DPIIT signals Copyright Act changes to address AI issues, Pg14

Key Highlights:

• The Union government plans to amend the Copyright Act within three years to address issues arising from Artificial Intelligence (AI).
• DPIIT has proposed a "blanket licensing" framework for websites crawled by Large Language Models (LLMs) like ChatGPT, ensuring royalty payouts via a copyright society.
• The proposed Copyright Royalties Collective for AI Training (CRCAT) will seek payments from AI firms upon commercialization of their models.
• Nasscom dissented from the DPIIT proposal, suggesting publishers should have the option to exclude their data from training models.

Detailed Insights:

• The DPIIT proposal aims to resolve tensions between online content publishers and AI firms regarding compensation for using content to train LLMs.
• The government will release another working paper to determine if AI-generated works are copyrightable and how authorship is decided.
• The proposed "blanket licensing" framework generally allows content scraping by AI developers while ensuring they compensate content publishers.
• Concerns have been raised that the DPIIT proposal reverses copyright jurisprudence by shifting the burden of proof of infringement to AI developers.
• News publishers, including ANI and The New York Times, have sued OpenAI for allegedly using their content in chat conversations.

Key Concepts Involved:

• Copyright Act: Law protecting the rights of creators of original works.
• Artificial Intelligence (AI): Simulation of human intelligence processes by computer systems.
• Large Language Models (LLMs): AI algorithms that use deep learning techniques and massive datasets to understand and generate human language.
• Blanket Licensing: An agreement that allows a user to access a large repertoire of copyrighted works in exchange for a fee.