ANI vs OpenAI: Delhi High Court Appoints Amicus Curiae As It Takes Up India’s First Major AI-Related Complex IP Dispute

ANI vs OpenAI: Delhi High Court Appoints Amicus Curiae As It Takes Up India’s First Major AI-Related Complex IP Dispute

Indian news agency Asian News International (ANI), joining other publishers worldwide, has filed a copyright infringement suit against OpenAI over its AI model, ChatGPT. The proceedings extend a series of legal challenges against AI companies globally and could have significant implications for how generative AI models use publicly available data and enforce copyright enforcement globally.

ANI claims that OpenAI used its content without permission to train its large language models (LLMs) and generate verbatim responses. However, OpenAI argues that only a fraction of its dataset includes news content, with ANI representing an even smaller subset.

ANI’s main contentions:

- OpenAI used ANI’s copyrighted content unauthorizedly/ without permission to train its AI model Chat GPT.

- ChatGPT outputs are substantially similar or reproductions to ANI’s content.

- Permanent storage of ANI’s content by Chat GPT.

-False attribution of content to ANI by Chat GPT.

OpenAI’s Preliminary Objections:

In the first hearing that took place on 19th November 2024, OpenAI raised preliminary objections on

- Delhi High court does not have territorial jurisdiction in the matter as Open AI have no physical presence or servers in India.

- Copyright only protects expression, not underlying facts.

- ANI’s failure to demonstrate a single instance of Chat GPT reproducing their specific expression.

TechLegis commented on the case, that proceedings and the decision in this case in India and elsewhere shall significantly impact the future of AI industry globally and that of the creative industry. From an Indian law perspective the case promises to deal with some very critical issue that may define and lay down future foundations of the intersection of IP law and the evolution of AI technology while considering propositions like if training AI and ML models using copyrighted content would tantamount to IP infringement and if ‘fair use’ defence is a available to companies, whether mere storage of copyrighted content for ML/ AI model trainings is permissible and can such use be considered transformative fair use under the India law.

It remains to be seen how courts in India and elsewhere would deal with such issues and what they will decide especially in light of contradictory decisions given by the courts in US on the ownership of copyrights in AI generated works of art.