Read Time: 07 minutes

Synopsis

Advocate Sidhanth Kumar, representing ANI, contended before the Delhi High Court that OpenAI using ANI’s content to train its language models leads to dilution of ANI’s market, resulting in unfair competition. These observations were made in a case filed by ANI Media, alleging that OpenAI had infringed upon its copyright by utilizing its content without authorization. 

ANI asserted that ChatGPT used its material to train the AI software, which is then stored and provided access to ANI’s content for users. However, OpenAI had first contended that since the company neither operates in India nor are its servers located in India, therefore no cause of action arises in India.

During the proceedings, Amicus Curiae Arul George Scaria appeared via video conferencing, while Amicus Curiae Adarsh Ramanujan was present in court. Advocate Ramanujan, presenting arguments on behalf of OpenAI, emphasized that language models were not designed as ‘truth machines’ but operated based on predictive algorithms. He informed the court that responses generated by such models varied depending on when the query was made, reinforcing the notion that these models relied on language patterns rather than factual knowledge.

The court sought clarity on whether the language model's operation was based on prediction rather than knowledge. Advocate Ramanujan responded by asserting that the model did not use ANI’s data directly. He highlighted that the issue involved data collection preceding training, making it necessary to examine the legal framework governing such processes. He further pointed out that ANI’s content was often published behind a subscription wall, yet its subscribers, who had varying levels of paywall restrictions, could republish the content.

Addressing the legal implications of public accessibility, Advocate Ramanujan argued that the mere availability of content in the public domain did not nullify copyright protections. He emphasized that OpenAI contended it had only used publicly available information, yet data collection inherently involved storage, which facilitated potential reproduction.

Expanding on text and data mining (TDM), he noted that other jurisdictions had introduced specific exceptions permitting such activities. However, Indian law did not explicitly recognize these exceptions outside the fair use doctrine. He compared OpenAI’s approach to that of search engines, explaining that data collection served as a precursor to data mining rather than mere storage.

A key argument centered on OpenAI’s ‘opt-out’ mechanism, which allowed content creators to exclude their data from future training. Advocate Ramanujan asserted that the burden of protection should not be placed on the plaintiff, arguing that past infringements could not be excused by an opt-out option under copyright law. He maintained that once a model was trained on specific data, it could not be undone.

Advocate Sidhanth Kumar, representing ANI, countered that ANI employed journalists globally to generate its content and maintained one of the largest video archives online. He explained that subscribers were permitted to republish content without modifications or use excerpts, but they were not authorized to resyndicate or sublicense it further. He urged the court to determine whether mere public accessibility granted OpenAI the right to use ANI’s content for training its models.

Amicus Curiae Arul George Scaria, in the last hearing, remarked that “Courts should adopt a two-step approach. Firstly, we should decide whether the concerned use falls under exceptions like private/personal use, criticism/review, or reporting of current events. Secondly, we should employ a fairness analysis”.

For Plaintiff: Advocate Sidhanth Kumar
For Respondent: Senior Advocate Akhil Sibal
Case Title: ANI Media Pvt Ltd v OpenAI Inc (CS COMM 1028/2024)