The Delhi High Court (“DHC”), in the past month, has seen multiple ongoing cases where social media intermediaries’ copyright strike mechanism has come into question (see here and here). With social media platforms now serving as a principal forum for public expression, and a handful of platforms controlling access to much of this speech, the consequences of automated content removal extend beyond private copyright disputes. Under such mechanisms, social media intermediaries automatically take down allegedly infringing content merely based on content similarity without application of mind or considering whether such use falls under the exceptions to copyright infringement. Such constant copyright censorship raises serious free speech questions. (Bindushree M and Shubham Thakare have previously examined this issue in their respective posts here and here). In this post, given the grave implications on free speech, I propose recommendations in the form of good-faith undertakings and damages as a potential solution to ensure such a take-down mechanism is not abused. I’d love to hear feedback and alternatives from readers on this.
What is the problem?
Before going into the problem, I will very briefly explain how such takedowns work. As a platform, copyright-related takedowns on Youtube are a common occurrence. Apart from rights-holders independently reporting specific content, Youtube also offers a feature called Content ID, which creates a unique digital fingerprint from the works rights-holders have submitted. This fingerprint is tracked against other content to check similarity, once established, depending on the rights-holders setting, it can automatically block the allegedly infringing video, monetize it in favour of the rights-holder, or track the video’s viewership statistics. Meta’s Rights Manager and its matching technology works in a similar manner. It in fact notes that, in limited instances “rights holders may automatically block matching content” though with certain qualifications protecting lawful or expressive content without detailing how it works. Thus, the current system not only ignores the risk of rights-holders abusing the takedown system but also provides automatic blocking of allegedly infringing content.
If any individual is suing for copyright infringement, they must establish that one of their exclusive rights is being infringed AND that it does not fall within the exceptions provided in the Copyright Act [most recently recognised in ANI v OpenAI (2026)].
Who is looking at whether or not allegedly infringing content falls under the exceptions in such automatic blocking mechanisms? Surprise surprise, nobody. The onus is now pushed upon the user to dispute such claims and prove that their use is permitted under the law. As Rishubh Agrawal notes, such automatic takedowns are largely abused, which involves weaponization of copyright by various actors, including government agencies!
Here is what is happening – large rights-holders prefer automation to save costs and ensure effective enforcement while platforms enable such mechanisms to possibly avoid intermediary liability. But where are the users’ free speech interests incorporated? Effective enforcement of a statutory right for commercial gains of the rights-holders is being deployed at the cost of the fundamental right of free speech.
Every use of a copyrighted work is not infringing. A finding of infringement requires that a particular use does not fall under Section 52 of the Copyright Act. The said section is not merely an exception, it has been interpreted broadly in conjunction with Article 19(1)(a) [Wiley Eastern v IIM (1995)], and has most recently been recognised as a user right/privilege [ANI v OpenAI (2026)]. Thus, rights-holders have automated platform-enabled enforcement without application of mind based merely on similarity at cost of the fundamental right of free speech and expression while conveniently shifting the onus on individuals whose free speech rights are infringed.
The problem is not just restricted to free speech concerns. Youtube notes that in 2025 that 90% Content ID claims were to monetize the views on the video (i.e, advertisement revenue from the allegedly infringing videos are diverted from the uploader and flow to the claimant/rights-holder). Thus, even where speech is not suppressed, revenue may be diverted based on an unproven claim of infringement. This is particularly problematic because commercial use does not, by itself, exclude a use from Section 52, except where the provision expressly says so. [ANI v OpenAI (2026), Super Cassettes v Hamar Televisions (2010)].
Good-faith Undertaking and Damages as the Solution – Expanding the Scope of Section 60
The takedown mechanism has its root in the US’ Digital Millennium Copyright Act (“DMCA”). Importantly, this statute also requires takedown notifications to contain a good-faith representation made by the complainant that “use of the material in the manner complained of is not authorized by the copyright owner, its agent, or the law” [17 USC § 512(c)(3)(A)(v)]. The part dealing with ‘authorized by the law’, in the aforementioned provision, was interpreted to include fair use in Lenz v. Universal Music, thus “a copyright holder must consider the existence of fair use before sending a takedown notification”.
The DMCA also provides for claiming damages if a claimant misrepresents that some alleged material/activity is infringing [17 USC § 512(f)]. Regarding such misrepresentation, the aforementioned Court observed the following: “To be clear, if a copyright holder ignores or neglects our unequivocal holding that it must consider fair use before sending a takedown notification, it is liable for damages under § 512(f).”
