Generative AI and Copyright Transparency in India


Author: Mrudula Mahesh Kale,
University of Mumbai Law Academy


Abstract

GAI or generative artificial intelligence is a growing field, and creativity is as well impacted, as computers mastered the possibility to generate texts, visuals, music and others. But this technological advancement poses serious questions about the laws governing the proprietary rights in the ongoing copyright domain. In India again where the copyright law is regulated under Copyright Act of 1957, issues related to authorship, ownership and liability linked with generative AI outputs have emerged as major concern. This study draws a discourse on the mapping of generative AI with copyright transparency, assessing the current legal dispensation in India, possible impacts, and, need for change.


To the Point

Generative AI systems like ChatGPT, DALL-E, and others produce outputs that mimic human creativity, raising critical questions about copyright ownership and infringement:
Ownership of AI-Generated Works: Who can be the copyright owner of the AI-generated content?
AI as an Author: To what extent non-human entities are allowed as authors under Indian law?
Transparency in Training Data: What ways can organizations match their AI training approach to copyright legislation?
There is a need to define a strategy that best protects policy makers, innovators, and holders of intellectual property rights, equally and sufficiently.


Use of Legal Jargon

India’s copyright law has its foundation in The Copyright Act, 1957. According to Section 2(d)(vi), author is someone who makes a work. This take out all the AI systems and put questions on issues to do with originality and authorship. Other key provisions include:

Section 13: Holds original literary, art and musical pieces.

Section 17: Sets legal ownership and ownership rights in an author.

Section 51: Forbids unauthorised copying or using close arrangement or translation, stressing on the infringements’ penalties.
These legal constructs are undermined by generative AI through posing new questions about derivative work and copyright protection for content generated by AI systems.


The Proof

Some of the areas which need to be made public include transparency in AI Training Data.
Training data of generative AI systems is usually obtained unclear and it primarily relies on big data. This raises concerns about:

Copyright Infringement: Use of copyrighted material in training models may fall foul of Section 51 of the Copyright Act, 1957.


Fair Use Doctrine: Section 52 permit limited use for things like research and in education sector but its suitability for the training of artificial intelligence is still ambiguous.


Ethical Concerns: And while opacity erodes principles of informed consent and accountability, there is no suggestion that contractors must be more selfish than they already are.


International Developments

European Union: The EU’s AI Act says that transparency is important and that the data used to train the AI must be disclosed.

United States: The Copyright Office announced that content created by AI without the help of a human being cannot be copyrighted.

Australia: Changes to the Canadian Copyright Act do allow text and data mining in reference to artificial intelligence for purposes of fair dealing.
Such examples are quite instructive for India’s increasingly complex legal context.


Case Laws

Indian Jurisprudence

While Indian courts have not directly addressed generative AI and copyright, the following cases provide guidance:

Eastern Book Company v. D.B. Modak (2008): Focuses on the necessity of creativity in the sphere of the protection of copyright that has intensified the difficulties connected with the AI creation of particular works.

Super Cassettes Industries Ltd. v. MySpace Inc. (2011): Intermediate liability, an issue of concern when AI platforms acts as a distributor of infringing works.


International Cases
Thaler v. Commissioner of Patents (Australia, 2021): Sustained no new procedural or substantive right that AI cannot be an inventor and denied non-human authorship.
Feist Publications, Inc. v. Rural Telephone Service Co. (US, 1991): Only underlined originality for obtaining a copyright.


Conclusion

The emergence of generative AI requires a fair level of copyright disclosure in India. Policymakers and stakeholders must consider:
Legislative Reforms: Introduce changes to the Copyright Act 1957 to cover creations produced by AI; Require data openness.
Clear Guidelines: Outline the boundaries of fair use applied to AI training, in reference to best practices seen internationally.
Ethical Development: Encourage licensing of training data and watermarking as a way of tracing copyrighted content.
Judicial Intervention: The court in India must provide some certainty by deciding cases that involve content created by artificial intelligence.
The measures mentioned above have the potential to turn India into a hi-tech innovative hub while protecting creators’ rights and legal certainty in the age of Artificial Intelligence.

Leave a Reply

Your email address will not be published. Required fields are marked *