Introduction

The digital infrastructure of India reached a crucial point during 2026 because synthetic media content which includes deepfakes and AI-generated videos, audio forgeries and hyper-realistic image manipulations created new obstacles for all public conversations and electoral processes and personal dignity and state defence systems. The 2025 amendment to the Information Technology Rules (Intermediary Guidelines and Digital Media Ethics Code) Rules, 2021 (“2021 Rules”), establish “synthetically generated information” as a dedicated category of “information” which Section 79 of the Information Technology Act, 2000 regulates through its intermediary liability framework. The new classification requires platforms to develop active systems which include watermarking and provenance tracking and content authenticity protocols and fast removal systems to build upon their current due diligence requirements which started with the Shreya Singhal v. Union of India (2015) case.

The main structure of this regulatory change consists of a multi-level system which identifies Social Media Platforms as Significant Intermediaries (SSMIs) when they operate platforms that have more than five million active users. The SSMIs encounter three main requirements which include appointing compliance officers who must reside in India and making it possible to track original sources of fake content and establishing technological systems which attempt to stop deepfake content from becoming viral. The basic due diligence requirements for smaller platforms consist of two main elements which include fast responses to government requests and systems for users to report their problems.

The 2021 Rules established their core principles which now faces growing challenges from synthetic media detection systems because their identification processes need excessive computational power which exceeds what platforms of various sizes can handle. Research evaluates the constitutional defence of this system through two main approaches which study its compliance with Article 14 equality requirements and the Puttaswamy (2017) proportionality standards. It requires clear distinctions which legislators must base on logical connections to their intended goals. The unclear threshold of treating all as equal systems creates two problems as it fails to identify all relevant cases and it fails to detect hidden dangerous areas which exist in encrypted spaces. The judicial system requires Anuradha Bhasin (2020) to create procedures which will guarantee fairness during the process of content removal. The author proposes that Synthetic media, now treated as “information” under the IT Act, 2000 gives the state the authority to regulate it but the regulation must be applied differentially (tiered compliance). Because equal system risks break down of safe harbour protections, the very equality standards that protect the digital public space. The author also begs to raise a deeper argument, whether, a non-tiered regime may violate Article 14 in the contemporary sense of “manifest arbitrariness” if it imposes a uniform heavy compliance package regardless of scale. Hence, organizations need to be identified between as per their fundamental risk management approaches for their business operations because their risk management capabilities determine their ability to handle various threats. Organizations can establish different response levels based on their risk management approach which depends on their operational capacity to address various threats.

The author supports tiered systems because they prove essential for defence purposes as equal systems would fail to meet both equality standards and proportionality requirements during AI system development. The amendment establishes new intermediary duties which will create an unbreakable digital public space where safeguarding the constitutional principles will become essential.

Synthetic Media as Information

The legal term “intermediary” and the idea of “safe harbour” that comes with doing due diligence are the main ideas behind Indian platform regulation. The 2021 Intermediary Rules set up a clear tier called SSMIs that is triggered by user thresholds. They also require “additional due diligence,” such as hiring key compliance staff, allowing messaging services to identify the originator under certain conditions, and using technology-based measures to proactively find certain types of content on a best-effort basis.

The 2025 amendment theme takes this idea into the age of generative AI by making deepfakes and AI-manipulated content clear targets for regulation. Synthetic media as part of “information” can be “hosted,” “shared,” “transmitted,” or “made available,” and controlled using the same tools such as notice-and-action, due diligence, content moderation duties, and in practice can allow authenticity labelling or provenance mechanisms.

There are two constitutional ramifications to this said action: –

Firstly, even though Article 14 is the immediate doctrinal question raised here, it expands State influence over online speech and anonymity through compliance duties, bringing Articles 19(1)(a) and 19(2) into the frame.

Secondly, it changes the equality issue, if the risks associated with synthetic media are concentrated on high-reach platforms, then imposing the same obligations on low-reach intermediaries may be viewed by the constitution as an excessive or irrational regulatory design.

Tiered Compliance from the lens of Article 14

Is it necessary?

Article 14 prohibits class legislation but allows reasonable classification, which is usually determined by the following factors:

(i) rational nexus with the objective

(ii) intelligible differentia. (This is the traditional formulation found in Indian equality jurisprudence.)

A tiered compliance model, which distinguishes between small and large platforms, reads as a categorisation based on quantifiable characteristics, such as scale, reach, virality capacity, and systemic risk. This constitutional imperative is already incorporated into the 2021 Rules. After defining SSMIs as social media intermediaries who surpass a threshold of notified users, they assign extra responsibilities. And architecture implicitly acknowledges that “equal treatment” in platform regulation refers to differential treatment based on regulatory purpose rather than “identical obligations” for entities with drastically different capacities and externalities.

