A Conversation with Professor (Dr.) Arul George Scaria

In November 2024, the Delhi High Court appointed Professor (Dr.) Scaria as an amicus curiae in the landmark case of ANI Media Private Limited v. OpenAI. Following the court’s recent interim judgment, this discussion explores his experience serving as an amicus, the critical role academics play in assisting the judiciary in complex technology disputes, and the potential broader implications of this ruling for copyright law and the generative AI landscape in India. The written submissions of Professor (Dr.) Scaria in this case can be accessed from this link.

1. For readers who are generally unfamiliar with court procedure, could you explain the exact role of an amicus curiae in a case, how you came to be appointed in this particular matter, and what the process of working as an amicus typically looks like?

The term amicus curiae translates to ‘friend of the court’. When someone is appointed to this position, they are expected to assist the court in an impartial manner on different dimensions of the case. Different jurisdictions have distinct traditions regarding these appointments. In the Indian context, it is a court-appointed position. Unlike the United States, where individuals or organizations proactively file amicus briefs, in India, the court must formally invite and appoint someone as an amicus curiae.

Historically, Indian courts have been heavily inclined toward appointing senior practitioners as amici. There have been very few instances where academics have been appointed. This may stem from a perception issue – a belief that academics are detached from the practical realities of active litigation.

In the intellectual property (‘IP’) domain, academic appointments are exceptionally rare. The only notable example I am aware of is that of late Professor Shamnad Basheer, who acted as an intervener and contributed immensely in the landmark Novartis case. He had filed an application for intervention in this case, which the court admitted recognising the immense value of his academic insights. While there have been a few academic appointments in other domains like criminal or constitutional law, in general, it is far less common than it arguably should be.

In the case that we are discussing – ANI v. OpenAI – the court appointed two amici: one practitioner, Advocate Adarsh Ramanujan, and myself, in view of the complexity and novelty of the technological and legal issues involved. The court wanted us to make both written and oral submissions. Interestingly, the court wanted to first hear the submissions of the amici, before moving to the submissions of parties and intervenors. So, I made my opening submissions in person in the Court in February 2025 and in the next hearing, the Court heard the submissions of Adarsh Ramanujan. We then had the arguments from the parties and intervenors. There were more than 20 extensive hearing sessions in this case over a period of more than one year. I joined them online to ensure that I could also hear all the arguments of the parties and intervenors. I made my concluding oral submissions in March 2026 and subsequently also submitted my (concluding) written submissions. Overall, this was a very fascinating learning experience for me.

2. How does the role of an amicus curiae differ from that of an independent expert in a case?

Across jurisdictions and even within different courts of the same jurisdiction, the roles of independent experts can vary significantly. I previously served as an independent expert before the IP Division of the Delhi High Court. In that instance, I submitted a written report (available online). Because the parties settled the case shortly after I submitted the report, it remains an open question whether the court would have required oral submissions and how those boundaries would have been defined.

Generally, the role of an independent expert is governed by the specific rules of the court. The IP Division Rules of the Delhi High Court specifically covers such appointments, and as the first academic appointed under those rules, the process in that case was also somewhat experimental. We will need more of such appointments to see more clearly how the Indian judiciary may distinguish these two roles in practice.

3. In the ANI v. OpenAI case, the court appointed two amici: yourself and Advocate Adarsh Ramanujan. How do such joint appointments work in practice? Do you collaborate, and how did your diverse backgrounds shape your submissions?

Adarsh is an exceptional practitioner and a good friend. In this case, the court’s objective in appointing both of us was to capture a wide range of perspectives. To maintain the integrity of this objective, we formulated our submissions independently, although we occasionally engaged in discussions and debated our differing stances on various legal aspects.

This independence resulted in distinct and sometimes contrasting arguments on several key legal issues. In hindsight, this was a highly effective approach by the court. It allowed the court to evaluate a rich diversity of viewpoints, which is far more beneficial than relying on a single, uniform perspective.

4. The high number of interveners in this case was quite notable. Was your role as an amicus influenced by their presence, and more generally, how do you see the role of interveners shaping proceedings and final decisions in complex technology and intellectual property litigation?

As India’s first major copyright-related Generative AI  litigation, interest in this case was extraordinary. Although the immediate subject matter concerned news reports, industries across the board were deeply anxious about how a ruling on copyright infringement liability might impact their operations. This prompted a wave of intervention applications.

Ultimately, the court admitted six interveners. Roughly three of them supported the plaintiff’s position, and three aligned with the defendant’s position, and each presented nuanced arguments.

They all constructively engaged with the amici. They listened patiently to our oral and written submissions. The submissions of intervenors clearly highlighted how different sectors and participants frame and approach these emerging legal challenges.

5. The judgment spends considerable time detailing how large language models (LLMs) operate before analysing the legal issues. How important is it for courts to deeply engage with the technicalities of the underlying technology in these disputes, and do you foresee a growing role for academic or technical experts here?

Appointing academics is highly beneficial because it injects new perspectives into judicial proceedings. This is true irrespective of whether the case is deals with a new technology or deals with purely legal questions in different fields of law like constitutional law or criminal law.

