Changes to Copyright Law in the Context of Artificial Intelligence

The Union government plans to amend copyright law due to demands from artificial intelligence developments. This process is expected to be completed within three years, as stated by a senior official on December 11, 2025.

Proposed Framework

• The Department for Promotion of Industry and Internal Trade (DPIIT) has proposed a “blanket licensing” framework.
• This framework suggests that websites with data used by large language models (LLMs) like ChatGPT will receive royalty payouts.
• Royalties will be managed by a copyright society, distributing money between content publishers.

Objectives of the Proposal

• To address global tensions between online content publishers and AI firms.
• Publishers argue for compensation due to their data aiding AI model development and improvement.
• AI developers may be allowed to scrape content, with eventual payments to publishers.

Implementation Steps

• A working paper will follow in two months, focusing on AI-generated works and copyrightability.
• The government plans an amendment to the Copyright Act, 1957, establishing this new regime.
• The proposal includes a Copyright Royalties Collective for AI Training (CRCAT).
• CRCAT will seek payments from AI firms only after model commercialization, not during data mining.

Contentious Issues and Responses

• Compensation of copyright holders is contentious, with lawsuits against AI firms like OpenAI by publishers like ANI and The New York Times.
• Tech industry body Nasscom, part of the drafting committee, opposed the blanket licensing model.
• Nasscom suggests publishers should opt-out of data inclusion in training models.
• Concerns about reversing copyright jurisprudence, placing the burden of proof on AI developers, were raised.

Future Considerations

The DPIIT acknowledges the feedback from firms and plans to consider these inputs as the consultation process continues.