Coming back to India, it is the need of the hour to incorporate similar mechanisms in India where every takedown notification issued by a rights-holder is accompanied by a good-faith undertaking that such alleged infringing use is not authorized by the law which would include Section 52 (exceptions) of the Copyright Act. Any misrepresentation or failure to consider Section 52 should result in the alleged infringer having a right to sue the rights-holder for damages.
The hook to recover such damages could be Section 60 of the Copyright Act. Here is a reproduction of Section 60 (Text version for accessibility: link)
The obvious push back here would be, whether copyright strike notices on social media platforms would qualify as “legal proceedings or liability” as per the text of Section 60? The DHC in a Section 60 application observed that “the issuance of Strike Notices to the Plaintiff, coupled with the absence of ‘any action’ as envisaged under Section 60 of the Act, renders the threats actionable within the meaning of Section 60 of the Act.” (Associated Broadcasting Company v. Google (2026), Subham has analysed this case here, this view was reiterated by the DHC in May 2026 in Avaia Ventures v Wildship Enterprises).
The damages recoverable under Section 60 are framed as “damages, if any, as he has sustained by reason of such threats.” The ambit of this provision should not be confined to legal costs incurred in responding to such threats, but should extend to the harm caused by the wrongful curtailment of the alleged infringer’s freedom of speech and expression, particularly where the claimant has failed to consider whether the impugned use was legally permissible. Since quantifying the financial loss resulting from the curtailment of lawful expression may be difficult, such harm could be recognised through an award of notional damages.
However, a challenge is presented by the proviso of Section 60 which exempts the application of the section if “the person making such threats, with due diligence, commences and prosecutes an action for infringement of the copyright claimed by him”. Wealthy rights-holders could easily file frivolous lawsuits to avoid liability under this section. However, the Bombay High Court has interpreted this proviso in a strict manner: “due diligence” would connote action initiated with good faith with care, caution and foresight, while “commences and prosecutes” means prosecution of the cause in the right earnest and does not include threats amounting to empty rhetoric [Manya Vejju v Sapna Bhog (2023), Anirud Raghav has previously analysed this case here]
Such a good-faith undertaking mechanism is not unknown to India. Rule 75 of the Copyright Rules 2013, provides the format of notification to be provided if a copyright owner is making a complaint under Section 52(1)(c) complaining that such transient or incidental storage is an infringement. The format under Rule 75(2)(c) explicitly provides that the written complaint should include details establishing that the work is an infringing copy AND “the allegedly infringing act is not covered under section 52 or any other act that is permitted by the Act”. This emphasis on ensuring that copyright exceptions are not overlooked is also reflected in the Copyright Act’s treatment of Technological Protection Measures. Section 65A(2)(a) expressly clarifies that the prohibition on circumventing such measures does not extend to acts undertaken for purposes “not expressly prohibited by this Act.”
Moreover, such a requirement also aligns with the text of Section 14 (exclusive rights of the copyright-holder) which notes that the exclusive rights are “subject to the provisions of this Act”, the provisions of the Act include exceptions to copyright (Section 52), thus, any assertion of infringement would necessarily entail ruling out the application of Section 52 [this argument aligns with the observations made in ANI v OpenAI (2026), Para 150-151].
Therefore, a good-faith undertaking must be mandated in takedown notifications, undertaking that such alleged infringing use is not authorized by the law. Non-consideration of the same should result in nominal damages, while the burden of proof regarding consideration of whether the alleged infringing use was permitted under the law should lie on the person claiming infringement.
Conclusion
The central problem with automated copyright takedowns is not the enforcement of copyright itself, but the absence of any meaningful consideration of whether the allegedly infringing use is legally permissible. By shifting the burden entirely onto users to contest automated claims, these mechanisms risk turning copyright enforcement into a tool of private censorship, particularly where the underlying use may be protected under Section 52 of the Copyright Act and Article 19(1)(a).
India should therefore move towards a system that requires rights-holders to make a good-faith representation that the complained-of use is not authorised by law, including under Section 52, before seeking a takedown. A necessary implication of this is minimization of automatic takedown, as such a representation requires considering whether the alleged infringing use is authorised by law. This should be accompanied by meaningful consequences, including notional damages, for knowing or reckless misrepresentation. Such a framework would not undermine legitimate copyright enforcement; rather, it would ensure that enforcement mechanisms respect the limits that copyright law itself places on the rights of copyright holders.
I would like to thank Swaraj Barooah for his valuable inputs and review of this post!
