OpenAI explained to the Delhi High Court on Tuesday that training ChatGPT with public data does not constitute commercial activity. OpenAI made this statement during a court hearing about a lawsuit filed by ANI Media which accused the company of stealing content to train its large language models (LLMs) including ChatGPT.
OpenAI through its CEO Sam Altman explained that training an LLM can exist as either a commercial or non-commercial operation. The company’s legal team explained that training activities including research with public data do not meet the criteria for commercial activities. A model becomes commercial when it functions as a service platform for other users.
The AI firm demonstrated that ChatGPT has actually increased website traffic for ANI which indicates no commercial damage occurred. OpenAI defended its position by stating that copyright protection applies only to creative expressions rather than basic facts or general knowledge which serves as the foundation for training language models.
Justice Amit Bansal marked down the arguments before setting the matter for additional hearing on May 16.
ANI previously argued that ChatGPT received its content training from ANI which resulted in the public availability of the AI model. The amicus curiae Adarsh Ramanujan clarified that LLMs operate as predictive statistical models instead of truth machines which generate content from statistical patterns. The majority of ANI’s content exists behind a subscription wall so subscribers possess the legal right to redistribute this content.
The ongoing case generates vital questions about AI training ethics together with copyright law and modern technology boundaries of public domain usage.