In a recent filing viewed by Reuters on Jan. 22, OpenAI informed an Indian court that any directive to remove the training data supporting its ChatGPT service would conflict with its legal obligations in the United States. The AI company, backed by Microsoft, also stated that Indian courts lacked jurisdiction to address a copyright infringement case initiated by local news agency ANI due to OpenAI's absence in the country.
In a high-profile lawsuit filed in November in Delhi, ANI alleged that OpenAI had utilized its published content without consent to train ChatGPT. OpenAI refuted the claims in an 86-page response submitted to the Delhi High Court, dated Jan. 10.
OpenAI, among other firms, has encountered multiple copyright violation lawsuits from prominent content creators accusing them of misusing their work to train AI models. The company has defended its position, stressing that its AI systems ethically leverage publicly accessible data.
As the case progresses, OpenAI highlighted its ongoing legal battle in the United States, emphasizing their obligation to retain the training data until legal proceedings conclude, as mandated by U.S. laws.
OpenAI argued that ANI's demands were outside the purview of Indian courts, stating that the company lacked a physical presence in India, and the servers housing ChatGPT's data were situated outside the country.
ANI, a news agency in which Reuters has a 26% stake, contended that the Delhi court has the authority to rule on the issue, vowing to provide a comprehensive response. The court is scheduled to review the case on Jan. 28.
OpenAI has been transitioning into a for-profit entity after securing substantial funding to remain competitive in the AI industry. The company recently forged partnerships with prominent media outlets to showcase content.
ANI expressed concerns about unfair competition due to OpenAI's commercial collaborations with other news organizations, alleging that ChatGPT reproduced ANI's content verbatim or with minor alterations upon user request. OpenAI countered by asserting that ANI had employed its own content to prompt ChatGPT deliberately.