the regulatory corollary of the principle that with great power comes great responsibility. A platform that shapes the information environment for half a billion citizens cannot claim the same privileges of non-interference as a neighbourhood bulletin board. The framework recognises that systemic impact warrants systemic accountability. This is constitutionally grounded in the doctrine of reasonable classification: platforms are not being treated differently because of their success but because their success creates qualitatively different risks and responsibilities. The tiered approach is thus not punitive but proportional—calibrated to the scale of potential harm rather than imposed uniformly regardless of context.
Q4: What is the relationship between the legally binding Intermediary Rules and the non-binding India AI Governance Guidelines, and why is this distinction important?
A4: The Intermediary Rules are enforceable legal obligations under the IT Act, carrying statutory consequences for non-compliance. The AI Governance Guidelines are policy guidance that articulate normative principles—transparency, accountability, human-centric design, risk awareness—without creating new legal obligations. The Guidelines explicitly operate “within the boundaries of existing law” and do not substitute for the binding rules. This distinction is important for several reasons. First, it preserves legal certainty: regulated entities know exactly which requirements are mandatory and which are advisory. Second, it enables adaptive governance: the Guidelines can be updated more easily than statutory rules as technology evolves and experience accumulates. Third, it reflects a sophisticated understanding of regulatory pluralism: different policy problems require different governance instruments. Binding rules are appropriate for establishing minimum standards of conduct and creating enforceable rights for citizens. Non-binding guidance is appropriate for shaping industry practice, disseminating best practices, and building a culture of responsible innovation. The framework thus employs a calibrated, multi-instrument approach rather than relying exclusively on command-and-control regulation. This is not regulatory weakness but regulatory sophistication.
Q5: What justifies Krishnan’s claim that India’s framework offers a “globally relevant model” for democratic AI governance, and how does it differ from the EU and Chinese approaches?
A5: Krishnan’s claim rests on the framework’s distinctive combination of features that position it as a third way between existing regulatory paradigms. The EU’s AI Act is a prescriptive, rules-based framework that imposes detailed, legally binding obligations on a wide range of AI systems. It provides strong legal certainty but is relatively rigid; its detailed provisions may prove difficult to amend as technology rapidly evolves. China’s approach is authoritarian and control-oriented, prioritising state security and social stability over individual rights and democratic accountability. It effectively suppresses dissent but is incapable of fostering the open, pluralistic public sphere that democratic societies require.
India’s framework occupies a distinctive third space. It is legally binding where necessary (labelling, takedowns, intermediary obligations) but flexibly guided where appropriate. It is grounded in constitutional rights and democratic values, not state security or bureaucratic convenience. It is designed to be adaptive and proportionate, not rigid and exhaustive. It recognises the legitimate interests of technology developers and platforms while insisting on the primacy of citizen welfare and public trust. It is, in short, a democratic, rights-respecting, innovation-friendly approach to governing one of the most disruptive technologies in human history. Whether it succeeds in practice remains to be demonstrated, but its conceptual architecture offers a compelling model for other democracies grappling with the same challenges.