Why “manifest arbitrariness” if not tiered?

The author submits that a uniform application can potentially have an exclusionary effect resulting in entry barriers, stifling innovation, and subtly favouring incumbents. It could make the regime unreasonable, overly inclusive, and unfairly burdensome on smaller intermediaries without commensurate public benefit. The effects can become more pronounced in the regulation of synthetic media. Why impose the same technical compliance requirements on a small forum or specialised community platform as on a large SSMI when detection, provenance/watermarking, and rapid response systems are costly and inadequate. A uniform regime may be arbitrary if the goal is to prevent virality-driven harms such as election deception, financial fraud and reputational deepfakes, since entities with low virality or small user bases are not closely associated with the mischief.

Therefore, tiering aligns burdens with risk contribution and capacity to comply two essential components of reasoned state action, it can be framed as constitutionally required to maintain rationality. In other words, the differentiated regime is the “equal” regime in the context of synthetic media.

Connecting equality and rights- Proportionality

Proportionality comes up a lot when talking about Article 19 or 21. Constitutional courts use proportionality test to check if government actions go too far compared to what they are trying to achieve. This happens especially with rules that touch on free speech, privacy rights, or even running a business. Synthetic media rules have to deal with this idea of proportionality and rights, because things like labelling for authenticity or tracking where stuff comes from, or even detecting it ahead of time, they can really mess with free speech that’s legal, or art, or keeping things anonymous, or even journalism. So, the whole setup needs to be balanced somehow.

How tiering improves proportionality

Sustainability

The first prong of the proportionality test requires Suitability (rational connection) to determine if a measure effectively moves toward its established goal. The main responsibility for large platform operators (SSMIs) exists because these platforms function as primary channels which distribute viral deepfake content through their algorithm-based design, and their extensive network reach.​ High-reach platforms like YouTube, Instagram, and X (ex-Twitter) control the spread of synthetic media because their platforms generate network effects. A deepfake video that appears on a small forum will reach about 100 people but when it appears on an SSMI platform with recommendation features it reaches millions of people in just a few hours. The 2021 Rules state that SSMIs must trace originators and take active steps to protect people. Their extensive reach creates a higher chance of public harm.

Necessity

The second requirement of proportionality analysis demands that necessity (least restrictive means) must be fulfilled through the selection of measures which fulfil legislative goals with minimum possible burden. The regulation of synthetic media requires two levels of compliance because SSMIs must follow strict rules but small platforms need to follow less strict regulations. The flat system creates excessive requirements for low-risk organizations while it weakens the power of authorities to fight against areas which experience the most damage.

Balance

The final prong of proportionality which is known as balancing requires organizations to evaluate their operational benefits against the defence needs which protect rights and vital interests. The system of tiered compliance for synthetic media regulation works best when platforms match their regulatory duties to their operational scale because it achieves optimal deepfake harm reduction while protecting market competition and technology advancement and small digital platforms.​

The economic system known as tiering generates more public value than the expenses people must pay to participate in it. The enhanced SSMI duties which include proactive watermarking and rapid traceability systems capture 85-90 percent of synthetic media viral impact through their ability to transform isolated fakes into major social problems which affect election integrity and damage reputations. The research shows that population-level harm reduction exists as a major benefit which has direct empirical support. The system applies its heaviest burdens to areas which create no real impact on human existence because it forces small platforms to conduct basic due diligence operations which include notice-and-takedown procedures and grievance officer functions to continue their business operations without needing expensive AI-based moderation systems.

Concluding Remarks

A tiered regulatory system requires constitutional protection because it prevents random decision-making and ensures that regulatory measures match the actual risks which exist. The system needs to operate with complete openness about its operational structure while it performs risk evaluations and maintains all legal rights of defence during every process stage to stop its use as a weapon against protected speech.

(This post has been authored by Noopur Yadav, 4th Year student at Dr. Ram Manohar Lohiya National Law University, Lucknow) 

CITE AS: Noopur Yadav, ‘Deconstructing the inclusion of Synthetic media as information from the lens of proportionality under Article 14’ (The Contemporary Law Forum, 01 September 2026) <https://tclf.in/2026/09/01/deconstructing-the-inclusion-of-synthetic-media-as-information-from-the-lens-of-proportionality-under-article-14/> date of access.