ChatGPT Data Removal Will Breach US Legal Obligations: OpenAI to Delhi HC

New Delhi: Tech giant OpenAI told an Indian court that any order to remove training data used to power ChatGPT's service would conflict with its legal obligations in the United States, Reuters reported citing a recent filing.

The most high-profile and closely-tracked lawsuit was filed on November 23, 2023, in New Delhi where Indian news agency ANI accused OpenAI of using its published content without its content permission to train its artificial intelligence models. Furthermore, the news agency also sought to remove the content from ChatGPT's training data, arguing that the content had been stored in the system's memory without authorisation.

In the 86-page filing at the Delhi Court dated January 10, the Microsoft-back tech company argued that the court did not have jurisdiction over the case, as the company does not have a presence in India. It further stated that its operations are based outside of India, and the servers storing the training data are also located abroad.

"OpenAI "is therefore under a legal obligation, under the laws of the United States to preserve, and not delete, the said training data", it said.

The company has "no office or permanent establishment in India ... the servers on which (ChatGPT) stores its training data are similarly situated outside of India," it added.

The company also emphasised that under US laws, it is required to preserve the training data while ongoing litigation in the US is unresolved.

This isn't the first case against OpenAI. The company is facing similar lawsuits in the US by several big media organisations including The New York Times. However, the tech giant has been denying all these allegations, stating that its AI systems make fair use of publicly available information.

The Delhi High Court is scheduled to hear the case on January 28. The outcome of this case could have significant implications for how AI companies handle copyright issues related to the training of their models.

(Inputs from Agencies)