In this specific litigation, the technical background was indispensable. The judge was very clear that he wanted to hear from the amici first to establish an understanding of how LLM technology functions before diving into the arguments of the parties. So during my opening remarks, substantial time was spent on detailing the mechanics of LLMs.

By the time we concluded the initial rounds of hearings, it was very evident that the court possessed a highly sophisticated grasp of the underlying technology. This permitted subsequent hearings to focus almost entirely on the legal questions rather than getting bogged down in technological definitions. Whenever a technical question did arise later, the judge could easily clarify it with counsels. Developing this level of technological literacy is vital for resolving modern tech disputes, and it also underscores the benefits for courts in actively engaging academic and technical experts, even in other areas of law.

6. Beyond intellectual property, are there other modern areas of law – such as competition law in digital markets or privacy law – where you see an urgent need for expert evidence and academic perspectives to be integrated?

Absolutely. In fields like competition law, digital regulation, and data privacy, academic expertise can offer immense value. However, a strict framework must govern these appointments to prevent conflicts of interest. For example, if an academic receives research funding from a technology company directly or indirectly, that financial relationship must be fully disclosed if they are invited to advise a court on matters affecting that company or its competitors. Proper disclosure and conflict avoidance are paramount for Courts to make an informed decision on whether such appointments are appropriate. This is equally applicable to appointment of practicing lawyers as amici. I must add here that this needs to be followed rigorously also on the policy making side. Unfortunately, many a times when lawyers or academicians become part of important policy making committees, they do not disclose potential conflict of interests. This issue needs to be addressed on priority.

7. There is a common assumption that academics operate at an arm’s length from the practical realities of litigation. In your experience, how do academic policy discussions differ from the courtroom approach, where judges must decide the concrete rights and liabilities of two opposing parties?

It is absolutely essential for academics to stay grounded in practical realities. In fact, keeping pace with ground-level developments is a prerequisite for meaningful teaching. If I do not understand how a technology operates in the real world or how active litigation is structured, I am not doing justice to the subject matter.

In intellectual property law, one must closely monitor not only courtroom disputes but also real-world commercial licensing and transactions on both a national and global scale. Failing to bridge this gap only reinforces the stereotype of the detached academic. This is equally true in other fields; for example, you cannot effectively teach criminal law without a deep comprehension of procedural realities on the ground.

8. As the first AI-related dispute adjudicated by an Indian court, this case represents a landmark. Looking ahead, what kinds of AI-related disputes do you expect to see in the courts? For instance, the Indian Patent Office recently rejected the inventor status for the DABUS AI. Do you foresee similar patent or other AI disputes being litigated in court soon?

Given the widespread adoption of LLMs, we are going to see a substantial increase in litigation. While ANI v. OpenAI centred on news articles, we may soon see disputes involving other creative mediums such as books, music, and art.

The legal outcomes of these future cases could vary significantly. In this particular matter, the court found no prima facie infringement on either the training or output side based on the specific facts presented. While I expect that the legal status of the training side might remain stable for some time, output-side liability is highly fact-specific and will have to be decided on a case-by-case basis.

Regarding patent law, now that the Indian Patent Office has officially rejected the DABUS application (which sought to recognise an AI as an inventor), that decision is bound to be heavily litigated.

My hope is that the courts will hear from a broad spectrum of stakeholders rather than limiting the dialogue to the immediate parties. The public interest dimensions of AI inventorship are massive. In ANI v. OpenAI, the court had the benefit of exhaustive submissions from diverse stakeholders and it helped the court to see diverse public interest dimensions also. I hope the courts adopt a similar, inclusive approach when tackling complex patent law questions like AI inventorship.

9. Are there any concluding thoughts or key takeaways you would like to emphasise regarding the wider impact of the ANI v. OpenAI decision?

The implications of this judgment extend far beyond generative AI and large language models. This is one of the few instances where an Indian court has very clearly and accurately articulated the integral structural balance of Indian copyright law. Specifically, the court can be seen highlighting that the user rights under copyright law are an integral, foundational component of the copyright system.

Furthermore, the court’s treatment of the public interest is exemplary. The manner in which it developed a unique fairness test, which balances the interests of both copyright holders and copyright users, instead of blindly adopting the four factors from the US fair use framework, is a remarkable contribution. I truly hope that this approach serves as a blueprint for future decisions across all areas of intellectual property law.

10. Since this judgment was in response to an interim application, the court’s findings were necessarily prima facie in nature. If this case goes to a final trial, do you think there will be an even deeper engagement with the underlying technology, or might the court’s stance evolve?

Two critical factors must be kept in mind here. First, as a matter of practical reality, the vast majority of intellectual property disputes in India are effectively decided at the interim stage. While cases can theoretically proceed to a full trial, most IP litigations in India conclude with the interim order.

Second, this interim proceeding was incredibly thorough, spanning over 20 hearings that lasted between 1.5 to 2 hours each. Given the extraordinary depth of these hearings, it is difficult to say what additional technical or legal ground would be broken during a full trial. I do not mean to endorse excessively long interim proceedings and ideally, interim applications should be decided swiftly with minimal oral arguments. But the reality of this case is that the court conducted an incredibly detailed review. We will have to see if the parties choose to take this matter to a full trial, and if so, whether the final findings deviate from this prima facie ruling.