Delhi High Court Rules Against ANI In Landmark ChatGPT Copyright Injunction Bid

DNI SUMMARY — KEY POINTS

• The Delhi High Court has officially dismissed the interim plea filed by ANI seeking to prevent OpenAI from utilizing its copyrighted news reports for training purposes.
• Justice Amit Bansal determined that the use of content for training Large Language Models is prima facie protected under the fair dealing exception of the Copyright Act.
• This legal battle represents the first major confrontation in India where a prominent media organization challenged a global generative AI giant over intellectual property rights.
• The court emphasized that granting an injunction would result in significant public harm while failing to establish that the chatbot reproduces original, copyrighted news content.
• While the interim injunction request was denied, the overarching copyright lawsuit remains active as the court continues to evaluate the merits of the broader legal arguments.

The Delhi High Court delivered a significant decision this week, refusing to grant an interim injunction to the news agency ANI in its high-profile copyright infringement lawsuit against OpenAI. The case, which marks a major legal milestone for the generative AI sector in India, centers on whether the ingestion of copyrighted news reports for training Large Language Models violates intellectual property protections. By denying the requested relief, the court has allowed the developer to continue its operations without immediate judicial restriction as the litigation proceeds on its merits.

Legal Threshold for Fair Dealing

Legal Threshold for Fair Dealing

Presiding over the case, Justice Amit Bansal observed that the act of storing and processing information for AI training constitutes a transformative use of data. The bench found that such activities fall under the purview of the fair dealing exception, specifically governed by Section 52 of the Copyright Act. This legal interpretation suggests that non-expressive use—where the machine learns patterns rather than simply reproducing text—is currently viewed as permissible under the existing framework, providing a vital regulatory victory for developers operating within the country.

Jurisdictional Claims and Technical Defenses

The court explicitly stated that the news agency failed to satisfy the essential triple test required for an interim injunction, which encompasses a prima facie case, balance of convenience, and the prevention of irreparable harm. Furthermore, the judge noted that restricting the functionality of advanced AI models would not only damage the interests of the creator, OpenAI, but also adversely affect the public interest. Broad access to emerging technologies, according to the court, remains a priority that outweighs the specific claims brought forward by the plaintiff during this procedural stage.

Jurisdictional Claims and Technical Defenses

Systemic Hallucinations and Editorial Integrity

While the court rejected the plea for an immediate halt, it addressed critical jurisdictional questions that had been raised throughout the proceedings. The Delhi High Court affirmed its authority to hear the case, dismissing arguments that the location of servers outside the country precluded legal oversight in India. Despite this ruling on jurisdiction, the court remained unconvinced by claims that the chatbot had effectively memorized or verbatim reproduced protected works, noting that the outputs generated by the system are not substantially similar to original journalism.

The lawsuit, which gained momentum in November 2024, expanded significantly when various industry bodies, including the Digital News Publishers Association, chose to join the legal challenge as intervenors. These entities argued that the unauthorized use of their intellectual property to improve AI models undermines the commercial viability of traditional media organizations. By standing in solidarity with the agency, these industry groups aimed to force a comprehensive re-evaluation of how international tech firms acquire data within the local digital marketplace, setting a tone for future digital policy debates.

Awaiting Final Judicial Guidance

Systemic Hallucinations and Editorial Integrity

Beyond the core copyright allegations, the news agency raised concerns regarding the potential for AI-generated hallucinations to damage its reputation. Counsel for the plaintiff argued that the model occasionally attributes inaccurate or fabricated stories back to the agency, thereby compromising its hard-earned credibility. Although the court acknowledged these concerns, it found insufficient evidence at this preliminary stage to conclude that such errors constitute a systemic infringement of intellectual property rights, opting to keep the primary trial focused on the legality of the model training process.

The legal proceedings are far from over, as the focus now shifts toward the substantive arguments that will be presented during the main trial phase. While the refusal of an interim injunction provides a temporary reprieve for the developer, the final judgment will likely define the long-term obligations of global technology companies working with domestic content in India. Experts anticipate that this case will serve as a foundational precedent, potentially guiding future legislative discussions on whether existing statutes require amendments to specifically address the unique complexities of machine learning.

Awaiting Final Judicial Guidance

Observers now look to the next hearing, where the developer is expected to offer a more detailed defense regarding its data ingestion practices and technical safeguards. The outcome of this case remains crucial for both the burgeoning AI ecosystem and the traditional publishing industry, which continues to grapple with the seismic shift in content distribution. As the court prepares for further deliberations, the tech community will continue to monitor the case, recognizing that the final ruling could fundamentally alter how innovation and copyright coexist in the modern digital age.