Back
Denies Against In Case Against
In a landmark ruling that settles significant uncertainty surrounding the deployment of generative artificial intelligence in India, the has refused to grant an against the United States-based technology firm . The decision, delivered by Justice Amit Bansal on , marks the conclusion of a high-stakes legal battle initiated by () in . The court’s order addresses whether the unauthorized usage of copyrighted news reports for training (LLMs) constitutes an infringement of , offering a clear interpretation of the "" exception within the digital age.
Background and Genesis of the Dispute
The lawsuit, the first of its kind initiated by an Indian media organization, challenged the practice of using news content to train artificial intelligence systems. contended that its copyrighted reports were being utilized by to fuel the commercial operations of its chatbot, ChatGPT, without authorization or compensation. The plaintiff further alleged that the AI-powered model frequently generated responses that either plagiarized original reports verbatim or created "hallucinated" news, falsely attributing content to the agency, thereby damaging its professional reputation and contributing to the spread of misinformation.
countered these allegations by asserting that its machine learning processes are in nature and that copyright laws do not extend to the raw facts contained in news reports. Crucially, the technology firm argued that publishers retain the ability to "" of training datasets by modifying their website metadata—a mechanism that has since implemented.
The Court’s Rationale on Jurisdictional Reach
Before arriving at a decision on the merits, the addressed a threshold challenge concerning . had argued that because its AI models were architected, trained, and hosted on servers located exclusively in the United States, the Indian court lacked the authority to oversee the dispute. Justice Bansal unequivocally rejected this contention, reaffirming that the court possesses the to entertain the suit, given the accessibility and impact of the technology within the Indian market.
Establishing the Exception
The core of the judgment rests on the application of . The court held that the storage and processing of literary works by to facilitate the training of its LLMs falls under the "" exception, provided the use remains .
Justice Bansal observed,
"
’s act of storing
’s original literary works for training its LLMs amounts to a
and
use, which is
protected under Section 52(1)(a) of the Copyright Act and consequently does not constitute infringement under
."
The bench emphasized the distinction between the "" use of creative content—which might attract copyright liability—and the "" use in machine learning, where the model extracts patterns and data structures rather than the literal expression of the work itself.
Analysis of (RAG)
A significant component of the legal challenge involved the use of (RAG) to provide real-time responses to user queries. The court closely inspected whether these responses infringed upon 's copyright. It found that the outputs were not "substantially similar" to the original works published by the media outlet.
According to the order,
"
has failed to satisfy this Court that any memorisation or regurgitation of
's copyrighted literary works has happened to the responses generated by ChatGPT."
By failing to prove that the chatbot provided verbatim copies of protected news reports, the plaintiff could not establish a
case of infringement regarding the model’s outputs.
Wider Implications and Contributions
Recognizing the technical complexity of this sector, the court appointed and of the as . Their contributions provided the court with a nuanced view of the intersection between copyright policy and machine learning. Dr. Scaria’s submissions regarding the necessity of distinguishing between and uses proved pivotal to the court’s decision-making process.
The proceedings drew extensive participation from other stakeholders in the industry, including the , the , and the . These organizations had filed intervention applications, expressing valid concerns regarding the unremunerated use of news reports, music, and academic literature for the training of proprietary AI models.
Preserving Public Interest and Innovation
In the final assessment, the court applied the "" for granting an : the existence of a case, the , and the potential for . The bench concluded that failed to satisfy these criteria.
The court explicitly noted that the public interest outweighed the potential for temporary loss of exclusive control claimed by the plaintiff. It observed that restraining a technological entity like at this stage would cause "" not only to the defendant but to the broader public, by effectively curtailing progress in artificial intelligence and limiting access to innovative tools that have become integral to the modern digital infrastructure.
Perspectives for the Future of Legal Practice
This interim order serves as a vital precedent for legal professionals navigating the intersection of technology and intellectual property. By signaling that courts are likely to perceive AI training as "," the has established a high bar for media organizations seeking to challenge AI developers on copyright grounds. Legal practitioners must now pivot toward strategies involving technological verification—such as evidencing actual "memorisation" or "regurgitation"—rather than relying solely on the fact of data ingestion. As AI continues to evolve, the distinction between functional training and verbatim reproduction of " content" will likely remain the focal point of future litigation in this rapidly shifting landscape.
