OpenAI, the creator of ChatGPT, has contested the jurisdiction of Indian courts in ANI’s copyright infringement lawsuit filed before the Delhi High Court. In its response, OpenAI argued that neither the company nor the cause of action falls within the territorial jurisdiction of Indian courts. It emphasised that ChatGPT’s servers and training data are located outside India, and the mere accessibility of the service within India does not automatically confer jurisdiction.
The Delhi High Court had issued summons to OpenAI on November 19 in response to a lawsuit by Asian News International (ANI) over the alleged unauthorised use of its content. The case is scheduled for a hearing on January 28.
In its submission, OpenAI contended that ANI’s request to delete training data conflicts with U.S. federal laws, which mandate the preservation of evidence due to ongoing litigation in the United States. It pointed to its ongoing legal battle with The New York Times over similar issues, arguing that granting ANI’s request would compel OpenAI to violate its obligations under U.S. law.
Additionally, OpenAI asserted that the Delhi High Court’s Commercial Division lacks jurisdiction, as ANI’s lawsuit combines multiple causes of action, some of which fall outside the scope of “commercial disputes” as defined under the Commercial Courts Act, 2015. It argued that ANI’s defamation claim does not qualify as a commercial dispute under the Act.
OpenAI further claimed that ANI, as a ChatGPT user, had agreed to the platform’s terms of use, which mandate arbitration and designate courts in San Francisco, California, as the exclusive forum for legal disputes. Consequently, ANI must pursue its claims by these terms.
Addressing copyright concerns, OpenAI countered ANI’s claim that storing its works for training a Large Language Model (LLM) constitutes copyright infringement under Indian law. It argued that Indian copyright law protects the expression of ideas, not the underlying facts or ideas themselves. In the digital realm, LLM training involves analysing patterns rather than storing or replicating data verbatim. OpenAI maintained that this process does not violate copyright, as it transforms freely available online information into simplified interpretations rather than directly copying it.
The organisation asserted that such use falls under the “fair use” exception, arguing that machine learning’s transformative nature does not amount to substantial reproduction of original works. It further contended that ANI’s reports primarily consist of factual content, which is not eligible for copyright protection under Indian law.
OpenAI also refuted ANI’s allegations that ChatGPT reproduces its content verbatim. It claimed that ANI had deliberately prompted ChatGPT with excerpts of its own articles in an attempt to manipulate the model into reproducing them. However, OpenAI argued that these efforts were unsuccessful, as ChatGPT generates unique responses based on input and learned patterns rather than direct replication.
Additionally, OpenAI addressed ChatGPT’s limitations, including the potential for “hallucinations”—instances where the AI generates incorrect or nonsensical information. It asserted that ANI’s defamation claims based on such hallucinations are unfounded, as there is no demonstrated harm to ANI’s reputation.
Lastly, OpenAI highlighted the broader societal benefits of AI technology and urged the court to dismiss ANI’s lawsuit.