By Tejas Basanwar

Introduction

The rise of artificial intelligence (AI) has changed the dynamics of information consumption and creation. The use of copyrighted material for AI training raises important questions concerning copyright protection and artificial intelligence.

Large Language Models (LLMs) like ChatGPT are trained on large-scale datasets containing publicly available information. In this process, the use of copyrighted material for AI training raises critical copyright law questions.

ANI Media Pvt. Ltd. v. OpenAI OpCo LLC is one of the significant copyright disputes in India examining the alleged unauthorised use of copyrighted news content for training large language models.

The case concerns whether the use of copyrighted news content for training AI constitutes copyright infringement and the extent to which copyright law can govern the use of copyrighted material during AI training.

Case Status: The case is still pending before the Delhi High Court and is attracting attention because of its potential implications for AI developers, news publishers, copyright owners, and policymakers.

Background

ANI Media Pvt. Ltd. (ANI) is a leading news company publishing news reports, photographs, videos and other journalistic content. The company has invested in producing news content protected under the Copyright Act, 1957.

OpenAI OpCo LLC is an artificial intelligence company associated with the development of ChatGPT, an AI chatbot capable of generating responses based on data and patterns on which its models have been trained.

ANI alleges that OpenAI used its copyrighted news content for training its AI models without permission and that ChatGPT can reproduce ANI's news reports, including allegedly reproducing certain content verbatim.

ANI approached the Delhi High Court alleging copyright infringement and claiming that unauthorised use of its content for AI training could adversely affect its licensing revenue and the commercial value associated with its journalism.

Arguments of the Parties

Arguments of ANI
  • ANI's news reports and journalistic content are protected by copyright.
  • OpenAI allegedly used ANI's content for AI training without permission or licence.
  • ANI argues that such use constitutes copyright infringement.
  • ANI alleges that ChatGPT can reproduce its copyrighted content verbatim.
  • Such use may diminish the commercial value of ANI's licensed content.
  • AI developers should not commercially benefit from copyrighted works without permission.
Arguments of OpenAI
  • ChatGPT is not merely a database of stored articles.
  • AI models learn statistical patterns in language.
  • ChatGPT generates responses rather than simply reproducing copyrighted material.
  • OpenAI disputes that AI training itself constitutes copyright infringement.
  • Issues concerning jurisdiction and applicability of the Copyright Act also require consideration.

Relevant Statutory Provisions

13

Section 13 – Subject Matter of Copyright

Section 13 provides the categories of works in which copyright subsists, including original literary, dramatic, musical and artistic works.

14

Section 14 – Meaning of Copyright

This provision provides copyright owners with specified exclusive rights, including the right to reproduce a protected work.

51

Section 51 – Copyright Infringement

Section 51 addresses circumstances in which acts involving rights reserved to a copyright owner may constitute infringement when carried out without authorisation.

52

Section 52 – Exceptions to Infringement

Section 52 provides circumstances in which specified uses of copyrighted works do not constitute copyright infringement, including certain fair-dealing exceptions.

Proceedings Before the Delhi High Court

The Delhi High Court has heard arguments concerning preliminary questions including maintainability and jurisdiction. The Court has also considered the wider implications of the dispute for copyright law and artificial intelligence.

The dispute raises significant questions extending beyond the immediate parties because its outcome may have consequences for technology developers, media companies, copyright owners and the development of copyright law in relation to AI.

Why this case matters: The litigation raises important questions concerning the relationship between AI training, copyrighted material and India's existing copyright framework.

Analysis

The dispute between ANI and OpenAI highlights the emerging tension between copyright protection and generative artificial intelligence.

From the copyright owner's perspective, unauthorised use of protected content during AI training may amount to exploitation of copyrighted works and may potentially affect their economic value.

AI developers, on the other hand, argue that model training involves analysis and learning from statistical patterns rather than conventional reproduction of copyrighted works.

The case therefore raises important questions about whether copying undertaken during the technical process of AI training falls within the exclusive rights provided under copyright law, whether statutory exceptions apply, and when an AI-generated output may potentially infringe an existing copyrighted work.

Potential Impact

The ANI v. OpenAI dispute may have implications for multiple stakeholders operating at the intersection of technology, media and intellectual property.

01

AI Copyright Disputes

It may influence future copyright disputes involving artificial intelligence.

02

Licensing Practices

It may influence licensing practices between AI developers and copyright owners.

03

Future Litigation

It may provide guidance for similar AI-related copyright litigation in India.

04

AI Regulation

The case may contribute to the development of India's regulatory landscape for artificial intelligence.

Conclusion

ANI Media Pvt. Ltd. v. OpenAI OpCo LLC is an important ongoing copyright dispute concerning the alleged unauthorised use of news content for training artificial intelligence systems.

The litigation raises fundamental copyright questions, including whether the use of copyrighted material during AI training constitutes infringement and how provisions of the Copyright Act, 1957 apply to AI-generated content.

The judicial determination may provide important guidance to AI developers, media companies, copyright owners, policymakers and other stakeholders regarding the use of copyrighted works for artificial intelligence development in India.

Have Questions About Intellectual Property?

Connect with our team to discuss patents, copyrights, trademarks, IP strategy, AI-related intellectual property issues, or other IP requirements.

service_page_form