OpenAI has told the Delhi High Court that complying with a request to delete training data used for its ChatGPT model would conflict with its legal obligations in the United States. This statement came in response to a copyright infringement lawsuit filed by the Indian news agency ANI in November.
It accuses the company OpenAI of using content it has published without its consent to train ChatGPT. The lawsuit seeks an injunction to delete ANI’s data already stored by the AI model and violate copyright claims.
Allegations by ANI and Defense from OpenAI
In an 86-page court filing dated January 10, OpenAI denied the allegations, emphasizing that its AI models make fair use of publicly available data. ANI claims that OpenAI continues to store and use its content despite assurances during a November hearing that ChatGPT would no longer rely on ANI’s material.
OpenAI explained that its training data is preserved under U.S. law as it faces similar lawsuits, including one filed by the New York Times. “OpenAI is therefore under a legal obligation, under the laws of the United States, to preserve, and not delete, the said training data,” the company stated.
Jurisdiction Dispute
OpenAI further argued before the Delhi High Court that it lacks the jurisdiction to deal with the matter at hand, as this company does not have a physical presence or infrastructure in India. “The servers upon which (ChatGPT) stores its training data are similarly situated outside of India”, OpenAI added in its submission.
ANI, in which Reuters holds a 26% interest, contends that the Indian court has the authority to adjudicate the matter. The news agency also expressed concerns about unfair competition due to OpenAI’s commercial partnerships with global media outlets like Time magazine and the Financial Times.
ANI Accuses ChatGPT of Verbatim Reproduction
ANI has claimed that ChatGPT reproduces verbatim or substantially similar extracts of its content in response to user prompts. OpenAI refuted this, stating ANI had deliberately used its articles as prompts to test the AI’s responses, calling it an attempt to “manipulate ChatGPT.”
High-Stakes Case for AI and Copyright
This case represents one of the most significant legal challenges to AI use in India, touching on critical issues of copyright, fair use, and jurisdiction. It also highlights the growing tension between AI firms and content creators worldwide.
The Delhi High Court will take up the case on January 28. OpenAI is, meanwhile, dealing with growing scrutiny as it moves from a nonprofit to a for-profit organization. In 2023, the company raised $6.6 billion to finance its AI progress and sealed deals with some of the world’s leading media houses.
As the date for the court approaches, the verdict in this case may mark a precedent for copyright disputes involving AI in India and elsewhere.