Delhi High Court: OpenAI ChatGPT Not Guilty of Copyright Infringement

• ANI Media (P) Ltd. v. OpenAI Inc. is the first lawsuit in India against a generative AI system for copyright violation by way of excessive use of copyrighted works to train a large language model, and while an interim injunction was dismissed, the main suit continues.
• The case raises novel questions about whether ingesting copyrighted content for AI training constitutes copyright infringement under the Indian Copyright Act, 1957.
• While the Delhi High Court's interim order prima facie found that OpenAI's actions fell under fair dealing, the final outcome of the main suit will determine whether OpenAI can escape liability via fair dealing under Section 52 of the Act.

The dispute centers on the ingestion of copyrighted content for AI training, raising novel questions about copyright infringement under the Indian Copyright Act, 1957.

In a landmark lawsuit, India's Delhi High Court recently dismissed an application for an interim injunction by Asian News International (ANI) against OpenAI over copyright infringement, though the main suit remains underway. ANI alleges that OpenAI used its proprietary journalism to train ChatGPT, a large language model, without permission. The plaintiff claims that ChatGPT reproduced certain content from ANI's articles, which are subject to paywalls, and is seeking Rs 2 crores in monetary damages and a permanent injunction against OpenAI's continued use of its content.

The dispute centers on the ingestion of copyrighted content for AI training, raising novel questions about copyright infringement under the Indian Copyright Act, 1957. The case has significant implications for AI companies operating in India, as the outcome will determine whether the use of copyrighted materials in AI training constitutes copyright infringement.

OpenAI's defense hinges on several arguments, including that LLM training does not involve traditional reproduction but rather develops statistical probabilities based on data. The Delhi High Court, in its interim order, rejected OpenAI's argument that it lacked jurisdiction over the matter, and prima facie found that any potential infringement would be excused under fair dealing as defined in Section 52 of the Act.

The Indian Copyright Act, 1957 provides exclusive rights to copyright owners, including the right to control reproduction. Section 14 of the Act explicitly includes electronic storage and reproduction within its scope. The Act also introduces fair dealing provisions in Section 52, which allows for limited use of copyrighted materials without permission. While the main suit is pending, the Delhi High Court's interim order has provided initial judicial guidance on the applicability of these provisions to AI training in Indian courts.

The case raises questions about the limits of fair dealing under Section 52 and whether it can accommodate the unique demands of AI training. The outcome will have far-reaching implications for creators, publishers, and AI companies operating in India, as well as the broader intellectual property landscape in an era dominated by AI.

The outcome of this case will carry significant weight not just for the parties involved but for every creator, publisher, and AI company operating in India. The ruling will determine whether the use of copyrighted materials in AI training constitutes copyright infringement under the Indian Copyright Act, 1957. This has profound implications for the development and deployment of large language models like ChatGPT.

The case also highlights the need for a more nuanced understanding of fair dealing provisions in the context of AI training. The court's ruling will set a precedent for future cases and shape the trajectory of intellectual property law in India as it relates to AI.

This case has significant implications for AI companies operating in India, as the outcome will determine whether the ingestion of copyrighted content to train large language models constitutes copyright infringement under the Indian Copyright Act. Lawyers should watch for the court's ruling on fair dealing and its potential impact on their clients' use of copyrighted materials in AI training.

Source: Original reporting via Briefly