Copyright’s generative AI experiment

Historical Evolution of Copyright Law

• Copyright law was created in 1710 to protect publishers from unauthorised publication and promote learning.

• It has evolved to accommodate various technologies, including the printing press, photocopying machines, recording devices, and the internet.

Generative AI and Copyright Law

• Generative AI companies, like Open AI, are at a crossroads with copyright law.

• Open AI uses Internet scraping technology, which trains the platform on all available knowledge.

• Copyright infringement cases are fought on subjects such as literature, music, and photographs.

Case of Open AI in India

• The Federation of Indian Publishers and Asian News International initiated copyright infringement claims against Open AI for training the AI platform with publishers’ works without their consent.

• Similar cases are pending in the U.S. courts, where Open AI has developed an opt-out mechanism for publishers to opt-out from data set training.

Court’s Role in Open AI Case

• Professor Dr. Arul George Scaria suggests the court should address the feasibility of unlearning information from training content.

• The Indian courts lack competence to hear the case, and the Indian Copyright Act adopts a different exception test than the ‘fair use’ test established in the U.S.

• The opt-out mechanism developed by Open AI could impact the future of generative AI, as the efficiency of the technology depends on the material it is trained upon.

Solutions to the Problem

• Copyright law should not be turned on its head to prohibit future creators from accessing this benefit.

• The arguments of publishers in the case could potentially view human creation and machine creation differently in the future.

• The foundational norms of copyright law offer solutions to the existing problem.