Copyright Act likely to be amended to address AI issues within three years, DPIIT official says

The Union government is working to make changes to copyright law, necessitated by the new demands of artificial intelligence, within the next three years, a senior official said on Thursday (December 11, 2025).

The Department for Promotion of Industry and Internal Trade (DPIIT) released a working paper on AI and copyright issues last week. It suggested a “blanket licensing” framework, where websites whose data is crawled by large language models (LLMs) like ChatGPT would receive royalty payouts through a copyright society which would split the money between them. 

The proposal aims to resolve the growing tensions worldwide between online content publishers (such as book publishers and news organisations) and AI firms, who “train” their LLMs with a large amount of text data, usually scraped from the public internet. Publishers have argued that they need to be compensated because of their contribution in developing and improving these models. 

The DPIIT proposal suggests generally allowing the scraping of content by AI developers while ensuring that they eventually pay content publishers.

Payout post-commercialisation

This will be followed in about two months by another working paper that will look into whether AI-generated works are copyrightable, and how authorship is decided for them, DPIIT Additional Secretary Himani Pande said in a news briefing. Following this, the government is likely to move an amendment to the Copyright Act, 1957 in Parliament to establish the new regime.

Ms. Pande said that the copyright society, known as the Copyright Royalties Collective for AI Training (CRCAT), would only seek payments from AI firms once they commercialise their models, as opposed to when they are mining data from the internet to develop or train them.

The compensation of copyright holders has been a contentious issue, with news publishers in particular suing AI firms around the world. Newswire agency ANI (from which The Hindu syndicates video content) and The New York Times are among those who have sued ChatGPT developer OpenAI for allegedly regurgitating their content in chat conversations. OpenAI has denied their accusations.

AI firms dissent

Tech industry body Nasscom, which had one seat in the committee that drafted this report, dissented from the model proposed by the DPIIT. The body, which represents Google, Meta, Amazon, and other large firms with significant AI investments, said that publishers should be able to opt out of including their data in training models, warning that a blanket licensing model could expose firms to further disputes.

In a note shared with The Hindu, a Big Tech firm that develops AI models raised specific concerns about the DPIIT proposal. “As per stated copyright jurisprudence, the burden of proof to establish infringement is on the copyright owner who is making the claim,” the note says. “The Hybrid proposal reverses this. It is well established in copyright law that in a copyright infringement suit, the plaintiff (content owner) must prove that the defendant (AI developer) [infringed their copyright]. If the burden of proof is on the AI Developer then they have to prove that they have not used the content owner’s material even if the output is similar. This is extremely onerous and technically infeasible because Gen AI tools are probabilistic and not deterministic.”

Ms. Pande said that firms’ inputs would be taken into consideration as the consultation process continues.